AIN SHAMS UNIVERSITY FACULTY OF ENGINEERING ... Thesis...AIN SHAMS UNIVERSITY FACULTY OF ENGINEERING...

AIN SHAMS UNIVERSITY

FACULTY OF ENGINEERING

IRRIGATION AND HYDRAULICS DEPARTMENT

MAPPING THE SPATIAL DISTRIBUTION OF WATER

QUALITY PARAMETERS IN LAKE NASSER USING

REMOTE SENSING TECHNIQUE

By

Eng. Mohammed Ali Hamed Ghareeb (B.Sc., Civil Engineering, Shoubra Faculty of

Engineering, Benha University, 2003)

(M.Sc., Civil Engineering Ain shams University, 2010)

A Thesis Submitted in Partial Fulfillment of the Requirements

for the Degree of Doctor of Philosophy in Civil Engineering

Supervised by

Prof. Dr. Aly Nabih El-Bahrawy Professor of Hydraulics

Irrigation and Hydraulics Department

Faculty of Engineering, Ain Shams University

Prof. Dr. Karima Mahmoud Attia Director

Water Resources Research Institute

National Water Research Center

Dr. Mohsen Mahmoud Yousry Secretary General

Nile Research Institute


Cairo, Egypt

2016

FACULTY OF ENGINEERING

IRRIGATION AND HYDRAULICS

MAPPING THE SPATIAL DISTRIBUTION OF WATER

QUALITY PARAMETERS IN LAKE NASSER USING

REMOTE SENSING TECHNIQUE

By

Eng. Mohammed Ali Hamed Ghareeb (B.Sc., Civil Engineering, Zagazig University, 2003)

(M.Sc., Civil Engineering, Ain shams University, 2010)

Examiners Committee

Signature

Prof. Dr. Medhat Saad Aziz ………….…. Director of Nile Research Institute


Ministry of Water Resources and Irrigation

Prof. Dr. Iman Elazizy ………….…. Professor of Hydraulics



Prof. Dr. Aly Nabih El-Bahrawy ………….…. Professor of Hydraulics



Prof. Dr. Karima Mahmoud Attia ………….…. Director of Water Resources Research Institute


Ministry of Water Resources and Irrigation

Date: -----/------/2016

(109سورة )الكهف( آية )

To my dear wife

To my lovely sons Ali and Ahmed

To my father and my mother

To my brother Ahmed and his family

STATEMENT

This dissertation is submitted to Ain Shams University for partial

fulfillment of the requirements for the Degree of Doctor of Philosophy in

Civil Engineering

The work included in this thesis was carried out by the author in the

department of Irrigation and Hydraulics, Ain Shams University.

No part of this thesis has been submitted for a degree or a qualification at

any other university or institution.

Date:

Name: Mohammed Ali Hamed

Signature:

CURRICULUM VITAE

Name

Mohammed Ali Hamed Ghareeb

Date of Birth

December, 06th, 1980

Present Position

Assistant Researcher, Nile Research Institute,

National Water Research Center.

Education

- From 1988 to 1992 Primary School

- From 1993 to 1995 preparatory School

- From 1996 to 1998 Secondary School

- From 1999 to 2003 Shoubra Faculty of

Engineering, Benha University

Degree Awarded

B.Sc. In Civil Engineering, Shoubra Faculty of

Engineering, Benha University, Egypt 2003. With

general grade Very Good with honor degree.

M.Sc. In Civil Engineering, Irrigation and Hydraulics,


Ain Shams University

Faculty of Engineering

Irrigation and Hydraulics

Name: Mohammed Ali Hamed Ghareeb

Thesis: Mapping the Spatial Distribution of Water Quality Parameters in

Lake Nasser Using Remote Sensing Technique

Abstract The main objective of this research is to use remote sensing in estimating

water quality parameters in Lake Nasser as it is considered the main and

strategic storage of fresh water in Egypt. Results of six field missions,

measured in different seasons, are used with Envisat/MERIS match-up

images to validate Case 2 water quality processors (Case2 Regional

(C2R), Eutrophic Lake and boreal Lake), with and without correcting

land adjacency effect using Improved Contrast between Ocean and Land

(ICOL) processor, to estimate optical parameters, Total Suspended Solids

(TSS) and Chlorophyll-a (Chl-a). Processors’ validation is based on using

of some statistical measures. For TSS, results show that, Eutrophic lake

processor without ICOL is suitable for the end of flood season, C2R and

eutrophic lake processor without ICOL is suitable for the falling period,

while all processors failed during rising period. For Chl-a, Eutrophic lake

processor with ICOL indicates good results during falling period and the

end of flood season. Also, regression models are created to estimate

optical and non-optical water quality parameters in different flood

seasons using atmospherically corrected MERIS images, where bands 5

and 6 are found to be the common predictors. Models are validated at the

end of flood season and applied to 913 images for period from 2002 to

2012 to create time series curves/maps for each parameter, after

excluding cloudy images. It is concluded that TSS and Total Phosphorus

concentrations indicate increasing until 2005 and started to decrease

again, Silica and TDS illustrate an opposite changing pattern and

transparency doesn’t have a specific pattern of change. It is also

concluded that period from September to November indicate the lowest

cloud probability and is suitable for preforming optical measurement

which helps in creating bio-optical model for Lake Nasser.

Keywords: Remote Sensing, Water Quality, Lake Nasser, Case2 Water

Quality Processors, regression models.

Supervisors : Prof. Dr. Aly Nabih El-Bahrawy

Prof. Dr. Karima Mahmoud Attia

Dr. Mohsen Mahmoud Yousry

ACKNOWLEDGEMENT

Praise is to Allah, Lord of the Worlds.

I will forever be grateful to many people either directly or indirectly for

the completion of this thesis.

I wish to express my deepest sense of gratitude and sincerest appreciation

to Prof. Dr. Aly El-Bahrawy, Professor of Hydraulics, Faculty of

Engineering, Ain Shams University, for his excellent advise, enthusiastic

guidance and continuous encouragement towards the successful

completion of this study.

This thesis would never have been completed without the supervision and

nurturing of my advisor, Prof. Dr. Karima Attia, Director, Water

Resources Research Institute, National Water Research Center. She

contributed tremendously to my scientific development. She always

challenged me to produce my best.

This thesis would never have been completed without the supervision and

nurturing of my advisor, Dr. Mohsen Yousry, Secretary-General, Nile

Research Institute, National Water Research Center.

I would also like to thank Prof. Dr. Medhat Aziz, Director of Nile

Research Institute, National Water Research Center, for his kind

assistance and providing the hardware and software facilities used to

perform this study.

I would like to express my deep thanks and appreciation to my dear wife,

Doha and my lovely sons, Ali and Ahmed, the source of my happiness,

encouragement and motivation.

I wish to express my deepest thanks, gratitude, and appreciation to my

mother and my father for their love, warm caring, support, praying for me

and great patience throughout the time of this study.

Last, I wish to express my deepest thanks, gratitude, and appreciation to

my twin brother Ahmed for his encouragement and for his support to

continue this thesis.

i

TABLE OF CONTENTS TABLE OF CONTENTS .........................................................................i

LIST OF FIGURES ................................................................................. v

LIST OF TABLES ..................................................................................ix

LIST OF ABBREVIATIONS ................................................................xi

LIST OF SYMBOLS ........................................................................... xiii

CHAPTER (1) INTRODUCTION ....................................................... 1

1.1. General ........................................................................................ 1

1.2. Problem definition ....................................................................... 1

1.2.1. Sedimentation in Lake Nasser .............................................. 1

1.2.2. Water quality and suspended sediment ................................ 4

1.3. Study objectives .......................................................................... 5

1.4. Outline of the thesis..................................................................... 5

CHAPTER (2) LITERATURE REVIEW ........................................... 9

2.1. Introduction ................................................................................. 9

2.2. Remote Sensing Theory .............................................................. 9

2.2.1. Remote Sensing History ....................................................... 9

2.2.2. Principles of Remote Sensing ............................................. 10

2.2.3. Types of Remote Sensing ................................................... 12

2.2.4. Scale and Resolution .......................................................... 15

2.2.5. Interaction of EMR with the earth’s surface ...................... 16

2.3. Optics of water .......................................................................... 18

2.3.1. Interaction between light and water .................................... 19

2.3.2. Case 1 and Case 2 waters ................................................... 21

2.4. Remote Sensing and water quality ............................................ 21

2.4.1. Retrieval approaches ........................................................... 21

2.4.2. Satellite Missions ................................................................ 22

2.5. Application ................................................................................ 24

2.5.1. Suspended Sediments ......................................................... 25

2.5.2. Turbidity and Transparency ................................................ 25

2.5.3. Chlorophyll-a ...................................................................... 26

ii

2.5.4. Total Phosphorus (TP) and Total Nitrogen (TN) ............... 27

2.6. Remote sensing studies in Lake Nasser .................................... 29

2.7. Discussion ................................................................................. 30

CHAPTER (3) STUDY AREA AND DATA COLLECTION ......... 31

3.1. General ...................................................................................... 31

3.2. Study Area ................................................................................. 31

3.2.1. Aswan High Dam (AHD) ................................................... 31

3.2.2. Aswan High Dam Reservoir (AHDR) ................................ 33

3.2.3. AHDR morphology ............................................................ 33

3.2.4. AHDR operation policy ...................................................... 35

3.2.5. AHDR Monitoring Activities ............................................. 35

3.2.6. AHDR hydrological characteristics .................................... 38

3.3. Data Collection .......................................................................... 42

3.3.1. Field data ............................................................................ 42

3.3.2. Remote sensing data ........................................................... 53

CHAPTER (4) METHODOLOGY .................................................... 57

4.1. General ...................................................................................... 57

4.2. Field work ................................................................................. 57

4.3. Satellite images acquisition ....................................................... 57

4.3.1. Image processing tools ....................................................... 58

4.3.2. MERIS Processing tool (MPT) ........................................... 60

4.4. Validation of water quality processors ...................................... 61

4.5. Regression ................................................................................. 62

4.6. Time Series ................................................................................ 63

4.7. Statistical performance measures .............................................. 64

4.7.1. Definitions .......................................................................... 65

4.7.2. Interpretation of Performance measures ............................. 67

4.7.3. Criteria for good processor ................................................. 67

4.7.4. Ranking and Evaluation ...................................................... 68

CHAPTER (5) RESULTS AND DISCUSSION ................................ 69

5.1. Introduction ............................................................................... 69

5.2. Validation of Case 2 water quality processors .......................... 69

iii

5.2.1. Validation of TSS ............................................................... 69

5.2.2. Validation of Chl-a ............................................................. 82

5.3. Regression ................................................................................. 87

5.3.1. Correlation between measured parameters ......................... 87

5.3.2. Stepwise regression Model ................................................. 90

5.3.3. Validation of retrieval model .............................................. 95

5.4. Time series analysis ................................................................ 100

5.4.1. Image selection criteria ..................................................... 100

5.4.2. Flood seasons .................................................................... 103

5.4.3. Time series of Water quality parameters .......................... 105

CHAPTER (6) CONCLUSION AND RECOMMENDATIONS .. 113

6.1. Conclusion ............................................................................... 113

6.2. Recommendations ................................................................... 117

REFERENCES ................................................................................... 119

APPENDIX (A) MERIS IMAGES CHARACTERISTICS AND

WATER QUALITY ANALYSIS ....................................................... 125

A.1. MERIS image characteristics .................................................. 125

A.1.1. MERIS product processing levels .................................... 126

A.1.2. Resolutions ....................................................................... 126

A.1.3. MERIS Case 2 water quality processors .......................... 129

A.2. Water Quality Analysis ........................................................... 130

A.2.1. Water sampling ....................................................................... 130

A.2.2. Preservation ...................................................................... 130

A.2.3. Types of analysis .............................................................. 130

A.2.4. Laboratory measurements ................................................. 131

A.2.5. Reagents ............................................................................ 131

A.2.6. Equipment ......................................................................... 131

A.2.7. Laboratory equipment ....................................................... 132

A.2.8. General equipment ............................................................ 132

APPENDIX (B) MERIS PROCESSING TOOL (MPT) AND EXCEL

CUSTOM FUNCTIONS ..................................................................... 133

B.1. Introduction ............................................................................. 133

iv

B.2. MERIS PROCESSING TOOL (MPT) .................................... 133

B.2.1. Main Idea of MPT ............................................................ 133

B.2.2. MPT description ............................................................... 134

B.2.3. MPT Code ......................................................................... 138

APPENDIX (C) Stepwise Regression charts ..................................... 171

v

LIST OF FIGURES

Figure ( 1-1) The Nile River Basin ....................................................... 2

Figure ( 1-2) Flow hydrograph of the main Nile and its tributaries ..... 3

Figure ( 1-3) Lake Nasser in two different seasons .............................. 4

Figure ( 2-1) Remote Sensing components (Miloud, 2012) ............... 11

Figure ( 2-2) The electromagnetic spectrum, (Tantirimudalige, 2002)

........................................................................................................... 12

Figure ( 2-3) The energy interactions, (Liliesand & Kiefer, 1993) .... 16

Figure ( 2-4) the spectral reflectance curves for clear water, vegetation

and bare soil, (Tantirimudalige, 2002) .............................................. 18

Figure ( 2-5) The relation between hydrologic optics and other

branches of optics (after Preisendorfer, 1976 and Kirk, 1994) ......... 19

Figure ( 2-6) Time line for Ocean color Sensors from 1978 to 2030

(CEOS, 2016) .................................................................................... 23

Figure ( 3-1) Layout of Aswan High Dam ......................................... 33

Figure ( 3-2) Lake Nasser Map........................................................... 34

Figure ( 3-3) Main component for different operation zones of AHDR

........................................................................................................... 35

Figure ( 3-4) Water level Upstream AHD from 1964 to 2011 ........... 40

Figure (3-5) AHDR inflow at Dongola station .................................. 40

Figure ( 3-6) Discharge and sediment concentration hydrographs at

Dongola station (1999-2003) ............................................................. 41

Figure ( 3-7) Water Quality samples’ Locations ................................ 44

Figure ( 3-8) Survey mission with respect to water level change in

period from 2002 to 2011 .................................................................. 45

Figure ( 3-9) Longitudinal Profile for TSS during different missions 47

Figure ( 3-10) Longitudinal Profile for Chl-a during different missions

........................................................................................................... 48

Figure ( 3-11) Longitudinal Profile for Transparency during different

missions ............................................................................................. 48

Figure ( 3-12) Longitudinal Profile for Turbidity during different

missions ............................................................................................. 49

vi

Figure ( 3-13) Longitudinal Profile for TP during different missions 50

Figure ( 3-14) Longitudinal Profile for SiO2 during different missions

........................................................................................................... 50

Figure ( 3-15) Longitudinal Profile for TDS during different missions

........................................................................................................... 51

Figure ( 3-16) Histogram of downloaded MERIS images ................. 56

Figure ( 4-1) Flow Chart of Research Methodology .......................... 58

Figure ( 4-2) Time series flowchart .................................................... 64

Figure ( 5-1) TSS for First Mission (mg/l) ......................................... 71

Figure ( 5-2) TSS for Third Mission (mg/l) ....................................... 72

Figure ( 5-3) TSS for Fourth Mission (mg/l) ..................................... 73

Figure ( 5-4) TSS for Fifth Mission (mg/l) ........................................ 74

Figure ( 5-5) TSS for Sixth Mission (mg/l) ........................................ 75

Figure ( 5-6) Measured verses predicted TSS Values for First, Third

and Fourth missions (mg/l) ................................................................ 76

Figure ( 5-7) Measured verses predicted TSS Values for Fifth and

Sixth missions (mg/l) ......................................................................... 77

Figure ( 5-8) Chl-a for Third Mission (mg/m3) .................................. 83

Figure ( 5-9) Chl-a for Fourth Mission (mg/m3) ................................ 84

Figure ( 5-10) Measured verses predicted Chl-a Values (mg/m3) ...... 85

Figure ( 5-11) Cloud probability summary surface of Lake Nasser . 102

Figure ( 5-12) Percentage of excluded images due to clouds and sun

glint .................................................................................................. 103

Figure ( 5-13) Percentage of excluded images due to clouds and sun

glint .................................................................................................. 104

Figure ( 5-14) Monthly average TSS at Arkeen Cross section as

calculated from MERIS Images ...................................................... 105

Figure ( 5-15) Spatial distribution maps of monthly average

concentration of TSS, Transparency, and TP (mg/l) ....................... 107

Figure ( 5-16) TSS during Period from October to December (mg/l)

......................................................................................................... 108

vii

Figure ( 5-17) Transparency during Period from October to December

(mg/l) ............................................................................................... 109

Figure ( 5-18) Silica during Period from October to December (mg/l)

......................................................................................................... 110

Figure ( 5-19) Total Phosphorus during Period from October to

December (mg/l) .............................................................................. 111

Figure ( 5-20) TDS during Period from October to December (mg/l)

......................................................................................................... 112

Figure ( A-1) Envisat Satellite Configuration and Payload Instruments

(ESA, 2006) ..................................................................................... 125

Figure ( A-2) MERIS Global coverage (ESA, 2006) ....................... 129

Figure ( B-1) Sample of A processing graph .................................... 134

Figure ( B-2) Flow chart of MPT programming............................... 135

Figure ( B-3) MPT interface ............................................................. 137

Figure ( B-4) ArcMap Model Builder used to cloud coverage

percentage ........................................................................................ 168

Figure ( B-5) ArcMap Model Builder used to Extract Lake Nasser

Surface ............................................................................................. 169

Figure ( C-1) Stepwise Regression Analysis Results For 2003 ....... 172






ix

LIST OF TABLES

Table ( 2-1) Selected remote sensing application for monitoring

Suspended Sediment in different water bodies (Chang et al., 2015) ....... 26

Table ( 2-2) Selected remote sensing application for monitoring

Chlorophyll-a in different water bodies (Chang et al., 2015) .................. 27

Table ( 2-3) Selected remote sensing application for monitoring Nutrients

(TN and TP) in different water bodies (Chang et al., 2015) .................... 28

Table ( 3-1) Definition of main monitoring locations surveyed (El

Sammany, 2002) ...................................................................................... 37

Table ( 3-2) Measured Parameters for each mission ................................ 43

Table ( 3-3) Descriptive statistics of measured water quality parameters 52

Table ( 3-4) Characteristics of Match-up images. .................................... 55

Table ( 5-1) Summary of statistical performance measures for TSS ........ 78

Table ( 5-2) Ranking values of water quality processors for TSS ............ 81

Table ( 5-3) Summary of statistical performance measures for Chl-a ...... 86

Table ( 5-4) Ranking values of water quality processors for Chl-a .......... 86

Table ( 5-5) Pearson correlation matrix .................................................... 89

Table ( 5-6) Rule of Thumb for Interpreting the Size of a Correlation

Coefficient (Hinkle et al., 2003) .............................................................. 90

Table ( 5-7) Retrieval Models for water quality parameters (First, Second

and third missions) ................................................................................... 93

Table ( 5-8) Retrieval Models for water quality parameters (Fourth, Fifth

and Sixth missions) .................................................................................. 94

Table ( 5-9) Retrieval Models Validation for different seasons (TSS, Chl-

a, Turbidity and transperancy) ................................................................. 98

Table ( 5-10) Retrieval Models Validation for different seasons (TP, SiO2,

and TDS) .................................................................................................. 99

Table ( 5-11) Start/End of Flood Seasons ............................................... 104

Table ( A-1) MERIS products description (ESA, 2006)......................... 127

Table ( A-2) MERIS spectral bands and applications (ESA, 2006). ...... 128

xi

LIST OF ABBREVIATIONS

TSS Total Suspended Solids

Chl-a Chlorophyll- a

TDS Total Dissolved Solids

TP Total Phosphorus

AHD Aswan High Dam

AHDR Aswan High Dam Reservoir

C2R Case 2 Regional

MPT MERIS Processing Tool

CIR Color Infra-Red

MSS Multispectral Scanner

TM Thematic Mapper

MR Microwave Radiometer

SAR Synthetic Aperture Radar

EMR Electromagnetic Radiation

NIR Near Infrared

ERTS Earth Resources Technology Satellite

CZCS Coastal Zone Color Scanner

OCTS Ocean Color Temperature Scanner

POLDER POLarization and Directionality of the Earth's

Reflectances

SeaWiFS Sea-viewing Wide Field-of-view Sensor

MERIS MEduim Resolution Imaging Spectrometer

MODIS Moderate Resolution Imaging Spectroradiometer

xii

SPM Suspended Particulate Matter

Rrs Remote Sensing Reflectance

GIS Geographic Information System

MSL Mean Sea Level

DGPS Differential Global Positioning System

ESA European Space Agency

DDS Data Dissemination System

ICOL Improved Contrast between Ocean and Land

SMAC Simplified Method for Atmospheric Corrections

APHA America Public Health Association

FB Fractional Bias

r Correlation Coefficient

NMSE Normalized Mean Square Error ()

MG Geometric Mean Bias

MV Geometric Mean Variance

Fac2 Factor of two

NDWI Normalized Difference Water Index

MNB Mean Normalized Bias

SSC Suspended Sediment Concentration

MIM Matrix Inversion Method

TN Total Nitrogen

TOC Temperature Total Organic Carbon

xiii

LIST OF SYMBOLS

EI () Incident Energy joule

ER () Reflected Energy joule

EA () Absorbed Energy joule

ET () Transmitted Energy joule

ρ (λ) Spectral reflectance ------

Kd The vertical attenuation coefficient m−1

c () The beam attenuation coefficient m−1

Ed (z, ) Depth of the ambient downwelling irradiance W/m2

1

CHAPTER (1)

INTRODUCTION

1.1. General

Traditional water sampling methods are able to quantify temporal

changes in water quality at specific points. However, they are unable to

effectively quantify water quality across the entire surface area. In lakes

with high spatial variation, like Lake Nasser, point samples are often not

representative of the whole water body, which may mislead the lake

management process. Also, due the expense of traditional monitoring

programs, often, only critical locations are selected for monitoring. The

advantage of remote sensing is that it can cover large areas of both

waterbodies and ground; hence water quality can be estimated over vast

greater area.

1.2. Problem definition

Lake Nasser is considered the main and strategic storage of fresh water in

Egypt, so that, the preservation and continues monitoring of water quality

in lake is very important and represent a national goal. Although the

water quality is still good, sediment load comes with flood and human

activities such as agriculture, fisheries and transport threaten lake water

quality. So that, more attention should be bayed to monitor, protect and

manage water resources in Lake Nasser using a scientific approach.

Remote sensing technique is considered a powerful tool for assessment

and monitoring of lake condition and can supply timely based and

reliable information to help in decision making.

1.2.1. Sedimentation in Lake Nasser

The Nile River water flow is received from three main distinct

watersheds, Figure ( 1-1); Equatorial lakes plateau in the South, Sudd

region in the Center, and Ethiopian Highlands in the East. Figure ( 1-2)

illustrates the nominal flow hydrographs of White Nile, Blue Nile, and

2

Atbara River in addition to the main Nile. These hydrographs show the

average flows over more than 80 years. It is noted that the coming flows

from Blue Nile is highly suspended with sediments, especially during

summer months. This case can contribute in changing the water quality in

Lake Nasser throughout the year.

Figure (‎1-1) The Nile River Basin

The Nile River has two distinct hydrological phenomena; a short high

flow period takes about three months represents the flood season, the

flow comes from Blue Nile and Atbara river represent the main source of

3

water in Lake Nasser, in this period water is muddy and carries a large

quantity of sediments, about 95% of sediment load in Lake Nasser; and a

long period takes about nine months, water in this period is clear and the

sediment concentration is low, the sediment load coming from the White

Nile and tributaries is relatively small about 5% of the total annual

sediment load of the main river.

Figure (‎1-2) Flow hydrograph of the main Nile and its tributaries

Two remote sensing images (MERIS images) are shown in Figure ( 1-3)

representing lake in two different seasons. The left image is representing

the lake in 05 May 2010 shows lake in low flow period and shows a clear

4

water with no suspended sediments, the right image represents lake in 03

September 2010 which shows in flood period, the sediment carried by

water can be obviously observed from the remote sensing image in the

southern part of lake.

05 May, 2010 (Before flood) 03 September, 2010 (After flood)

Figure (‎1-3) Lake Nasser in two different seasons

1.2.2. Water quality and suspended sediment

Suspended Sediment in water has two opposite effects on water quality.

The first effect is that the suspended sediments in water, especially the

finer ones, absorb some pollutants such as phosphorus, nitrogen, organic

compounds and pathogenic, that can help in improving water quality to a

certain degree. The second effect is that sediment can be source of

pollutant, as it can carry and store other pollutants, such as bacteria,

pesticides, residues and viruses which will affect the water purity,

transparency and quality (Yang, 2003).

5

1.3. Study objectives

The main objectives of this study are to develop a method using remote

sensing techniques to be used in spatial mapping of water quality

parameter to monitor water quality in Lake Nasser. The Medium

Resolution Imaging Spectrometer (MERIS) will be used to fulfill the

research objectives.

The objectives can be summarized in the following points:

- Validate the use of the existing Case 2 water quality processors

(Case 2 Regional, Boreal Lakes, and Eutrophic lakes) in lake

Nasser, and to what extent it can be applied in different seasons to

obtain optical water quality parameters such as Total Suspended

Solids (TSS), Chlorophyll-a (Chl-a).

- Determining a regression model that correlates different water

quality parameters and remote sensing reflectance of different

MERIS image bands.

- Time series analysis and Spatial and temporal mapping of water

quality parameters to determine seasonal variations.

- A cloud coverage and sun glint analysis for the lake Nasser area

will be performed to find the most suitable months for coupling

between RS and water quality; this will be helpful for future

researches.

1.4. Outline of the thesis

CHAPTER (1): Introduction

This chapter contains a general introduction of problem statement, effect

of sedimentation and different water quality parameters, research

objectives and contents of every chapter.

CHAPTER (2): Literature review

This chapter gives a literature review for using remote sensing in

assessment and monitoring, especially for water quality of water bodies

6

all over the world, also gives a literature about the previous studies for

Lake Nasser with remote sensing such as evaporation.

CHAPTER (3): Study area and data collection

This chapter will contain a description of Lake Nasser as a study area; its

location and the current monitoring program, in addition to its physical

characteristics, hydrologic and sedimentation characteristics. Also, it

contains a description of both field and remote sensing data used in the

research

CHAPTER (4): Methodology

This chapter will give a full description of the methodology used

throughout the research in both validation of case 2 water quality

processors and regression analysis between water quality parameters and

different bands reflectance, in addition to description of tools used in

through the current research. Also, it will contain a description of

statistical performance measures used to evaluate results.

CHAPTER (5): Results and discussion

This chapter illustrates in details the results obtained through the research

for validation of Case 2 water quality processor to estimate TSS and Chl-

a in Lake Nasser through different seasons, the results of stepwise

regression will be illustrated to get a regression model for each water

quality parameter and time series analysis. The chapter also contains

analysis of cloud coverage percentage over the surface area of the lake

which will help in excluding the cloudy images from time series analysis;

also it can help in identifying the best month for coupling water quality

measurements and remote sensing. Also, the chapter will contain results

of time series analysis for different water quality parameters.

CHAPTER (6): Conclusion and recommendations

This chapter summarizes the study scope and procedures in addition to

most important conclusions. Further studies and future works are also

recommended in this chapter.

7

REFERENCES

This chapter includes the biography and references used in the study

APPENDIX (A): MERIS Images Characteristics and water quality

analysis

It contains a description of MERIS image characteristics in means of its

spatial and temporal resolution, its processing levels and common uses of

it. In addition to a detailed description of Case 2 water quality processors

used through the research. It also includes methods and standards used in

sampling.

APPENDIX (B): MERIS Processing Tool (MPT) and Excel Custom

Functions

It contains a description of MERIS Processing Tool; which have been

programed in order to automate the processing of MERIS images. Also it

contains the programing code for excel custom functions to calculate

different statistical measures. In addition to some tools created by

ArcMap to automate the processing of images.

APPENDIX (c): Stepwise Regression charts.

9

CHAPTER (2)

LITERATURE REVIEW

2.1. Introduction

The aim of this chapter is to give a detailed review of remote sensing

theory and its application in lakes/reservoirs, rivers, estuaries and coastal

zones, especially in water quality. It will also give a review about the

researches pertain the use of remote sensing techniques in Lake Nasser.

2.2. Remote Sensing Theory

Remote sensing, also called earth observation, is a technique used to have

information about objects by collecting and analyzing data without being

in direct contact with the object or area. In general, remote sensing can be

defined as a tool used for observing and studying the Earth from space, it

can observe land surface, seas and oceans, the atmosphere and its

dynamics. It offers a means of obtaining large amounts of data, but its

value is in the information that is obtained from the acquired data and

how it is used, stored and expanded in the spatial, temporal and spectral

modes.

2.2.1. Remote Sensing History

Earth observation is an activity that started a long time ago, may be with

the land surveys conducted physically using various measuring

instruments. Gradually observation from air was possible and then

developed to satellite observation, (Tantirimudalige, 2002).

The first aerial photograph was taken by the Parisian photographer called

Gaspard Tournachon in 1858, a, from a balloon at a height of 80 m. The

earliest existing aerial photograph was taken in 1860 from a balloon over

Boston. In order to obtain meteorological data kites were used to obtain

aerial photographs from about 1882.

During World War I, Photography from aircraft received highest

attention especially in the interest of military reconnaissance to observe

10

enemy troop movements. Aerial photographs were used, during World

War II, to map sea bed conditions along the coastline to better plan the

maneuvering of landing craft. Furthermore, infrared film made it possible

to distinguish between green vegetation camouflage nets.

After the wars in the 1950s, color infra-Red (CIR) photography was

found. In 1956, Colwell conducted experiments on the use of CIR to

recognize and classify of vegetation types and to detect the diseased and

stressed or damaged vegetation. The significant progress in radar

technology was found also in 1950s. Hence, Studying the Earth from

space developed from pure research, to even daily applications.

The first satellite to be used for observing and monitoring of the earth

was launched in July, 1972. Early indications from NASA say that the

satellite is sending good pictures back to Earth, (Mitsch, 1973).

Today, remote sensing satellites have several applications such as

Weather, Environmental monitoring, water quality and Mapping.

2.2.2. Principles of Remote Sensing

In Remote Sensing, four elements are essential, see Figure ( 2-1):

1- Platform to hold instruments, for example, aircrafts and satellites.

2- A target object to be observed, the target is the earth.

3- An instrument or sensor to observe the Earth such as cameras,

scanners, radars, etc.

4- The information obtained from analysis of acquired data can help

in increasing our knowledge about the earth such as the cloud

cover, land use, and land cover, characteristics of waterbodies,

and much more).

11

Figure (‎2-1) Remote Sensing components (Miloud, 2012)

Detection and classification of different objects on the earth surface

means detecting and recording of radiant energy emitted or reflected by

these objects or surface material, which depends on the physical,

chemical, and structural properties of material, surface roughness and

radiant energy characteristics such as, angle of incidence, intensity, and

wavelength of radiant energy.

A general understanding of the electromagnetic radiation (EMR) and

electromagnetic spectrum is necessary to consider the technology of

remote sensing. EMR can be defined as the energy propagated through

space between electric and magnetic fields. Visible light was shown by

Maxwell as a part of the electromagnetic spectrum which also contains

gamma rays, X-rays, ultraviolet rays, infrared rays, microwaves, radar

and radio waves. Figure ( 2-2)

In general when electromagnetic energy hits any matter, the energy is

conserved according to basic physical principles in one or more of the

following states:

- Absorbed energy by the matter and it contributes in heating it

12

- Emitted energy, or commonly remitted energy, it is depended on

the structure and temperature of the matter. It has different values

depending on the matter and wave length.

- Scattered, deflected to one side and lost ultimately to absorption

of further scatter, or

- Reflected, returned unchanged to the medium (Colwell et al.,

1963).

Figure (‎2-2) The electromagnetic spectrum, (Tantirimudalige, 2002)

2.2.3. Types of Remote Sensing

Remote sensing can be classified according to either energy source or

sensor type into passive or active systems. Active systems use their own

energy source such as microwaves (radar), where the passive systems

depend upon external source of illumination, such as sun or self-emission

of the observed object.

Also remote can be classified according to the wavelengths into three

categories as follows:

1- Visible and near infrared Remote Sensing, this type can be

classified as a passive system where it depends on the reflected

energy from the sun. It helps in understanding of land surface

conditions such as vegetation, rivers, lakes, and urban areas

13

distribution. Also it can be used in field of water quality

monitoring.

2- Thermal Infrared Remote Sensing, this type classified as a passive

system where it depends on collecting the thermal infrared rays

radiated from land surface heated by sunlight. It can be used in

identifying the high temperature area such as volcanic activities

and fires areas.

3- Microwave (radar) Remote Sensing, this type can be classified as

either active or active microwave systems. In active microwave

system, the observation satellite emits microwaves and observes

microwaves reflected by different object in the earth, this type is

suitable to observe mountains and valleys. While the passive

microwave system measures the naturally radiated microwaves

from land surface, this type can be useful in observing sea surface

temperature, snow accumulation and thickness of ice.

There are many types of both passive and active sensors; each type has its

own application, the different types will be illustrated below as listed in

(NASA, 2015).

Passive sensors include different types of radiometers and spectrometers

as follows:

- Accelerometer: it is an instrument that able to measure both

translational or angular acceleration.

- Radiometer: this instrument can quantitatively measure the

intensity of electromagnetic radiation in some bands within the

spectrum.

- Imaging radiometer: it is a radiometer that can provide a two-

dimensional array of pixels to produce an image. It uses an array

of detectors for Scanning images and it can be performed

mechanically or electronically.

- Spectrometer: it is used to detect, measure, and analyze the

spectral characteristics of incident electromagnetic radiation.

14

- Spectroradiometer: it is a radiometer that can measure the

intensity of radiation in multiple wavelength bands (i.e.,

multispectral).

- Hyperspectral radiometer: it is an advanced multispectral

sensor and can detect hundreds of very narrow spectral bands in

the visible, near-infrared, and mid-infrared portions of the

electromagnetic spectrum.

- Sounder: it measures vertical distributions of atmospheric

parameters such as temperature, pressure, and composition from

multispectral information

Most active sensors operate in the microwave portion of the

electromagnetic spectrum, which enables them to penetrate the

atmosphere under most conditions. The below list is describing the

different types of active sensors:

- Radar: it can be an airborne or spaceborne which emits a series

of microwave pulses from an antenna. It detects and measures the

backscattered microwaves reflected back to sensor. The distance

or range to the target can be determined using time required for

the energy to travel to the target and return back to the sensor. A

two dimensional image can be produced by recording both range

and magnitude of the reflected energy from all targets.

- Ranging Instrument: this system uses a different technique which

depending on a pair of identical microwave instruments on two

different platforms. Signals are transmitted from each instrument

to the other, with the distance between the two determined from

the difference between the received signal phase and transmitted

(reference) phase.

- Scatterometer: it is high-frequency microwave radar used

especially to measure backscattered radiation. It can be used to

derive maps of surface wind speed and direction over ocean

surfaces.

15

- Lidar: it is a system used in light detection and works on the

principle of radar, but uses light from a laser.

- Laser altimeter: it is an instrument that uses a LIDAR to

measure the height of the platform above the mean Earth’s

surface; it can be used to determine the topography of the

underlying surface.

- Sounder: it can measure the vertical distribution of precipitation

and other atmospheric characteristics such as temperature,

humidity, and cloud composition

Today, there is a big list of remote sensing satellites that use passive and

active sensors. They are now onboard which provide the scientists with

hundreds of daily images for earth surface with different spatial, spectral,

temporal, and radiometric resolution

2.2.4. Scale and Resolution

Scale and resolution are very important properties of both aerial

photographs and satellite images, these properties can be calculated by

characteristics of lens and the platform flying height.

- Scale

Remote sensing information is scale dependent. Scale of remote sensing

image can depend on many factors such as the effective focal length of

the optical device used in acquiring image, altitude of the remote sensing

satellite and the factor of magnification used to reproduce the image.

- Resolution

The acquired remote sensing images can be described by three different

types of resolutions such as spatial, spectral and radiometric resolution.

The fourth one is depending on the space vehicle which is the temporal

resolution.

Four types of resolutions considered in remote sensing work will be

described as follows:

16

1- Spatial Resolution is by the minimum object size that can be

detected or recorded by sensor.

2- Spectral Resolution is defined by the used channel bandwidth.

High spectral resolution was characterized by a narrow

bandwidth; it provides more accurate discrimination of objects.

3- Radiometric Resolution is determined by the number of discrete

levels into which signal strength maybe divided (quantization).

4- Temporal Resolution is can be defined as the time interval

between two successive visits at the same place by the remote

sensing satellite.

2.2.5. Interaction‎of‎EMR‎with‎the‎earth’s‎surface

The incident radiation from the sun can be reflected either by the surface,

transmitted into the surface or absorbed and emitted by the surface,

Figure ( 2-3). Hence a several changes were happened to the EMR in its

direction, magnitude, wavelength, polarization and phase. These changes

can be detected by the remote sensor and can be interpreted by scientists

to obtain useful information about the object of interest.

Incident energy = reflected energy + absorbed energy + transmitted

energy

EI () = ER () + EA () + ET () (1)

Figure (‎2-3) The energy interactions, (Liliesand & Kiefer, 1993)

17

Remote sensing are principally concerned with the reflected radiation

which emitted by the different objects in the earth. The magnitude of

reflected radiation depends on the type of surface reflecting it. This leads

to what is called spectral signatures.

The ratio of reflected energy with respect to the incident energy is known

as spectral reflectance, [ρ (λ)], it is a wavelength dependent. Each feature

in the earth has its own spectral reflectance characteristics. it is

responsible for the color of different objects, for example tree color is

green as they reflect more in the green wavelength.

The spectral reflectance is dependent on wavelength; it has different

values at different wavelengths for a particular terrain feature. The

reflectance characteristics of the earth’s surface features are expressed by

spectral reflectance, which is given by:

ρ (λ) = [ER(λ) / EI(λ)] x 100 (2)

Where,

ρ (λ) = Spectral reflectance at a specific wavelength.

ER (λ) = the reflected energy from object at a specific wavelength.

EI (λ) = the incident energy upon the object as a function of wavelength.

The spectral reflectance curve is the relation between ρ (λ) and λ. The

variation in the curve depends on the physical conditions and chemical

decomposition of each land feature. Wide range of values was observed

for the same material, so that, averaging is required to get a generalized

spectral response for the objects under study. Spectral response pattern

can be identified by spectral signature which indicates the characteristic

of each land feature. Figure ( 2-4) represents the spectral reflectance

curves for clear water, vegetation and bare soil.

18

Figure (‎2-4) the spectral reflectance curves for clear water,

vegetation and bare soil, (Tantirimudalige, 2002)

2.3. Optics of water

This section will explain the fundamentals of water optical properties,

including the interaction of light with natural waters in lakes, rivers,

estuaries, and how this process can used to acquire information about the

constituents in the water. Optics is that branch of Physics which studies

the characteristics of light (Visible Part of EMR). Since the behavior of

light is significantly affected by the nature and characteristics of the

medium through which it is passing, there are different branches of optics

dealing with different kinds of physical systems. Figure ( 2-5) presents the

different physical systems that interact with the EMR (Kirk, 1994).

Hydrological optics is the branch of optics that quantitatively studies the

interactions of radiant energy from the Earth’s oceans, estuaries, lakes,

reservoirs, and other water bodies (Mobley, 1995). Hydrological optics

can be subdivided into oceanographic and limnological optics depending

on the type of water under consideration whether it is salty, marine or

fresh, inland waters (Kirk, 1994).

19

Figure (‎2-5) The relation between hydrologic optics and other

branches of optics (after Preisendorfer, 1976 and Kirk, 1994)

2.3.1. Interaction between light and water

The interaction of the visible light in water depends on the characteristics

of the water and its optical properties. The water optical properties are

connected with the ambient light under the radiative transfer theory

(Mobley, 1995). According to (Mobley, 1995), Preisendorfer classified

the optical properties of water in two classes: inherent and apparent.

2.3.1.1. The inherent optical properties

The inherent optical properties depend on the medium and its

characteristics. There are two things that can happen to light passing

through water: it can be absorbed or scattered. Thus, if the effects

happened to light due to absorption and scattering were identified, the

characteristics of the water constituents can be known. Some

measurements were needed to know the quantity of light absorbed and

scatted by water. The ability of aquatic medium to scatter and absorb the

light at a given wavelength can be specified in terms of the scattering

20

coefficient, absorption coefficient, and the volume scattering function

(Kirk, 1994).

2.3.1.2. The apparent properties

The apparent properties depend upon both the medium and the ambient

light field. There are two critical apparent optical properties of water, the

vertical attenuation coefficient (Kd), also called spectral diffuse

attenuation coefficient, and the spectral irradiance reflectance (𝜆). The

beam attenuation coefficient, 𝑐 (𝜆), is different than the diffuse

attenuation coefficient, (Kd). 𝑐 (𝜆) refers to the radiant power lost from a

single, narrow beam of photons while the Kd refers to the decrease with

depth of the ambient downwelling irradiance 𝐸d (𝑧, 𝜆), which includes

photons heading in all downward directions (Mobley, 1995).

The most optically significant constituents in natural waters (lakes, ocean

and rivers) include, dissolved substances and particulate matter, including

phytoplankton (Mobley, 1995). These constituents will affect the

absorption and scattering processes of the light energy passes through

water.

The solar radiation that penetrates water is crucial to support life in

aquatic ecosystems. The response of water in the EMR differs by

wavelength and by the type of particles in it. Clear water reflects more in

short wavelengths and continuously reduces its reflection with increasing

wavelength. Therefore, clear water is commonly studied in the visible

spectrum, since it is there where water reflects the most. In the Near

Infrared (NIR) the reflectance of deep, clear water is nearly zero.

The challenges for satellite remote sensing are to estimate concentrations

of optical water quality parameters from reflectance and to differentiate

the reflectance generated by each parameter. Given the change of color

and consequently the change in reflectance of water, depending upon the

constituents it has, reflectance can in theory determine those constituents.

For example, in low concentration of chlorophyll, reflectance is highest

21

at blue wavelengths, this being the reason that clean ocean water has a

blue color. When there is greater concentration of chlorophyll, the

maximum R(λ) shifts from the blue to the green wavelengths; hence the

color of the water turns green. In high-sediment concentrations, R(λ) are

relatively flat from blue to yellow. Waters with high concentrations of

yellow matter will have high reflectance in the yellow part of the

spectrum.

2.3.2. Case 1 and Case 2 waters

(Morel & Prieur, 1977) defined water in ocean, coastal zones and inland

lakes into Case 1 and Case 2 waters, depending on its optical properties,

and refined by (Gordon & Morel, 1983). Case 1 waters are those

influenced by phytoplankton and their derivative products which play a

dominant role in defining the optical properties of the water. In Case 2

waters, the optical properties of water are influenced mainly by

suspended sediments transported by flood in river and reservoir, or from

particles and/or dissolved color. In Case 2 waters, phytoplankton

products may also be presented in significant amount or not.

2.4. Remote Sensing and water quality

Water quality parameters were studied using remote sensing earlier by

1970s. Many approaches and algorithms were developed using data

collected either by remote sensing satellites or by field optical

measurements. This section will discuss the approaches used, in addition

to a brief illustration about the past, present and future satellite missions.

2.4.1. Retrieval approaches

It is essential for retrieve water quality using remote sensing to have a

model between the different parameters concentrations and information

acquired using sensing satellites such as radiance or reflectance. Three

approaches were used in relating water quality to remote sensing,

empirical, semi-analytical and analytical approaches as stated by (Morel

& Gordon, 1980). The description of each approach is presented below:

22

1- Empirical approach

It depends on constructing statistical models between measured spectral

values and measured water quality parameters, these models can be

obtained using a linear or nonlinear regression. This approach is the most

widely used in estimating water quality parameters.

2- Semi-analytical approach

This approach typically uses an analytical model to derive the in-water

optical properties of backscattering and absorption from remote sensing

reflectance, but also require the use of empirical relationships between

these optical properties and seawater constituents. Semi-analytical

approaches are typically used in optically complex Case 2 waters such as

coastal and inland lakes and used to estimate chlorophyll-a and colored

dissolved organic matter.

3- Analytical approach

This approach is based on theoretical algorithms which uses apparent

optical properties and inherent optical properties to model the reflectance

and vice versa. There are a very few purely analytical algorithms because

they require detailed knowledge of a host of complex and often poorly

understood relationships between water components and their specific

optical properties (Morel, 1980; Morel & Maritorena, 2001).

2.4.2. Satellite Missions

Remote sensing was used in water quality mapping for first time in

1970’s with lunching with the launch of the Earth Resources Technology

Satellite (ERTS-1) by NASA to obtain a periodical multispectral images

over environmental targets and use these images to some investigations

that dealing with a variety of variables pertinent to Earth’s resources,

(Bukata, 2005).

23

Figure (‎2-6) Time line for Ocean color Sensors from 1978 to 2030

(CEOS, 2016)

In 1978, NASA had been successfully launched the first ocean color

sensor called Coastal Zone Color Scanner (CZCS). Although CZCS was

an experimental mission, it continued to provide data for selected sites

until 1986. A gap of ten years was found until lunching of Ocean Color

Temperature Scanner (OCTS) and POLarization and Directionality of the

Earth's Reflectances (POLDER) sensors in 1996 and Sea-viewing Wide

Field-of-view Sensor (SeaWiFS) in 1997. In the last years, some color

sensors were put onboard such as MERIS, MODIS Aqua and Terra, and

PROBA. Figure ( 2-6) illustrates past, current, and future water-color

missions. Data used in the figure were extracted form (CEOS, 2016).

Various types of remote sensing satellite rather than presented in Figure

( 2-6) also were used in monitoring water quality status in different water

bodies such as Landsat, IKONOS, ASTER, ALOS and SPOT.

GCOM-C3

FY-3G

FY-3RM

GCOM-C2

Meteor-M N3

Sentinel-3 C

FY-3F

FY-3E

Sentinel-3 B

FY-3D

GCOM-C

Sentinel-3 A

FY-3C

FY-3B

OCEANSAT-2

HJ-1A

FY-3A

Aqua

PROBA

MERIS

Terra

SeaWiFS

OCTS

CZCS

19

78

19

82

19

86

19

90

19

94

19

98

20

02

20

06

20

10

20

14

20

18

20

22

20

26

20

30

24

2.5. Application

The previous sections discuss the principles of remote sensing and the

interaction between the light and water which is effective in getting water

quality parameters via remote sensing techniques. Here in this section, a

various application of using remote sensing in retrieving water quality

parameters will be discussed:

A valuable literature review paper was prepared by (Chang et al., 2015)

to address the critical researches for monitoring surface water quality

status and ecosystem state in relation to the nutrient cycle in a 40-year

perspective. They do a great effort to collect the researches that have

been done in different water bodies to estimate water quality parameters

using empirical, semi-analytical and analytical approaches.

Regarding empirical approach, (Chang et al., 2015) reviewed selected

research which uses the empirical approach to retrieve water quality

using remote sensing platforms and sensors. The selected researches were

covering the period from 1998 to 2012, different empirical methods such

as regression, Artificial Neural Networks and Genetic Algorithms.

For semi-analytical approach, the papers selected in review of (Chang et

al., 2015) were covering the period from 1996 to 2012 to monitor various

types of water bodies using both air- and space-borne multispectral and

hyperspectral data.

Several selected studies that uses the analytical approach was also

reviewed in (Chang et al., 2015), using different techniques in finding the

relationship between inherent and optical properties such as Matrix

Inversion Method (MIM), Analytical Optical Modelling and Bio-optical

Model. The selected researches covered to period from 1998 to 2011.

In addition, they listed the researches that have been done to estimate

concentrations water quality parameters such as suspended sediments,

Chlorophyll-a, Total Phosphorus (TP), Total Nitrogen (TN), Temperature

Total Organic Carbon (TOC) and Microcystin.

25

2.5.1. Suspended Sediments

Suspended sediment is considered one of optical water quality parameters

so that most of the researches were tries to estimate its value using

different images and different modelling approaches. The review that

done by (Chang et al., 2015) cited some of these researches. Table ( 2-1)

gives a more information about these research such as which water body

it was conducted for, the used remote sensing platform, the inversion

method used to correlate between suspended sediment and remote

sensing reflectance, and the correlation coefficient.

Some study cases were added to Table ( 2-1) used to estimate suspended

sediment in two water bodies in Egypt. (Farag, 2011) constructed an

empirical Log-linear equation to estimate suspended sediment in Lake

Burullus, Egypt using MERIS image. (Azab, 2012) was studied

suspended sediment in Lake Edko, Egypt using MODIS Image.

Suspended sediment was correlated to remote sensing data acquired form

worldview-2 satellite in Rosetta branch, Egypt by (El Saadi et al. 2014).

Multiple band regression was used.

Table ( 2-1) indicates that the correlation coefficient is very good for most

of the reviewed researches except for small cases it gives a bad

correlation.

2.5.2. Turbidity and Transparency

Turbidity and transparency have been affected by the suspended sediment

concentration, so that it is noted both turbidity and transparency have a

good correlation with remote sensing data. (Allan et al., 2007) conducted

a study to evaluate different water quality parameters such as (TP, TN,

Chl-a, turbidity and transparency) in Rotorua Lakes using remote sensing

techniques. During this research, landsat TM was used in two different

seasons and found that Turbidity and Transparency have a very good

correlation with remote sensing data, where the correlation coefficient

between natural log of Turbidity and Transparency is -0.931 and 0.939

respectively.

26

Table (‎2-1) Selected remote sensing application for monitoring

Suspended Sediment in different water bodies (Chang et al., 2015)

Location Reference

Remote

sensing

Platforms

Inversion

method R

2

Six northern

Mississippi

reservoirs, U.S.

Ritchie et al., 1976 Landsat Linear

regression 0.88

Lake Chicot,

Arkansas, U.S.

Ritchie & Schiebe,

1986 Landsat MSS

Linear

Regression 0.78

Multiple

Regression 0.83

Delaware Bay-

U.S. Keiner & Yan, 1998 Landsat TM

Log-linear

equation 0.54

Neural

Network 0.97

North Sea,

Europe Gomez, 2000 AVHRR

Linear

Regression

Summer

(0.94)

Winter

(0.024)

The Great Miami

River, Ohio Senay et al., 2001 CASI

Linear

Regression 0.79

Lake Balaton,

Central Europe Tyler et al.,2006 Landsat TM

Linear

Regression 0.89

Istanbul, Turkey Ekercin,2007 IKONOS Linear

Regression 0.83-0.97

The Peace-

Athabasca Delta,

Canada

Pavelsky & Smith,

2009 SPOT, ASTER

Empirical

Formula 0.78-0.82

Mayaguez Bay,

Puerto Rico

Rodríguez Guzmán

& Gilbes Santaella,

2009

MODIS Exponential

Equation 0.73

Lake Mazalah,

Egypt Farag, 2011 MERIS

Log-linear

equation 0.56

Lake Edko

Egypt Azab, 2012 MODIS

Exponential

Regression 0.85

Rosetta Branch,

Nile River, Egypt El Saadia et al. 2014 WorldView-2

Multiple

Regression 0.388

2.5.3. Chlorophyll-a

Chlorophyll-a is also optical water quality parameters, so that, it indicate

a very good correlation with remote sensing data as indicated in Table

( 2-2) cited by (Chang et al., 2015) for some of selected studies.

27


Chlorophyll-a in different water bodies (Chang et al., 2015)

Location Reference Remote Sensing

Platforms

Inversion

Method R

2

A group of lakes in

Finland

Kallio et al.,

2001

Airborne Imaging

Spectrometer for

different Applications

(AISA)

Two-band Ratio

Equation 0.72-0.89

Black Sea Kopelevich et

al., 2002

Coastal Zone Color

Scanner (CZCS) Regression 0.81

A group of

Minnesota lakes,

U.S.

Brezonik et

al., 2005 Landsat TM

Linear

Regression 0.88

West Florida Shelf

and Grand Bahamas

Bank

Cannizzaro

and Carder,

2006

Lambertian “graycard”

reflector

Semi- analytical

Model _

Istanbul, Turkey Ekercin, 2007 IKONOS Linear


West lake, China Torbick et al.,

2008 Landsat 7 ETM

Linear

Regression 0.82

Lake Malawi Chavula et

al., 2009 MODIS

Two-band and

three- band

Models

0.58

Burdekin Falls

reservoir, Australia

Campbell et

al., 2011 MERIS

Analytical

model 0.92

Group for

Mediterrane an

Lakes

Gomez et al.,

2011 CHRIS and MERIS

Single Band

Regression 0.99

Group of Lakes in

Eastern China

Duan et al.,

2010 MERIS

Two-band


Fremont Lakes,

Nebraska, US

Moses et al.,

2012

Airborne Imaging

Spectrometer for

different Applications

(AISA)

Two-band or

Three-band

Models

0.83-0.98

Lake Okeechobee,

Florida, U.S.

Chang et al.,

2012a MODIS

Genetic

Programming 0.72

Rosetta Branch,

Nile River, Egypt

El Saadia et

al. 2014 WorldView-2

Multiple

Regression 0.388

2.5.4. Total Phosphorus (TP) and Total Nitrogen (TN)

Total phosphorus (TP) and Total Nitrogen (TN) were known as nutrients,

the presence of nutrients have a direct effect in increasing the growth of

28

algae and in turn may increase the eutrophication of water body, so that

many researches were conducted during years to estimate values of both

parameters using remote sensing techniques Table ( 2-3) highlights some

of researches done to estimate both TP and TN as cited by (Chang et al.,

2015). It is noted that the correlation between both TP and TN with

remote sensing data is very good in some cases and in other cases is poor.


Nutrients (TN and TP) in different water bodies (Chang et al., 2015)

Location Reference Remote Sensing

Platforms

Inversion

Method R2

Manzala

Lagoon, Egypt

Dewidar &

Khedar, 2001 Landsat 5 TM

Multiple

Linear

Regression

Model

TN (0.298)

TP (0.421)

Guanting

Reservoir,

Beijing, China

He et al.,

2008 Landsat 5 TM Regression

TN (0.75)

TP (0.61)

Lake Kemp,

Texas, U.S.

El-Masri &

Rahman,

2008

MODIS Linear

Regression TP (0.56)

Kucukcek

Lake, Turkey

Alparslan et

al., 2009

Landsat-5 TM +

SPOT-Pan, IRS-

1C/D LISS+ Pan,

and Landsat-5 TM

Multiple

Regression

TN (0.835)

TP (0.788)

Koise River,

Japan

He et al.,

2010 QuickBird

Linear

Regression TN(0.94)

Quintang

River, China

Wu et al.,

2010 Landsat TM Regression TP (0.77)

Baltic Sea,

Eastern part of

Sweden

Anderson,

2012 Landsat TM Regression TP (0.41)

Taihu Lake,

China

Chen &

Quan, 2012

Landsat TM (Band

1,2,3,4)

Linear

Regression

TN (0.24)

TP (0.63)

Tampa Bay,

USA

Chang et al.,

2012b MODIS

Genetic

Programming TN (0.75)

Tampa Bay,

USA

Chang et al.,

2013a MODIS

Genetic

Programming TP (0.58)

Rosetta

Branch, Nile

River, Egypt

El Saadia et

al. 2014 WorldView-2

Multiple

Regression TP (0.58)

29

2.6. Remote sensing studies in Lake Nasser

This part of literature will focus on the researches that use remote sensing

to describe the behavior of Lake Nasser in many driplines. The first study

of Lake Nasser using remote sensing have been done by (Aguib et al.,

1989) to monitor sedimentation in AHDR, they proposed a technique for

sediment process calculation using a hydrodynamic model and remote

sensing data. Since this date, many researches were conducted to study

the Lake in many fields.

Evaporation losses were studied by (Shafik, 2004) using two sets of

Landsat images to represent the whole lake surface in two different dates,

the images were processed to calculate the lake surface area at different

water levels. Evaporation losses were calculated using surface area and

other meteorological data.

(Ebaid & Ismail, 2010) conducted a study to evaluate the reduction of

evaporation in Lake Nasser by disconnecting the secondary Channels

(Khores). The study estimates the evaporation using the integration

between remote sensing and GIS techniques. The aerodynamic method

was used in order to estimate the evaporation volume from surface of

lake and evaporation from secondary channels. The results show that 2.4

billion m3/year can be reduced from the total evaporation losses by

disconnecting the secondary channels.

(Hassan, 2013) estimate the daily evaporation rate in Lake Nasser using

Surface Energy Balance Algorithm for Land. The algorithm was applied

to different seven Landsat TM for lake, the results were compared with

six traditional methods used for estimating the evaporation rate. A good

performance of the mass transfer method was noted.

(Mostafa & Sousa, 2006) used Geographic Information System (GIS) and

remote sensing techniques to propose a suitability map of the area around

lake based on different environmental aspects, three Landsat images were

used in different date (1984, 1996 and 2001) to track the changes in lake

30

shape in different years and the water level fluctuation which helps in

identifying the agriculture areas around the lake.

(Hassan & Fahmy, 2005) estimated surface suspended sediment using the

second order polynomial equation developed by (Ritchie et al., 1991).

The model was applied to two Landsat images represent the high and low

water levels. The author stated that the proposed equation at low water

levels suspended sediment in surface layer were ranging between 2 to 5

mg/l, while it isn’t representing the suspended sediment value during

flood.

2.7. Discussion

The main objectives of the literature review chapter is to explain what is

meant by remote sensing and its history especially in field of estimating

surface water quality, in addition to the used techniques in evaluating

different water quality parameters. Also the chapter contains the

researches that had been done in Lake Nasser in different fields which

guide us that there is only one published research in the field of water

quality, this research used an equation developed by (Ritchie et al., 1991)

to estimate TSS in Lake Nasser while the current research will introduce

a full study to retrieve water quality parameters via remote sensing

techniques. The current study will introduce a different regression models

that can be used especially for Lake Nasser in different seasons.

31

CHAPTER (3)

STUDY AREA AND DATA COLLECTION

3.1. General

The current chapter will illustrate the study area characteristics and a full

description of data collected to achieve the research objectives.

3.2. Study Area

The Aswan High Dam (AHD) has its great effect on the economic

development of Egypt by means of storing water during flood in the

upstream reservoir, known as Aswan High Dam Reservoir (AHDR), to

supply water needed for irrigation, producing hydroelectric power and

protecting the downstream reaches from flood damage. Protection of

water in AHDR from pollution is a big challenge, so that, Egypt

government take some important decisions to protect it against pollution,

such as, isolating it from any human activities. Decree 203/2002

identified a 2 km buffer zone around, where it is not allowed to have any

agricultural, industrial and tourist activities. Wadi Alaqi, as a part of Lake

Nasser was considered as a Biosphere Reserve of international

importance according to Law 102/1983 which stated that this area should

remain free of any development, changes and disturbance in land-use or

activity that may affect the natural site (Zaghloul et al., 2012). This

research will focus on study of the spatial and temporal concentration of

different water quality parameters such as suspended sediment,

chlorophyll, total phosphorus, etc. A complete description of study area

characteristics is carried out within this chapter.

3.2.1. Aswan High Dam (AHD)

The AHD was constructed in the period between 1964 and 1971, created

a large reservoir, upstream the dam, the reservoir is extended in Egypt

and Sudan, the Egyptian part is known as Lake Nasser while the

Sudanese part is known as Lake Nubia. This research will focus on the

32

Egyptian part of reservoir (Lake Nasser) as a study area. Lake Nasser is

considered the second largest man-made lake all over the world (ICOLD,

1984). Thus, it has to be set under comprehensive studies particularly the

water quality parameters to make sure of the quality and clarity of the

stored water using a scientific approach to help in monitoring, protecting

and managing water resources. Satellite remote sensing techniques can be

used for assessment and monitoring of lake conditions where it can

supply timely and reliable information to decision-maker to help in

putting sustainable management and development plans.

AHD was located 7 Km upstream Aswan Old Dam (AOD). The

construction of AHD took place in two steps. The first step was the

diverting the river into a diversion channel in the eastern bank of the Nile

and the construction of cofferdams along the dam path to close off that

part of its course where the main body of the dam was to be built (Said,

1993). The second step was the building of the main body of the dam.

This step continued till 1968 and the hydroelectric power plant from 1967

to 1971 (Shahin, 2002). The dam is a rock-fill dam made of granite rocks

and sands and provided with a vertical cutoff wall consisting of very

impermeable clay. The dam has a trapezoidal cross section, its crest

width is 40 m and its base width is 980 m. A layout of the dam can be

seen in Figure ( 3-1).

33

Figure (‎3-1) Layout of Aswan High Dam

3.2.2. Aswan High Dam Reservoir (AHDR)

After the construction of the upstream cofferdam in 1964, the reservoir

slowly started to fill. The AHDR is designed to store water at a maximum

water level of 182 m above (MSL). The mean depth of the reservoir is 25

m while the maximum depth reaches up to 110 m at its maximum level.

The total capacity of AHDR is 162 km3 at its maximum water with a total

length of 500 km where 333 km in Egypt and 167 km Sudan. The

average width of AHDR is 13 km, and its surface area is about 6500 km2

at its maximum water level.

3.2.3. AHDR morphology

In general, the morphology of a lake is described by bathymetric map

which is required to determine the major morphological parameters. This

map is prepared by a survey of the shoreline by standard surveying

methods combined with aerial photography or remote sensing images.

34

Lake Nasser has a number of side channels called in Arabic Khors

extending in both the Egyptian and Sudanese parts as shown in Figure ( 3-

2). Hence, the reservoir is considered dendritic in shape characteristics.

The total number of important Khors is about 100 in both Lake Nasser

and Nubia Lake. Khors Allaqi, Kalabsha and Toshka are the largest in

area.

Figure (‎3-2) Lake Nasser Map

35

3.2.4. AHDR operation policy

The reservoir total capacity is divided into three zones, according to its

operation policy. The dead storage zone of 31 km3, as shown in Figure

( 3-3), which is used for sediment depositions which comes with the river

flow during the flood, the dead zone is identified as the volume of the

reservoir allocated below level of 147 m above (MSL). The second zone

is the live storage zone, it extended between levels of 147 m and 175 m

above (MSL) divided into a buffer zone lies between elevation 147 and

150 m and a conservation zone between 150 to 175 m, the total capacity

of live zone is about 90 km3. The third zone is used as for flood buffer

and control; its volume is about 41 km3 and it is allocated between levels

of 175 m and 182 m above (MSL), which is considered the maximum

water level of the reservoir. The crest of emergency spillway is found at

level of 178 m which is considered as the separation level between the

flood buffer zone and flood control zone (Fahmy, 2001; Sadek et al.,

1997). The policy of AHDR is to maintain the level of the reservoir at

175 m at the end of the hydrologic year (July 31st).

Figure (‎3-3) Main component for different operation zones of AHDR

3.2.5. AHDR Monitoring Activities

Since 1973, the first mission was conducted by High Aswan Dam

Authority (HADA) incorporation with Nile Research Institute (NRI), to

Inflow

(178 m)

(182 m)

(175 m)

(150 m)

(147 m)

Dead Storage (31 BCM)

Conservation Zone

Flood Buffer zone

Flood control zone

Buffer Zone

Live Storage

(90 BCM)

Flood Storage

(41 BCM)

36

collect data about the sediment deposition in Lake Nasser. The main

goals of this mission were designed to conduct hydrographic survey,

measure flow currents, and collect suspended matters and bed material

samples at specific monitoring cross sections along the lake in both Egypt

and Sudan parts.

Total number of 29 cross sections was hydrographically surveyed in

order to estimate sedimentation in AHDR. Table ( 3-1) indicates names

and location from upstream AHD gauge. In earlier time of monitoring

program, the survey equipment used in the survey was theodolites and

primary echo-sounders. So that, only the indicated cross sections was

used in monitoring the sedimentation. By the time, the survey equipment

and techniques were developed. A real time data acquisition system, such

as Mini-Ranger systems (Falcon IV), was used since 1990. Differential

global positioning system (DGPS) has been used in survey since 1998.

This development in survey equipment enables the survey team to

increase monitoring cross section and also increase the accuracy of

estimating the sedimentation volume.

At the beginning of the monitoring program only hydrographic survey

was conducted in addition to measuring velocity, suspended matters and

bed material samples which usually collected at different vertical

locations depending on the total depth at sampling point. Samples were

collected at east, middle and west of each cross section. At the end of

1970s, water quality was taken into account in the monitoring program, a

fewer parameters, in particular physiochemical, were measured at the

beginning and by the time the measured parameters were increased to

include more water quality variables such as nutrients and major ions. (El

Sammany, 2002) gives a detailed description of all the actual variables

measured at each cross section as well as the number of times a particular

variable was measured at each cross section.

37

Table (‎3-1) Definition of main monitoring locations surveyed (El

Sammany, 2002)

Region Serial

No. Cross Section Name

Cross

Section

Number

Location U.S.

AHD (km)

Su

da

nes

e B

ord

ers

1 EL-Daka 23 487.000

2 Okma 19 466.000

3 Malek EL-Nasser 16 448.000

4 EL-Dowaishat 13 431.000

5 Ateere 10 415.500

6 Semna 8 403.500

7 Kajnarity 6 394.000

8 Morshed 3 378.500

9 Gomai D 372.000

10 Madeek Amka 28 368.000

11 Amka 27 364.000

12 Second Cataract 26 357.000

13 Abdel-Qader 25 352.000

14 Doghame 24 347.000

15 Dabrosa 22 337.500

Eg

yp

tia

n B

ord

ers

16 Arkin 21 331.100

17 Sara 20 325.000

18 Adendan N/A 307.000

19 Abu-Simbel N/A 282.000

20 Toushka N/A 256.000

21 Ibreem N/A 228.000

22 Korosko N/A 182.500

23 Wadi- Alarab N/A 171.000

24 EL-Madeek N/A 130.000

25 Garf Hussien N/A 90.000

26 Morwaw N/A 60.000

27 Kalabsha N/A 41.000

28 Dahmeet N/A 30.000

29 High Dam N/A 0.500

38

The AHDR monitoring program is basically proposed and designed for

sediment deposition in the reservoir. Since sediment deposition is

influenced by flood conditions, estimating of sediment deposition

frequency has been chosen to be once per year for both sediment

deposition and water quality variables. In addition, the timing for

sampling within the day of sampling has not been defined according to a

certain criteria and/or understanding of timing effects on water quality

sampling. For example, thermal stratification pattern is changing from

time to another within the same day, as it is affected by incoming solar

radiation. In addition, photosynthesis process, carried out by algae and

aquatic plants, is largely affected by the timing of the day since the

available light penetration is changing from time to another within the

day of sampling.

3.2.6. AHDR hydrological characteristics

AHDR was designed essentially for purposes of flood protection, water

conservation and hydropower generation. The released water from dam is

designed to meet the irrigation requirements, which varies from season to

another accordingly with the crop pattern. In addition, the reservoir is

operated to control the annual flood by drawing down the reservoir to a

predetermined level on a specified date each year. Additional constraints

on the operation of the reservoir are imposed by the need to limit the

magnitude of releases so as to avoid downstream degradation and/or

hindrance to navigation.

3.2.6.1. Water levels

There is only one gauging station measuring daily water level all over the

year and it is located just upstream the AHD. But, during the survey

missions, the water levels at the sampling locations are measured and

estimated from the vertical elevation of the nearest landmark to each

sampling location. Figure ( 3-4) illustrates the daily water levels U.S.

AHD from 1964 to 2011. Water level data were extracted from Nile

research Institute (NRI) database, the figure shows that reservoir started

39

to fill in 1964 until it reached its full storage capacity at level of 175 m

according to the indicated by the operation policy of the reservoir. Period

from 1975 to 1981, the maximum and minimum water levels in the

reservoir is extremely fixed as the level ranging from 171.13 m to 173.03

for minimum and from 175.71 m to 177.47 m. reservoir reach its

minimum capacity in 1988 as the water level reached to value of 150.62

m, just above the buffer zone limit. A rapid increase in reservoir water

level occurred in the same year due to high flood resulted in increasing

the water level to 168.82 m. the water level increased gradually until it

reaches its maximum value 181.60 m. The rising stage starts by the end

of July and reaches its peak around the middle of September and the

falling stage, of which the discharge starts to decrease during months

October to June.

3.2.6.2. Discharges

The average monthly inflow discharges at Dongola gauging station is

shown in Figure ( 3-5). Dongola is the first main gauge station located on

the main Nile below the confluence of Atbara river, it is located 750 km

U.S. AHD, Dongola play a key role in calculating water balance of

AHDR. Figure represents the inflow from 1964/1965 to 2009/2010, the

data used in the figure was extracted form (Abdel-Latif and Yacoub,

2011). The relatively high flood occurred in 1964/1965, 1974/1975 and

1988/1999.

40

Figure (‎3-4) Water level Upstream AHD from 1964 to 2011

Figure (‎3-5) AHDR inflow at Dongola station

This research will focus in the period from 2002 to 2011. In this period,

the reservoir reached its lowest level of 168.57 m in July 2006, but

fortunately, the large flood of during the successive years in 2006/2007

41

and 2007/2008 allowed the reservoir to rise again and reach up to a

maximum storage level of 180.11 m in October 2007. The minimum

flood was occurred in 2003/2004 and the maximum occurred in

2006/2007.

3.2.6.3. Sediment Data

After the construction of AHD, Dongola Station has been selected as a

measuring station to monitor suspended sediment. The rising and falling

rating curves for AHDR were constructed using sediment data available

for the period 1966-1982 (EL-Moattassem and Abdel-Aziz, 1988).

Figure (‎3-6) Discharge and sediment concentration hydrographs at

Dongola station (1999-2003)

The equations for estimating the suspended sediment hydrograph at

Dongola Station are as follows:

(i) For rising stage flow discharge hydrograph:

𝑄𝑠 = 5.753 × 10−6 𝑄1.98 (2)

(ii) For falling stage flow discharge hydrograph:

𝑄𝑠 = 2.695 × 10−7 𝑄2.347 (3)

Where Q is the discharge at Dongola Station in million m3/day and Qs is

the sediment load in 109 kg/day. By applying these equations, the

Discharge and Sediment Concentration at Dongola Station

( 1999 - 2003)

0

200

400

600

800

1000

1200

1400

1600

1800

2000

AU

G.

SE

P.

OC

T.

NO

V.

DE

C.

JA

N

FE

B.

MA

R.

AP

R.

MA

YJU

N.

JU

L.

AU

G.

SE

P.

OC

T.

NO

V.

DE

C.

JA

N

FE

B.

MA

R.

AP

R.

MA

YJU

N.

JU

L.

AU

G.

SE

P.

OC

T.

NO

V.

DE

C.

JA

NF

EB

.

MA

R.

AP

R.

MA

Y

JU

N.

JU

L.

AU

G.

SE

P.

OC

T.

NO

V.

DE

C.

JA

NF

EB

.

MA

R.

AP

R.

MA

Y

JU

N.

JU

L.

1999/2000 2000/2001 2001/2002 2002/2003

Months

Dis

ch

arg

e (

M.m

^3/d

ay)

0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

Sed

imen

t C

on

cen

trati

on

pp

m

Flow Discharge

Sediment Concentration

42

suspended sediment concentration hydrographs at the inlet boundary of

the study area of the reservoir (Dongola Station) can be estimated for the

period from 1999 to 2003 as shown in Figure ( 3-6).

From Figure ( 3-6), it is clear that the concentration of suspend sediment

entering Aswan High Dam Reservoir also has a seasonal variation like

the flow hydrograph, the peak discharge and peak suspended sediment

concentration do not occur simultaneously. The suspended sediment

concentration rises to a maximum (5000 ppm) many days before the peak

of water discharge. The lag time between the peak of the water discharge

and the suspended sediment concentration varies from year to year, and

on average is approximately 10 days.

3.3. Data Collection

3.3.1. Field data

Water quality data used through this research was collected during annual

field survey for Lake Nasser carried out by High Aswan Dam Authority

(HADA) incorporation with Nile Research Institute (NRI), the main goal

of this annual trip to perform a hydrographic survey, water quality

measurements, and study of growing weeds in lake and around the lake

shore line. (El Sammany, 2002) gives a full description and main

drawbacks of the methodology used in these field trips.

Figure ( 3-7a) shows samples’ location for water quality monitoring, and

the location of each sample upstream AHD was shown as well, in

addition to its geodetic coordinates. The indicated location is for all

survey missions except for 2010 mission as samples were taken every

five kilometers from the Egyptian-Sudanese border until reaches near

Allaqi section as indicated in Figure ( 3-7b). Five vertical samples were

taken at 0.50 m below water surface, 25%, 50%, 65% and 80% of water

depth at each location. Water sample at depth 0.5 m under water surface

is used through this research as it is the effective layer in remote sensing.

43

Water quality data used through this research was collected during six

survey missions in the period from 2003 to 2011, as shown in Table

( 3-2). Seven water quality parameters are measured during the indicated

field mission. The parameters are Total suspended sediment (TSS),

Chlorophyll-a (Chl-a), Turbidity, Transparency, Total phosphorus (TP),

Silica (SiO2), Total dissolved Solids (TDS). Figure ( 3-8), shows the

period of each survey with changing water levels in Lake Nasser during

the period 2001- 2011. The first three survey missions and the sixth

mission were conducted in the period in which lake is in the case of

maximum level after the end of flood. The fourth mission was conducted

during the Lake level falling period, while the fifth mission was in the

rising period. Table ( 3-2) shows dates and the measured water quality

during each mission. As discussed before, not all water quality

parameters had been measured every year, a description for different

water quality parameters measured and the lake condition in each mission

will be illustrated in the following sections in details by means of water

level and measure water quality parameters during every survey mission.

Table (‎3-2) Measured Parameters for each mission

Dates TSS Chl-a Turbidity Transp. TP SiO2 TDS

First Mission 23Oct 03

15 Nov 03 ●

● ● ● ● ●

Second mission 27 Nov06

15 Dec 06 ● ● ● ● ●

Third Mission 15 Nov 07

24 Nov 07 ● ●

● ● ● ●

Fourth Mission 24 Apr 09

08 May 09 ● ●

● ● ●

Fifth mission 21 Aug 10

31 Aug 10 ●

● ● ●

●

Sixth Mission 04 Oct 11

10 Oct 11 ●

44

a)

b)

Figure (‎3-7) Water‎Quality‎samples’‎Locations

45

Figure (‎3-8) Survey mission with respect to water level change in

period from 2002 to 2011

First mission was conducted from October 23, 2003 to November 15,

2003. Table ( 3-2) indicates that six parameters were measured during this

mission. Water level in this period ranging from 177.28 m to 177.91 m

above MSL, this period water level in lake reaches its maximum value.

Table ( 3-2) indicates that TSS, Turbidity, Transparency, TP, SiO2 and

Total Dissolved Solids (TDS) were measured during first mission.

Parameters measured during second mission were only five parameters

(Turbidity, Transparency, TP, SiO2 and TDS). Mission duration was

between November 27, 2006 and December 12, 2006. The recorded

Water level at upstream AHD gauge had a constant value of 176.46 m

above MSL for all survey periods. This period was lied in the end of

flood period as the water level begun to decrease and suspended sediment

started to settle in the bottom of lake and by time water became clearer,

and hence the water quality changed accordingly. During 2006, Lake

reached to its minimum water level value of 168.57 m, which represented

the lowest water level occurred in recent years.

23

-Oct

-03

15

-Nov-0

3

27

-Nov-0

6

12

-Dec

-06

15

-Nov-0

7

24

-Nov-0

7

24

-Ap

r-09

08

-May

-09

21

-Au

g-1

0

31

-Au

g-1

0

04

-Oct

-11

10

-Oct

-11

Fir

st M

issi

on

Sec

on

d M

issi

on

Th

ird

Mis

sion

Fou

rth

Mis

sion

Fif

th M

issi

on

Six

th M

issi

on

46

The period of measurement for third mission was lied between November

15, 2007 to November 24, 2007. The flood during 2007 had its maximum

value in recent years as it reached to 88.78 BCM which resulted in

increasing water level upstream AHD to maximum value of 180.11 m

occurred in October 19, 2007. Water level through the period of

measurements was ranging from 179.85 m to 179.71 m at the start and

end of mission respectively, which means that it lies in the start of fall

period. Measured water quality parameters were (TSS, Chl-a,

Transparency, TP, SiO2, and TDS)

Fourth mission was surveyed in the period from April 24, 2009 to May

08, 2009. Water level was ranging from 175.62 m to175.1 m above MSL,

that period was lied, as previously illustrated in Figure ( 3-8), in the

middle of falling period. Table ( 3-2) shows the ranges of TSS, Chl-a,

Secchi depth, TP and SiO2 measured during that survey.

Parameters measured during fifth mission were only five parameters

(TSS, Turbidity, Transparency, TP, and TDS), and the measurements

were taken in period from August 21, 2010 to August 31, 2010. Water

level at upstream AHD gauge increased from 171.70 m in the beginning

of measuring period to 171.90 m above MSL. that period represented the

start of rising stage which occurred due to flood, water level started to

increase until it reaches its maximum value at the end of rising stage and

water is carrying large amount of sediments during this period.

Sixth mission was conducted during period from October 4, 2011 to

October 10, 2011. Only TSS was measured during that mission. Water

level values were ranging from 173.93 m to 173.95 m above MSL.

Table ( 3-3) presents a descriptive statistics for all missions, table shows

the maximum and minimum values for each parameter in addition to its

standard deviation and mean. Also, it shows the total number of samples

taken during each field mission.

The maximum TSS value occurred during the fifth mission which took

place during the start of flood period where water was fully loaded by

47

sediments. The minimum value of TSS occurred in the fifth mission as

well, see Figure ( 3-9), the reason of this is that the mission was just

before the flood and all sediments settled down. The maximum value of

standard deviation occurred in the fifth mission as there were big spatial

variations in concentration of sediment between start and end of the lake.

The maximum TSS concentrations for other mission was ranging from 14

to 24 mg/l, which was very small compared to the maximum value during

flood. The minimum values ranged from 2 to 6 mg/l, the variation in the

minimum concentration of TSS was not big as the northern part of the

lake is not influenced by flood sedimentation. The measured

concentrations of Chl-a during third and fourth missions were ranging

from 10 to 13 mg/m3 for maximum and 3.5 to 4.5 mg/m

3 for minimum,

Figure ( 3-10) indicates that there is a small variation of Chl-a

concentration during these two missions.

It is clear that transparency is affected by both turbidity and TSS, as

transparency is a measure of water clarity. Figure ( 3-11) indicted that the

maximum and minimum values were altered along the lake compared to

TSS and Turbidity. It is obvious from Table ( 3-3) that values of

maximum and minimum transparency are 1.30 m and 0.17 m during start

of flood, fifth mission. While it ranges between 1.64 m to 4.70 m during

other missions.

Figure (‎3-9) Longitudinal Profile for TSS during different missions

0

20

40

60

80

100

0 50 100 150 200 250 300 350

TS

S (

mg

/l)

Distance from AHD (Km)

First Mission

Third Mission

Fourth Mission

Fifth Mission

Sixth Mission

48

Figure (‎3-10) Longitudinal Profile for Chl-a during different

missions

Turbidity had high values during fifth mission, where maximum and

minimum values were 78.40 NTU and 1.35 NTU respectively. While

during other two missions; first and second, it approximately had the

same longitudinal profile as they were conducted during the same lake

condition, see Figure ( 3-12).

Figure (‎3-11) Longitudinal Profile for Transparency during different

missions

0

2

4

6

8

10

12

14

0 50 100 150 200 250 300 350

Ch

l-s

(mg/m

3)


Third Mission

Fourth Mission

0

1

2

3

4

5

0 50 100 150 200 250 300 350

Tra

nsp

are

ncy

(m

)


First Mission

Second Mission

Third Mission

Fourth Mission

Fifth Mission

49

Figure (‎3-12) Longitudinal Profile for Turbidity during different

missions

Total phosphorus (TP) ranging from minimum value of 0.07 mg/l to 0.15

mg/l during the start of flood season, fifth mission, while it had the

lowest concentrations compared to other missions. Maximum TP

concentrations were observed during the second mission, after end of

flood, where it ranged from 0.11 mg/l to 0.52 mg/l. it is observed that

profile of TP during end of flood season took the same trend as TSS

while it was approximately flat during other seasons, Figure ( 3-13).

Figure ( 3-14) presents the longitudinal profile for Silica (SiO2). During

first three missions (end of flood season) SiO2 ranges from 6.10 mg/l to

18.10 mg/l while during falling period it had low values compared with

the others. SiO2 ranges from 4.20 mg/l to 7.80 mg/l as shown in Table

( 3-3)

0

20

40

60

80

100

0 50 100 150 200 250 300 350

Tu

rbid

ity (

NT

U)


First Mission

Second Mission

Fifth Mission

50

Figure (‎3-13) Longitudinal Profile for TP during different missions

Figure (‎3-14) Longitudinal Profile for SiO2 during different missions

0

0.1

0.2

0.3

0.4

0.5

0.6

0 50 100 150 200 250 300 350

TP

(m

g/l

)


First Mission

Second Mission

Third Mission

Fourth Mission

Fifth Mission

0

10

20

30

40

0 50 100 150 200 250 300 350

SiO

2 (

mg

/l)


First Mission Second Mission Third Mission Fourth Mission

51

Figure (‎3-15) Longitudinal Profile for TDS during different missions

According to Table ( 3-3), TDS ranged from 131.0 mg/l observed during

first mission to 170.3 mg/l during the end of flood season, TDS

longitudinal profiles, Figure, indicates that the concentration during

flood, TDS had higher values in the southern part of the lake compared to

the southern part. That case was altered while it starts to decrease in the

southern part until the end of flood season.

120

130

140

150

160

170

180

0 50 100 150 200 250 300 350

TD

S (

mg/l

)


First Mission Second Mission Third Mission Fifth Mission

52

Table (‎3-3) Descriptive statistics of measured water quality

parameters

First

Mission

Second

Mission

Third

Mission

Fourth

Mission

Fifth

Mission

Sixth

Mission

n 10.00 12.00 11.00 11.00 16.00 9.00

TSS (mg/l)

Max 20.00 - 28.00 14.00 84.00 24.00

Min 6.00 - 3.00 5.00 2.00 2.00

Mean 10.20 - 12.82 7.45 23.06 8.67

Std Dev. 5.20 - 8.87 2.70 25.99 8.40

Chl-a (mg/m3)

Min - - 13.00 10.00 - -

Max - - 3.50 4.50 - -

Mean - - 7.50 8.05 - -

Std Dev. - - 3.07 2.16 - -

Transparency

(m)

Min 0.80 0.40 0.45 1.25 0.17 -

Max 4.70 3.50 4.00 2.00 3.10 -

Mean 2.80 2.08 2.05 1.64 1.30 -

Std Dev. 1.43 1.22 1.41 0.22 0.97 -

Turbidity

(NTU)

Min 0.75 1.06 - - 1.35 -

Max 16.50 37.30 - - 78.40 -

Mean 5.13 10.77 - - 21.27 -

Std Dev. 6.07 11.67 - - 24.54 -

TP

(mg/l)

Min 0.08 0.11 0.12 0.09 0.07 -

Max 0.19 0.52 0.40 0.15 0.15 -

Mean 0.11 0.25 0.25 0.12 0.09 -

Std Dev. 0.04 0.14 0.11 0.02 0.02 -

SiO2

(mg/l)

Min 6.10 7.90 12.80 4.20 - -

Max 14.20 18.10 16.40 7.80 - -

Mean 9.97 13.93 14.67 5.74 - -

Std Dev. 2.61 3.34 1.17 1.04 - -

TDS

(mg/l)

Min 138.0 142.35 131.00 - 140.16 -

Max 153.0 170.30 164.00 - 160.00 -

Mean 148.4 154.48 143.73 - 151.08 -

Std Dev. 5.13 9.40 12.92 - 6.99 -

53

3.3.2. Remote sensing data

The Medium Resolution Imaging Spectrometer (MERIS) will be used to

fulfill the objectives of the research. MERIS instrument is a multispectral

sensor that produces images with 15 spectral bands acquired in the visible

and near infra-red wavelengths. Each band has a programmable width

and position within the 390 nm to 1040 nm spectral range of the

electromagnetic spectrum (ESA, 2006). The primary mission of MERIS

is to measure the color of water in the oceans, seas, coastal zones and

inland lakes. Knowledge of water color can be interpreted into a

measurement of water quality parameters such as concentration of

suspended sediment and chlorophyll pigment, and atmospheric aerosol

loads over water (ESA, 2006). MERIS was provided in three main spatial

resolutions; Full-Resolution (FR), Reduced Resolution (RR) and Low

Resolution (LR). Each pixel in an FR image represents a ground area of

260 m × 290 m, while it represents an area of 1,040 m × 1,160 m in an

RR image, and in LR the pixel represents an area of 4,160 m × 4,640 m.

More details about MERIS images will be found in Appendix (A).

3.3.2.1. Data Sources

Remote sensing images used in the current research are available from

European Space Agency (ESA) through TIGER Initiative project no. 14

under title of “Enhancing the Sediment Transport Modelling in the Nile

Basin Reservoirs” and it was downloaded from different sources as

follow:

1- EOLI-SA is a free interactive tool that allows users to access and

download many remote sensing images from different platforms

especially ESA’s Earth Observation data products.

2- The ESA satellite-based Earth Observation Data Dissemination

System (DDS) installed in Nile Research Institute (NRI) which

provided for the dissemination of Envisat data products especially

MERIS Images.

54

3- MERIS FRS regional extraction tool (merisfrs-merci-

ds.eo.esa.int).

3.3.2.2. Used MERIS images

MERIS images used through this research is divided into two groups: the

first group is match-up images and the other one is data used for creating

time series for water quality parameters in Lake Nasser, each group can

be described as follows:

- Match-up images

MERIS full resolution (FR) level 1B scenes were used in this research.

Table ( 3-4) shows the match-up between MERIS images and field trips,

one image per each field trip was selected, the specific dates were chosen

taking into account to select the most clear and cloud free images and to

be at the middle of the period as much as it can. There are three reasons

to follow this approach in matching up process; first reason is because

this research was proposed after samples was taken, the second is that the

change of water quality parameters occurred slowly except in flood

season, the third reason is because the sampling date affected by the

whole survey plan of monitoring program such as hydrographic survey,

so that, it is not possible to get samples on the basis of scheduled MERIS

overpasses. Six images were selected for purposes of validation of Case 2

water quality processors and regression analysis between reflectance and

other water quality parameters.

- Images used for time series

Remote sensing images (MERIS images) for period of ten years from

December, 2002 to March, 2012 were downloaded with permission of

European Space Agency (ESA) under the umbrella of TIGER II project.

A full swath Level 1 MERIS images were downloaded using MERIS

FRS regional extraction tool, it is an internet site which provides a

services for accessing Envisat MERIS Level 1 and Level 2 Full

Resolution Full Swath data products.

55

MERIS images were downloaded by specifying Lake Nasser bounding

box coordinates in the extraction tool, the total downloaded images was

1824 images, a filter process were conducted to the downloaded images

in order to get rid of the images that not cover all area of Lake Nasser and

repeated image. A subset process for filtered images had been done to

reduce the extents of image to area of Lake Nasser, which helps in

reducing time needed for image processing.

Table (‎3-4) Characteristics of Match-up images.

First

Mission

Second

Mission

Third

Mission

Fourth

Mission

Fifth

Mission

Sixth

Mission

Period From 23-Oct-03 27-Nov-06 15-Nov-07 24-Apr-09 21-Aug-10 04-Oct-11

To 15-Nov-03 12-Dec-06 24-Nov-07 08-May-09 31-Aug-10 15-Oct-11

Matchup

image 07-Nov-03 01-Dec-06 19-Nov-07 01-May-09 21-Aug-10 10-Oct-11

Sun zenith 44.65-43.58 50.24- 49.1 47.16-45.98 28.97-

27.92

29.54-

28.65

35.04-

34.13

Sun azimuth 152.11-

149.45

153.94 :

151.69

155.19-

152.76

102.99 :

98.52

109.27-

104.76

148.34-

144.72

View zenith 17.64- 8.37 17.69 : 8.43 27.87-

19.64

30.39-

21.87 14.92- 5.20

30.44-

22.36

View

azimuth

283.77-

283.22

283.77-

283.21

284.34-

283.75

101.83-

101.41

102.66-

102.18

284.41-

283.81

Figure ( 3-16) shows a histogram for Downloaded MERIS images for

period from December 2002 until February, 2012. The total number of

images used in time series, after performing filter process, is 913 images.

The frequency of images per month, as shown in figure, is ranging from

12 images to one image per month. There are some months that have zero

frequency, especially in the first two years of lunching Envisat.

56

Figure (‎3-16) Histogram of downloaded MERIS images

0

2

4

6

8

10

12

14

De

c-2

00

2

May

-20

03

Oct

-20

03

Mar

-20

04

Au

g-2

00

4

Jan

-20

05

Jun

-20

05

No

v-2

00

5

Ap

r-2

00

6

Sep

-20

06

Feb

-20

07

Jul-

20

07

De

c-2

00

7

May

-20

08

Oct

-20

08

Mar

-20

09

Au

g-2

00

9

Jan

-20

10

Jun

-20

10

No

v-2

01

0

Ap

r-2

01

1

Sep

-20

11

Feb

-20

12

Fre

qu

en

cy

Months

57

CHAPTER (4)

METHODOLOGY

4.1. General

This chapter is devoted to present description of research methodology

used to fulfill the objectives of this research, whether used in sampling

and field/lab analysis, acquisition of MERIS images, or in the validation

and regression analysis. In addition, the tools used for images processing

will be described as well. Figure ( 4-1) represents a flow chart for research

methodology which consists of two main parts; field work and satellite

images acquisition. Figure highlights only the main steps of

methodology; the detailed steps of methodology will be described in the

following sections.

4.2. Field work

Field work was carried out through survey missions of Lake Nasser,

water samples were taken at specified locations in Lake Nasser. Water

samples were collected vertically to represent water column at each

sampling point, five vertical samples were taken at 0.50 m below water

surface, 25%, 50%, 65% and 80% of water depth at each location. Water

sample at depth 0.5 m under water surface will be used through this

research as it is the effective in remote sensing. A laboratory tests for

different parameters were done according to the Standard Methods for

the Examination of Water and Wastewater (APHA, 2012) to get

concentration of each parameter. The sampling procedure and both field

and laboratory equipment needed in measuring different in situ water

quality parameters will be described later in Appendix A.

4.3. Satellite images acquisition

Satellite images acquisition is considered the main part of methodology

to fulfill the objectives of this research. The spatial coverage of the image

was greater than the current study area, so that, subset process for images

58

was performed to reduce image spatial coverage to the bounding box

coordinates of Lake Nasser, hence, the time needed for image processing

was reduced.

There two main objectives of this thesis, the first is to validate the use of

Case2 water quality processors in Lake Nasser, while the second is to

find a relationship between different water quality parameters and

Remote sensing reflectance extracted from MERIS images.

Match-up images were specified based on the dates of field survey as

indicated in Table ( 3-4). Both validation and regression models will be

based on comparing measured data with data extracted from the

atmospherically corrected match-up images.

Figure (‎4-1) Flow Chart of Research Methodology

4.3.1. Image processing tools

Many tools will be used in order to process MERIS image, it will be

described as follows:

59

- BEAM (V4.10)

BEAM is an open-source toolbox used for processing, analyzing and

viewing of remote sensing images. It is originally developed to facilitate

and utilize image data acquired from Envisat's optical instruments

(Brockmann, 2014). It is composed of the three main components, 1)

VISAT which is an application used in processing, analyzing and

visualization of remote sensing raster images. 2) A set of tools that can be

run either from BEAM command line or found in VISAT, 3) A rich Java

API that can be used in programming of new BEAM plugins or

developing new remote sensing applications.

- Case 2 water processors

These processors are a BEAM plugins; it has been developed by GKSS

Research Center, Institute for Coastal Research, and Brockmann Consult,

under contract of the European Space Agency. It uses an algorithm

developed by (Doerffer and Schiller, 2008), the algorithm is found as a

neural network which relates the radiance detected by MERIS to surface

reflectances and then calculates the concentration of optical water quality

constituents.

Case 2 Regional (C2R) processor (V1.5.8) is A BEAM plugin will be

used through the current research, similarly, Lakes processor (V1.58) will

be used too. Lakes processors include two inland water modules for

boreal and eutrophic lakes.

Processors were validated using several remote sensing campaigns that

were conducted during the summer of 2007 in Finland (3 campaigns and

4 lakes), Germany (3 campaigns and 1 lake), and Spain (8 campaigns and

6 lakes). In addition to European lakes, validation activities were also

conducted in Lake Victoria, Africa, using data collected before the start

of the project, in 2003, 2004 and 2005 and in Lake Manzalah, Egypt,

with data collected in 2007 (Koponen et al., 2008).

http://www.brockmann-consult.de/cms/web/beam/

60

- MERIS ICOL Processor (V2.9.2)

Improved Contrast between Ocean and Land (ICOL) is BEAM plugin

was developed by (Santer and Zagolski, 2009), can be implemented with

the case2 water quality processors to correct for adjacency effects

(increased radiance due to scattering and reflection of photons). It was

developed as a result of a detailed analysis of the adjacency effect

carrued out by (Santer and Schmechtig, 2000). More details about ICOL

processor can be found in Appendix (A).

- SMAC Processor (V1.5.203)

Simplified Method for Atmospheric Corrections (SMAC) is a semi-

empirical algorithm used to correct MERIS image from the atmosphere

effect by introducing some approximations to the radiative transfer in the

atmosphere. The algorithm is developed by (Rahman and Dedieu, 1994).

- Cloud Probability Processor (V1.5.203)

It is a BEAM plugin used to implement the detection of clouds in MERIS

L1b products.

- ArcMap

ArcMap is used through this research in many processes such as

computing the percentage of cloudy areas over the lake surface,

calculating the monthly average water quality parameter over Lake

Nasser surface based on the results of regression models, creating

shapefiles for lake surface and extracting the monthly average water

quality parameter based on it. All these processes were applied to the 913

images used in the time series either by using ArcMap Model Builder or

by writing scripts by python programming language. Python is a free and

open source language, it used as a scripting language in ArcGIS

geoprocessing.

4.3.2. MERIS Processing tool (MPT)

MERIS Processing tool (MPT) was created using Visual Basic for

Application (VBA) found in Excel in order to automate the process of

61

subset, reproject, and preform water quality processors. In addition to

preform atmospheric correction, band math and cloud probability. The

description of MPT will be shown later on in Appendix (B). It helps in

processing large number of MERIS images needed for validation of

water quality processors, regression analysis and time series.

4.4. Validation of water quality processors

Water quality processors are mainly developed in order to estimate the

optically active substances found in water such as Chlorophyll-a (Chl-a),

Total Suspended Sediment (TSS). Case2 water processors such as C2R,

eutrophic lake and boreal lake processors will be validated in Lake

Nasser and the effect of ICOL processor will be taken into account.

Different processors will be applied to the match up images for different

field missions as follows:

1- ICOL processor was carried out to match-up images in order to

correct the adjacency effect of water pixels.

2- Case 2 water processors (Case 2 Regional, eutrophic lake and

boreal lake processors) were applied to the raw match-up images,

i.e., without using ICOL processor. Also, Case 2 water processors

were applied to images resulted from ICOL processor.

3- Data were extracted from the resulted images for two cases (with

and without using ICOL processor) at the same locations of in situ

measurements.

4- A comparison between the predicted values of TSS and Chl-a

with the field measurements will be conducted to validate

processors in Lake Nasser.

5- An evaluation of each processor will be performed in order to

select the best water quality processor to be used in the lake and

to what extent it can be applied in different seasons to obtain

water quality.

6- The evaluation procedure will be based on many statistical

performance measures like Fractional Bias (FB), Correlation

62

Coefficient (r), Normalized Mean Square Error (NMSE),

Geometric Mean Bias (MG), Geometric Mean Variance (MV)

and Factor of two (Fac2).

7- The description of statistical measures terms in addition to the

evaluation and ranking methods of water quality processors will

be shown below in section 4.7

4.5. Regression

Regression gives, in many cases, strong and best relationships between

surface reflectance and different optical water quality parameters. Many

examples can be found in the remote sensing literature that claim the

ability to measure water quality variables that do not directly affect light

reflectance or are present in natural waters at such low concentrations

that they do not affect reflectance signals measured by satellite sensors.

Through current research, many parameters can be correlated to the

extracted surface reflectance of 15 bands of match-up MERIS images.

According to Table ( A-2) which shows bands application it can be notice

that bands from 11 to 15 contain data for O2 absorption and it can be used

in atmospheric correction (ESA, 2006), so that only first ten bands will

be used in regression analysis, method of analysis was presented in

Figure ( 4-1) and it can be described below:

1- Surface reflectance will be calculated by applying a Simplified

Model for Atmospheric Correction (SMAC) to the raw match-up

images. SMAC is found also as plugin in BEAM

2- Extracting surface reflectance from the atmospherically corrected

MERIS images at the location of in situ measurements.

3- Pearson correlation matrix between measured water quality

parameters will be calculated to check influence of optical water

quality parameters such as TSS and Chl-a on other parameters.

4- Regression analysis using stepwise regression will be applied

separately for each field mission in order to automate the choice

of variable and regression model based on

63

5- Validating the resulted regression model, the validation will be

performed only for the end of flood as this period has different

measurements during this period.

4.6. Time Series

This section will describe the method and tools created in order to make a

time series curves of each water quality parameters. Concentration of

each will be calculated using the validated models resulted from stepwise

regression. Dates of each season will be defined using the upstream AHD

hydrograph; the images will be categorized for each season to apply the

different regression models. Figure ( 4-2) represents a flow chart for steps

used to create time series curves.

Three different processes will be performed to MERIS image using MPT

as follows:

1- Atmospheric correction using Simplified Method for Atmospheric

Correction (SMAC) to calculate surface reflectance of each image

band.

2- Band Math will be used in order to extract the lake surface using

Normalized Difference Water Index (NDWI)

𝑁𝐷𝑊𝐼 =𝑟𝑎𝑑𝑖𝑎𝑛𝑐𝑒_5−𝑟𝑎𝑑𝑖𝑎𝑛𝑐𝑒_10

𝑟𝑎𝑑𝑖𝑎𝑛𝑐𝑒_5+𝑟𝑎𝑑𝑖𝑎𝑛𝑐𝑒_10 (4)

This index was found by author by comparing the spectrum

characteristics of land pixels and water pixels, and it is found that

NDWI values greater than 0.1 which is identifying the water

surface.

3- Cloud probability processor will be used to classify clouds over

each pixel into three classes as cloud free, cloud uncertain or

cloudy.

Outputs from the previously described processes will pass through many

steps to create the time series curves. The cloud free percentage over the

whole lake surface will be calculated using an ArcMap model builder

64

represented in Appendix (B), Figure ( B-4). Images that have a cloud free

percentage lower than 95% of the lake surface will be excluded from

time series. Sun glint effect will be used also to exclude images from

time series. The lake surface shapefile will be extracted using an ArcMap

model builder represented in Appendix (B), Figure ( B-5). Regression

models for each water quality parameters will be applied to SMAC

output using MPT. A python code was written to calculate the average

monthly concentration for each parameter.

Figure (‎4-2) Time series flowchart

4.7. Statistical performance measures

The statistical methods used here to evaluate the use of water quality

processors was originated by (Fox, 1984) and modified by (Hanna,

1989), the method was used to evaluate several air quality models Hanna

(1993) and also used in statistical analysis on wheat experiments (Patryl

& Galeriua, 2011). The main idea is to find out the reliable and accurate

model when compared to the real and actual measured values.

65

Two types of performance measures will be used in evaluating and

selecting of water quality processor to be used in Lake Nasser, the first is

measure of difference which represents a quantitative estimate of the size

of difference between predicted and observed values, while the second is

measurement of correlation, it is a quantitative measure of association

between predicted and observed values. The correlation coefficient is a

popular method for evaluating models, perfect correlation between the

predicted and observed is a necessary condition for a perfect model, but it

is still not sufficient as the only criterion for model evaluation.

4.7.1. Definitions

In order to compare the computed water quality parameters from

different processors and observed values, several statistical performances

measures will be used such as:

Model Bias

Model Bias can be defined as the difference between computed

concentration (CP) and observed concentration (CO). It is given by:

OP CCBiasModel (5)

Correlation Coefficient (r)

The correlation coefficient is a numerical value that describes the

quantitative relation between two variables. A good model should

indicate correlation coefficient close to unity.

The correlation coefficient can be calculated as follows:

PoP CC

PPOO CCCCr

(6)

Mean Normalized Bias (MNB)

It averages the residual between computed and observed values

normalized by observed values. A best model should have a MNB value

close to zero.

66

%1001

o

op

C

CC

NMNB (7)

Fractional Bias (FB)

The fractional bias is a dimensionless number and is based on a linear

scale and refers to the arithmetic difference between CP and Co. It

indicates only systematic errors which lead to always under estimate or

overestimate the measured values (Patryl & Galeriua, 2011). It is written

in symbolic form as:

PO

PO

CC

CCFB 2 (8)

Values for the fractional bias range between -2.0 (extreme under-

prediction) to +2.0 (extreme over-prediction)

Normalized Mean Square Error (NMSE)

Normalized Mean Square Error (NMSE) is used to measure the scattering

of data. It can indicate both systematic and random error. The Smaller

value of NMSE indicates a better performance of the model. The NMSE

is given by:

PO

PO

CC

CCNMSE

2

(9)

Geometric Mean Bias (MG)

The geometric mean bias (MG) like FB, indicates only systematic errors

which leads to always under estimate or overestimate the measured

values but it is based on logarithmic scale (Patryl & Galeriua, 2011). MG

is given by:

PO CCMG lnlnexp (10)

Geometric Mean Variance (VG)

67

The geometric mean variance (VG) can indicate both systematic and

random error and is given by:

2lnlnexp PO CCVG (11)

Factor of two (Fac2)

The factor of two (Fac2) is defined as the percentage of the predictions

within a factor of two of the observed values. The ideal value for the

factor of two should be 1 (100%).

Fac2 = Fraction of data that satisfy 0.5 CP/CO 2

In order to calculate all statistical performance measure, a custom excel

function was created to save time needed for manual calculations.

Appendix (B) contains the programming code for these custom functions.

4.7.2. Interpretation of Performance measures

The three Water quality processors tested through this research will be

validated using specific criteria for each statistical measure. The quality

of an ideal and perfect processor should indicate values of statistical

measures as follows:

The values of both fractional bias and normalized mean square

error should equal to zero.

The fractional bias gives a positive weight for under predictions

and negative weight for over predictions.

The fractional bias for cases with factors of 2 under prediction

and over prediction are -0.67 and +0.67, respectively.

Geometric mean bias and geometric mean variance value should

equal to 1.

4.7.3. Criteria for good processor

The following criterion was suggested by (Kumar et al., 1993) for good

model and to accept its performance:

𝑁𝑀𝑆𝐸 ≤ 0.5 (12)

68

−0.5 ≤ 𝐹𝐵 ≤ +0.5 (13)

𝐹𝑎𝑐2 > 0.80 (14)

Also two additional criteria for MG and VG were given by (Ahuja &

Kumar, 1996) as a useful test of model performance:

0.75 ≤ 𝑀𝐺 ≤ 1.25 (15)

And 0.75 ≤ 𝑉𝐺 ≤ 1.25 (16)

4.7.4. Ranking and Evaluation

In order to evaluate which water quality processor is valid, the following

ranking method was implemented:

All statistical performance measures, declared above, will be

taken in the ranking process except Mean Normalize Bias as there

is no defined criterion for best model found in literature. Also

correlation coefficient will not be used in ranking.

One degree will be given for each performance measure that

complies with the criteria for best processor, and zero for which

not comply.

The sum of these degrees for each processor divided by number

of statistical measures, used in raking, expresses the ranking

degree each processor.

𝑅𝑎𝑛𝑘𝑖𝑛𝑔 𝑉𝑎𝑙𝑢𝑒 =𝑁𝑜. 𝑜𝑓 𝐶𝑜𝑚𝑝𝑙𝑖𝑒𝑑 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑠

𝑁𝑜. 𝑜𝑓 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑠 (17)

Six ranking degrees are proposed in the following list:

- Not valid 0

- Poor 0.2

- Accepted 0.4

- Good 0.6

- Very good 0.8

- Excellent 1

69

CHAPTER (5)

RESULTS AND DISCUSSION

5.1. Introduction

In this chapter, the most important results will be illustrated. First,

predicted concentration of TSS and Chl-a using Case 2 water quality

processors will be compared with the in situ measurements. Several

longitudinal profiles and scatter plots have been made and analyzed to

evaluate the behavior of each processor; the selection of the suitable

processor for Lake Nasser in different seasons will be based on using

some statistical performance measures. Second, Regression analysis

between measured water quality parameters and remote sensing

reflectance extracted from atmospherically corrected MERIS image will

be conducted. Regression analysis will be made for optical and non-

optical water quality parameters. Finally, the time series for each water

quality parameter will be created and analyzed.

5.2. Validation of Case 2 water quality processors

Validation of water quality processor will include only two parameters;

TSS and Chl-a as these processors were created to evaluate only

parameters whose change in concentration affects the color of water,

which called “optical parameters”. The following two sections will

discuss the results of validation.

5.2.1. Validation of TSS

Comparison between measured water quality parameters were discussed

previously in CHAPTER (3(. Figure ( 5-1) to Figure ( 5-5) show the

longitudinal profiles of TSS for C2R, Eutrophic and Boreal Lake

processors compared with the in situ measurements, in addition to the

spatial distribution of TSS. The upper part of each figure represents the

predicted TSS values without using ICOL processor and the lower is after

using ICOL.

70

The comparison between predicted TSS longitudinal profiles from the

three processors shows that Eutrophic lake processor gives higher

predicted values while Boreal lake processor gives a lowest value for all

survey missions, this for with and without using ICOL processor. The

spatial distribution of TSS represented in the same set of figures also

supports the previous notice. Figure ( 5-3) shows a quite different

behavior of boreal lake processor when using ICOL processor; most of

predicted values along the longitudinal profile are the lowest value but in

some locations in the lake gives values higher than the other two

processors.

Scatter plots between measured and predicted TSS values for three

processors are shown in Figure ( 5-6) and Figure ( 5-7). Summary of the

statistical performance measures are shown in, Table ( 5-1). The

underlined bold numbers cells indicate that the value lies within the limits

of good model. Table ( 5-2) shows the ranking of each processor

according to proposed ranking procedure.

71

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or

Wit

h u

sin

g I

CO

L P

roce

sso

r

Figure (‎5-1) TSS for First Mission (mg/l)

Wit

hout

usi

ng I

CO

L P

roce

ssor

Case 2 Regional Processor Eutrophic Lake Processor Boreal Lake Processor

Wit

h u

sing I

CO

L P

roce

ssor


0

5

10

15

20

25

30

35

40

45

50

0 50 100 150 200 250 300 350

TS

S (

mg/l

)


First Mission

C2R EUT BOR In situ

0

5

10

15

20

25

30

35

40

45

50

0 50 100 150 200 250 300 350

TS

S (

mg/l

)


First Mission

C2R EUT BOR In situ

72

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or

Wit

h u

sin

g I

CO

L P

roce

sso

r

Figure (‎5-2) TSS for Third Mission (mg/l)

Wit

hout

usi

ng I

CO

L P

roce

ssor


Wit

h u

sing I

CO

L P

roce

ssor


0

5

10

15

20

25

30

35

0 50 100 150 200 250 300 350

TS

S (

mg

/l)


Third Mission

C2R EUT BOR In situ

0

5

10

15

20

25

30

35

0 50 100 150 200 250 300 350

TS

S (

mg

/l)


Third Mission

C2R EUT BOR In situ

73

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or

Wit

h u

sin

g I

CO

L P

roce

sso

r

Figure (‎5-3) TSS for Fourth Mission (mg/l)

Wit

hout

usi

ng I

CO

L P

roce

ssor


Wit

h u

sing I

CO

L P

roce

ssor


0

5

10

15

20

25

0 50 100 150 200 250 300 350

TS

S (

mg

/l)


Fourth Mission

C2R EUT BOR In situ

0

2

4

6

8

10

12

14

16

18

20

0 50 100 150 200 250 300 350

TS

S (

mg

/l)


Fourth Mission

C2R EUT BOR In situ

74

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or

Wit

h u

sin

g I

CO

L P

roce

sso

r

Figure (‎5-4) TSS for Fifth Mission (mg/l)

Wit

hout

usi

ng I

CO

L P

roce

ssor


Wit

h u

sing I

CO

L P

roce

ssor


75

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or

Wit

h u

sin

g I

CO

L P

roce

sso

r

Figure (‎5-5) TSS for Sixth Mission (mg/l)

Wit

hout

usi

ng I

CO

L P

roce

ssor


Wit

h u

sing I

CO

L P

roce

ssor


0

5

10

15

20

25

30

35

40

0 50 100 150 200 250 300 350

TS

S (

mg

/l)


Sixth Mission

C2R EUT BOR In situ

0

5

10

15

20

25

30

35

40

0 50 100 150 200 250 300 350

TS

S (

mg

/l)


Sixth Mission

C2R EUT BOR In situ

76

a) 1

st Mission

b) 3

rd mission

c) 4

th mission

Figure (‎5-6) Measured verses predicted TSS Values for First, Third

and Fourth missions (mg/l)

77

a) 5

th mission

b) 6

th mission

Figure (‎5-7) Measured verses predicted TSS Values for Fifth and

Sixth missions (mg/l)

78

Table (‎5-1) Summary of statistical performance measures for TSS

Without ICOL With ICOL

C2R Eut_Lakes Bor_Lakes C2R Eut_Lakes Bor_Lakes

FIR

ST

MIS

SIO

N

No

v-0

3

r 0.971 0.949 0.956 0.922 0.931 0.747

MNB -52.324 -26.397 -63.823 -69.039 -56.623 -70.556

FB 0.645 0.264 0.882 0.916 0.635 1.044

NMSE 0.517 0.087 1.119 1.156 0.48 1.784

MG 2.144 1.389 2.853 3.646 2.595 3.94

VG. 1.869 1.165 3.205 6.792 3.16 8.683

Fac2 0.500 1.000 0.000 0.1 0.3 0.1

TH

IRD

MIS

SIO

N

No

v-2

00

7

r 0.97 0.966 0.969 0.963 0.959 0.975

MNB -14.667 15.212 -38.916 -26.077 -1.028 -45.592

FB 0.122 -0.146 0.452 0.234 -0.035 0.528

NMSE 0.03 0.054 0.305 0.078 0.035 0.402

MG 1.182 0.876 1.651 1.383 1.033 1.865

VG. 1.047 1.035 1.308 1.164 1.049 1.521

Fac2 1.000 1.000 1.000 0.909 1.000 0.545

FO

UR

TH

MIS

SIO

N

May

-20

09

r 0.671 0.672 0.684 0.694 0.698 0.689

MNB -6.497 27.673 -33.075 -37.627 -17.839 -46.841

FB 0.05 -0.259 0.367 0.421 0.147 0.579

NMSE 0.078 0.188 0.205 0.262 0.146 0.433

MG 1.122 0.82 1.586 1.718 1.31 2.02

VG. 1.118 1.14 1.398 1.55 1.257 1.929

Fac2 1.000 0.909 0.636 0.636 0.818 0.545

FIF

TH

MIS

SIO

N

Oct

-20

10

r 0.48 0.442 0.36 0.472 0.435 0.376

MNB 67.3 114.402 7.768 33.039 63.493 -8.247

FB 0.076 -0.084 0.656 0.142 -0.013 0.694

NMSE 0.681 0.619 2.287 0.748 0.667 2.384

MG 0.757 0.6 1.216 0.974 0.793 1.382

VG. 2.001 2.542 2.164 1.848 1.983 2.05

Fac2 0.750 0.375 0.563 0.813 0.625 0.625

SIX

TH

MIS

SIO

N

Oct

-20

11

r 0.976 0.972 0.968 0.973 0.967 0.968

MNB 52.622 133.157 14.522 18.2 69.197 -18.627

FB -0.257 -0.583 0.093 -0.135 -0.455 0.223

NMSE 0.092 0.468 0.107 0.046 0.359 0.138

MG 0.67 0.448 0.906 0.858 0.6 1.25

VG. 1.227 2.081 1.089 1.053 1.337 1.086

Fac2 1.000 0.556 1.000 1.000 0.778 1.000

79

5.2.1.1. Validation of first mission

This mission has been performed in the period of maximum water level

at the end of the flood. The behavior of water quality processors is shown

in Figure ( 5-6a). The values of Mean Normal Bias are smaller than zero

for all processors with and without using ICOL, which indicates a general

underestimation. The maximum value of correlation coefficient (r) is

observed for C2R without using ICOL processor. The higher value of (r)

is not evident that it is the best case as the other statistical measures for

the same case is outside the accepted limits. The statistical measures

indicates that Eutrophic Lake processor without ICOL is the most

appropriate to use during this field mission, the ranking value of that

processors is 0.8 indicating a very good degree for application, the values

is inside the accepted limits except for Geometric Mean Bias (MG), but

its value still the smallest compared to other processors.

5.2.1.2. Validation of third mission

Figure ( 5-6b) shows that the relation between predicted and observed

TSS values is around the 1:1 line for Eutrophic lake processor with and

without using ICOL. The statistical measures for all cases indicate that

there are three cases suitable for predicting TSS; C2R without using

ICOL and Eutrophic lakes processor with and without ICOL. The ranking

value for these processors, according to Table ( 5-2), is equal to one

which indicates an excellent degree of ranking. The correlation

coefficient for three processors ranges from 0.959 to 0.97, which

indicates a good correlation for all three cases. C2R without using ICOL

processor also gives a very good ranking as the ranking value is equal to

0.8 and the correlation between measured and estimated values equal to

0.963. So, it can also be used to estimate TSS in the third mission.

5.2.1.3. Validation of fourth mission

Although R values is relatively small for all processors, Eutrophic Lake

processor and C2R processor without using ICOL processor indicate an

80

excellent ranking of statistical performance measures, see Table ( 5-1),

which comply with the limits of good model. The most appropriate

processor according to the evaluation criteria is Eutrophic lake processor

without using ICOL.

5.2.1.4. Validation of fifth mission

The scatter plot represented in Figure ( 5-7a) shows the results of fifth

mission which performed at the start of flood period. The figure indicates

a different behavior for all processor that observed during the mission,

the relation between measured and predicted TSS values indicate that all

processors fail to estimate TSS value when it exceed 40 mg/l, where it

gives approximately a constant estimated value. Waters that carry a

heavy load of sediments can be classified as Case 3 waters (Liqin, 2014).

So that, in Lake Nasser both Case 2 and Case 3 water can be found

during the start of flood season. For values below 40 mg/l gives a good

trend with measured data. Some of statistical performance measures, such

as FB, GM, and Fac2, were complied with the criteria of good model but

it cannot give a good indication for best model. Also, the correlation

between estimated and measured is not good as it does not exceed 0.48.

5.2.1.5. Validation of sixth mission

Table ( 5-1) shows that there are three processors suitable for estimating

TSS in Lake Nasser and give an excellent ranking degree during the six

mission; the processors are boreal lake processor with and without using

ICOL and C2R processor with using ICOL. According to Table ( 5-2),

C2R without using ICOL processor can be taken into account for this

mission as the ranking value is 0.8 which means a very good comply with

the criteria of good model described before in chapter 4. Correlation

coefficient (r) for all four selected processors is ranging from 0.968 to

0.976, which gives a good correlation between field and estimated values.

81

Table (‎5-2) Ranking values of water quality processors for TSS



First Mission 0 0.8 0.0 0.0 0.2 0.0

Third Mission 1.0 1.0 0.6 0.8 1.0 0.2

Fourth Mission 1.0 1.0 0.4 0.4 0.6 0.2

Fifth Mission 0.4 0.2 0.2 0.6 0.2 0.0

Sixth Mission 0.8 0.2 1.0 1.0 0.4 1.0

5.2.1.6. Overall evaluation of TSS

The overall evaluation will be based on the ranking values, missions will

be grouped depending on the situation of lake and flood condition to

specify which processor will be used during different seasons.

First, third, and six missions were conducted during maximum water

level at the end of flood, the evaluation and selection of water quality

processors, for these mission, will be based on the sum of ranking values

divided by three. The overall ranking values are 0.6 for C2R without

ICOL, 0.67 for eutrophic lake processor without ICOL, and 0.6 for C2R

without ICOL processor. So that, the most appropriate processor that can

be used in this period is Eutrophic lake processor without using ICOL.

Fourth mission was conducted during the falling period, C2R and

eutrophic lake processor without ICOL can be used to estimate TSS

values as the overall rank value equal to one.

The evaluation of water quality processor to estimate TSS during rising

period, fifth mission, did not indicate a specific processor to estimate TSS

for all lake during this rising period as all processors failed to estimate

high concentrations.

The intensity of field mission during different seasons is not enough to

specify the suitable processor that can be used all over the year, for

example, during the rising and falling periods there are only one mission

per each period, so that the conclusion will be based on the results of one

82

mission. This is not enough to assure that it is the selected processor is

the best.

5.2.2. Validation of Chl-a

Validation of water quality processor to estimate Chl-a concentration in

Lake Nasser will be based on the results from two field missions (third

and fourth). The validation will use the same procedure as followed for

TSS. Figures (5-8) and (5-9) show the longitudinal profiles of Chl-a,

estimate using three water quality processors, third and fourth missions

respectively.

The comparison of the three longitudinal profiles show that boreal lake

processor gives a higher values than C2R and eutrophic lake processors

with and without using ICOL processor while eutrophic lake processor

give lower values, that can be supported visually using the spatial

distribution maps shown in the same figures.

Figure ( 5-10) represents the scatter plot comparison between measured

and estimated Chl-a values. Tables (5-3) and (5-4) show different

statistical measured and ranking values respectively. The evaluation of

each mission will be discussed below for each mission.

83

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or

Wit

h u

sin

g I

CO

L P

roce

sso

r

Figure (‎5-8) Chl-a for Third Mission (mg/m3)

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or


Wit

h u

sing

IC

OL

Pro

cess

or


0

10

20

30

40

50

60

0 50 100 150 200 250 300 350

Ch

l-a

(m

g/l

)


Third Mission

C2R EUT BOR In situ

0

5

10

15

20

25

30

35

40

45

0 50 100 150 200 250 300 350

Ch

l-a

(m

g/l

)


Third Mission

C2R EUT BOR In situ

84

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or

Wit

h u

sin

g I

CO

L P

roce

sso

r

Figure (‎5-9) Chl-a for Fourth Mission (mg/m3)

Wit

ho

ut

usi

ng

IC

OL

Pro

cess

or


Wit

h u

sing

IC

OL

Pro

cess

or


0

10

20

30

40

50

60

0 50 100 150 200 250 300 350

Ch

l-a (

mg/l

)


Fourth Mission

C2R EUT BOR In situ

0

5

10

15

20

25

30

35

40

45

0 50 100 150 200 250 300 350

Ch

l-a

(m

g/l

)


Fourth Mission

C2R EUT BOR In situ

85

a) 3

rd mission

b) 4

th mission

Figure (‎5-10) Measured verses predicted Chl-a Values (mg/m3)

5.2.2.1. Validation of third mission

The correlation between measured and estimated value is very poor, as

indicated in Table ( 5-3), the values are not exceed 0.0231, also the values

of mean normalized bias are very high and indicate that there is a big

scatter in the data and the positive sign indicates that the processors give

an over prediction for all processed cases except for eutrophic lake

processor with using ICOL, the mean normalized bias has the lowest

value, in addition to the comply of good model criteria for FB, NMSE

and MG.

86

Table ( 5-4) indicates a ranking value of 0.6 for eutrophic lake processor

with using ICOL, which mean, according to the ranking list shown

before, that it gives good results compared to the in situ measurements.

5.2.2.2. Validation of fourth mission

Table ( 5-3) indicates that C2R and eutrophic lake processors with using

ICOL comply with the criteria of good model for FB, NMSE and MG.

The overall ranking values indicate the same value of 0.6 for both

processors, which indicate a good ranking, while the correlation is very

poor and did not exceed 0.35 between measured and estimated Chl-a

values.

Table (‎5-3) Summary of statistical performance measures for Chl-a



TH

IRD

MIS

SIO

N

No

v-2

00

7

R 0.0231 0.0021 0.000011 0.001 0.010 0.018

MNB 79.390 72.679 324.959 62.751 -7.671 115.335

FB -0.417 -0.362 -1.114 -0.301 -0.230 -0.996

NMSE 0.387 0.346 2.360 0.341 0.331 2.029

MG 0.630 0.666 0.286 0.719 0.774 0.352

VG. 1.578 1.543 7.040 1.520 1.517 5.182

Fac2 0.636 0.636 0.273 0.700 0.700 0.200

FO

UR

TH

MIS

SIO

N

May

-20

09

R 0.080 0.089 0.103 0.308 0.350 0.352

MNB 87.795 74.755 341.599 28.854 16.981 155.765

FB -0.513 -0.446 -1.190 -0.103 -0.005 -0.720

NMSE 0.383 0.305 2.418 0.284 0.264 1.278

MG 0.576 0.617 0.250 0.922 1.013 0.523

VG. 1.567 1.450 8.225 1.408 1.389 2.776

Fac2 0.727 0.727 0.000 0.636 0.636 0.545

Table (‎5-4) Ranking values of water quality processors for Chl-a



Third Mission 0.4 0.4 0 0.4 0.6 0

Fourth Mission 0.2 0.4 0 0.6 0.6 0

87

5.2.2.3. Overall evaluation of Chl-a

The statistical performance measures and overall ranking of Chl-a for

third and fourth missions indicate that eutrophic lake processor with

using ICOL process indicates good results.

Although the results are good for estimating Chl-a values, it cannot be

concluded that the processor is valid to estimate Chl-a concentration for

the entire year.

5.3. Regression

The relation between reflectance values and measured water parameters

can be found by three approaches empirical, analytical, semi analytical.

Through this research, the empirical approach will be used; regression

analysis will be conducted between different water quality parameters

and reflectance of different bands of MERIS image. The regression

analysis will not only be for optical parameters but it will also include the

non-optical parameters. Pearson correlation matrix will be calculated to

find the effect of optical parameter on others; strong correlation can be an

evident that non-optical parameter can be correlated to different bands’

reflectance.

5.3.1. Correlation between measured parameters

The linear correlation (Pearson correlation) between different water

quality parameters is shown in Table ( 5-5). The analysis will be shown

below, for each water quality parameter, to show its behavior with the

change of optical parameters such as TSS and Chl-a. The interpretation

of correlation coefficient and to how extent the parameters were linearly

correlated will be described using the guidelines guideline suggested by

Hinkle et al. (2003) as illustrated in Table ( 5-6).

Turbidity is a measure of water clarity and how much the suspended

material in water decreases the passage of light through the water. So

that, a very strong positive relationship between TSS and Turbidity was

88

found, the correlation factor is 0.994 and 0.999 for First and fifth

missions respectively.

Transparency means how deep sunlight penetrates through the water and

measured with a Secchi disk. It is also considered as measurement of

water clarity as it depends on the amount of particles in the water. These

particles can be algae or suspended sediments, the more particles, the less

water transparency. So that transparency has a strong negative correlation

with TSS as shown in Table ( 5-5), Pearson correlation factor ranging

from a minimum value of -0.807 in fifth mission to -0.928 in third

mission.

Pearson correlation factor between Total phosphorus (TP) and other

water quality parameters is shown in Table ( 5-5), the correlation between

TSS and TP is a positive strong to very strong relation for second to fifth

mission, it ranges from 0.846 for fourth mission and 0.974 in third

mission while first mission indicates correlation factor of 0.696.

Silica (SiO2) indicates a positive correlation with TSS for first, second

and third missions as the values of correlation factor are 0.759, 0.576 and

0.748 for the three missions respectively. A different behavior of SiO2 is

found during fourth mission, as it gives a negative correlation with TSS

with a value of -0.570. According to the indicated values the relation can

be described as strong to moderately strong relationship.

Total Dissolved Salts (TDS) is showing, in common, a strong negative

correlation with TSS, as indicated in Table ( 5-5), for first and third

missions while gives a positive correlation for fifth mission. Values of

correlation are -0.970, -0.837 and 0.835 for the indicated missions

respectively.

89

Table (‎5-5) Pearson correlation matrix

a- First Mission

TSS Turbidity Transparency TP Sio2 TDS

TSS 1 - - - - -

Turbidity 0.994 1 - - - -

Transparency -0.917 -0.889 1 - - -

TP 0.696 0.696 -0.737 1 - -

SiO2 0.759 0.713 -0.839 0.329 1 -

TDS -0.970 -0.947 0.897 -0.565 -0.799 1

b- Second Mission

Turbidity Transparency TP Sio2 TDS

Turbidity 1 - - - -

Transparency -0.882 1 - - -

TP 0.971 -0.933 1 - -

SiO2 0.576 -0.856 0.644 1 -

TDS -0.686 0.617 -0.607 -0.431 1

c- Third Mission

TSS Chl-a

Transparenc

y TP SiO2 TDS

TSS 1 - - - - -

Chl-a -0.687 1 - - - -

Transparenc

y -0.928 0.537 1 - - -

TP 0.974 -0.656 -0.938 1 - -

SiO2 0.748 -0.571 -0.831 0.801 1 -

TDS -0.837 0.524 0.943 -0.871 -0.878 1

d- Fourth Mission

TSS Chl-a Transparency TP SiO2

TSS 1 - - - -

Chl-a -0.733 1 - - -

Transparency -0.814 0.653 1 - -

TP 0.846 -0.498 -0.877 1 -

SiO2 -0.570 0.335 0.654 -0.610 1

e- Fifth Mission

TSS Transparency Turbidity TP TDS

TSS 1 - - - -

Transparency -0.807 1 - - -

Turbidity 0.999 -0.819 1 - -

TP 0.852 -0.660 0.847 1 -

TDS 0.835 -0.902 0.841 0.750 1

90

The correlation between Chl-a and TSS as indicated in Table ( 5-5) is a

negative correlation with values -0.687 and -0.733 for both third and

fourth missions respectively. While it gives a positive correlation with

transparency with correlation not exceeds value of 0.653. TP also gives a

negative correlation with Chl-a with values of -0.656 and -0.498. Silica

(SiO2) is indicating a different behavior as it gives a negative correlation

with Chl-a in third mission with value of -0.571, and a positive

correlation is noted in fourth mission with value of 0.335. In general, the

correlation between Chl-a can be described as a moderately strong

correlation.

Table (‎5-6) Rule of Thumb for Interpreting the Size of a Correlation

Coefficient (Hinkle et al., 2003)

Size of Correlation

(r)

Interpretation

0.90 to 1.00 Very strong correlation

0.70 to 0.90 Strong correlation

0.50 to 0.70 Moderately strong correlation

0.30 to 0.50 Weak correlation

0.00 to 0.30 Negligible correlation

According to the previous analysis, it is found that there is a correlation

between optical water quality parameters and other non-optical

parameters, so that, the non-optical water quality parameters can be

correlated to remote sensing reflectance calculated from MERIS images.

5.3.2. Stepwise regression Model

Simple regression analysis between water quality parameters, measured

during six missions, and remote sensing reflectance calculated from

MERIS images will be performed using stepwise regression. Stepwise

Regression is an iterative process used to construct a regression model. It

is semi-automatic selection process of independent variables carried out

in two ways by including independent variables in the regression model

one by one at a time if they are statistically significant, or by including all

91

the independent variables initially and then removing them one by one if

they prove to be statistically insignificant. Ten MERIS bands will be used

in stepwise regression, Figure ( C-1) to Figure ( C-6) in appendix (C)

represent Matlab output of stepwise regression for each water quality

parameter. Each figure contains a graph showing the coefficient with

error bar, t-stat, p-value for each band reflectance, where the bands

reflectance is designated by (X1 to X10). In addition to other factors that

can be used in evaluating the regression models such as RMSE, R-

squared, F and p values for overall model. Summary of regression

analysis are presented in Tables ( 5 7) and ( 5 8), retrieval models for

different water quality parameter are grouped by mission, where each

parameter has a different retrieval model for each mission. A description

of retrieval models for each water quality parameter will be discussed

below.

Retrieval models between TSS and surface reflectance show that TSS can

be correlated to different bands, as indicated in Table ( 5-7). It is found

that TSS can be correlated to reflec_5 for both first and fourth mission.

While reflec_6 and reflec_8 can be used to predict TSS for third and fifth

missions respectively. For sixth mission, TSS can be obtained by a linear

equation between both reflec_1 and reflec_5. The R-squared value is very

strong for all TSS retrieval models for all missions except fourth mission;

it has a minimum value of 0.881 and maximum value of 0.969. While for

fourth mission R-squared value is 0.578, the low value for fourth mission is

attributed to conducting this mission during falling period where the influence

of TSS is very low.

Chl-a was measured only twice; in third and fourth missions. The results

of stepwise regression show that reflec_8 can be used as a predictor of

Chl-a during the third mission, but with low R-squared value (0.431).

Stepwise regression is failed to get an equation for fourth mission.

Turbidity was measured in three missions; the retrieval models show that

R-squared value is not less than 0.823 which indicate a very good

92

correlation. Also, it shows that turbidity can be correlated to many bands

reflectance depending on the season of the mission.

Transparency is considered as a measure of water clarity, so that, its

retrieval models for it have a strong R-Squared values ranges from 0.893

to 0.984, it is noted that, retrieval models for different missions indicate

that reflec_5 can be used as a predictor for transparency, it can be used

alone as in fifth mission or it can be used with other band reflectance

such as reflec_4 and reflec_2 for first, second and third missions. Fourth

mission indicates a combination of other two bands which are reflec_3

and reflec_10.

Stepwise regression between TP and surface reflectance of different

MERIS images bands indicate that reflec_8 is the suitable for TP

prediction for both fourth and fifth missions. Where reflec_5 and reflec_2

are suitable for second and third missions respectively. The R-squared

value is ranging from 0.55 for fourth mission to 0.986 for third mission.

Silica (SiO2) was measured during four mission. Retrieval model for first

mission shows that reflec_8 give R-squared value of 0.47 and fourth

mission has R-squared value of 0.45 with reflec_1, this indicates that the

relation is not good but still is the best to predict SiO2 concentration

during these two missions. While second and third mission were give an

R-squared values of 0.796 and 0.847 respectively. Where the retrieval

models shows that reflec_5 and reflec_6 are used to estimate SiO2 for

second mission, while reflec_4 and reflec_9 for third mission. The

resulted equation did not show correlation to a common band reflectance.

Retrieval models for TDS indicate that reflec_6 can be used as a

predictor for first, second and fifth missions, the R-squared values are

0.808, 0.446 and 0.814 for these missions respectively. While both

reflec_2 and reflec_4 can be used during third mission.

93

Table (‎5-7) Retrieval Models for water quality parameters (First, Second and third missions)

Mission Parameter Retrieval Model R-Sq Ajd R-Sq RMSE F P

1st TDS 164.488-395.906*reflec_6 0.808 0.784 2.38 33.719 0.0004

Transparency 0.207+436.153*reflect_4-346.301*reflec_5 0.889 0.857 0.542 28.009 0.00046

SiO2 2.082+225.371*reflec_8 0.47 0.4 2.015 7.088 0.0287

Turbidity -0.056+10.066*reflec_4-9.331*reflec_9 0.823 0.773 0.0379 16.329 0.0023

TSS -11.412+316.079*reflec_5 0.917 0.906 1.858 87.943 1.37E-05

2nd

TDS 164.505-150.108*reflec_6 0.446 0.391 7.308 8.066 0.017

TP -0.0124+3.275*reflec_5 0.878 0.866 0.0495 72.259 6.89E-06

SiO2 6.542+233.258*reflec_5-155.372*reflec_6 0.796 0.751 1.664 17.594 0.00077

Transparency 2.45+50.634*reflec_2-38.832*reflec_5 0.983 0.977 0.178 254.316 1.21E-08

Turbidity 8.189-365.413*reflec_2+334.423*reflec_6 0.913 0.894 3.805 47.227 1.69E-05

3rd

Chl-a 9.81-76.269*reflec_8 0.431 0.638 2.444 6.825 0.028

TSS 1.914+258.003*reflec_6 0.969 0.965 1.66 276.867 4.57E-08

TP -0.019+8.246*reflec_2 0.986 0.984 0.014 654.597 1.03E-09

Transparency 3.297+99.144*reflec_2-71.577*reflec_5 0.978 0.972 0.235 174.839 2.50E-07

SiO2 11.769+102.322*reflec_4-85.277*reflec_9 0.847 0.809 0.51 22.193 5.40E-04

TDS 154.985+2038.49*reflec_2-1701.5*reflec_4 0.923 0.904 4.01 47.934 3.52E-05

94

Table (‎5-8) Retrieval Models for water quality parameters (Fourth, Fifth and Sixth missions)

Mission Parameter Retrieval Model R-Sq Ajd R-Sq RMSE F P

4th TSS 0.954+120.622*reflec_5 0.578 0.531 1.847 12.324 0.00666

SiO2 9.295-180.949*reflec_1 0.45 0.389 0.813 7.354 0.0239

TP 0.0321+3.121*reflec_4-1.144*reflec_8 0.94 0.926 0.0058 63.13 1.26E-05

Transparency 2.336-35.723*reflec_3+19.756*reflec_10 0.893 0.867 0.0815 33.553 0.000129

5th TSS -33.274+632.144*reflec_8 0.881 0.873 9.266 104.013 7.32E-08

TDS 134.257+176.93*reflec_6 0.814 0.8 3.122 61.215 1.77E-06

TP 0.052+0.437*reflec_8 0.55 0.518 0.016 17.108 0.0011

Transparency 5.409.39.844*reflec_5 0.984 0.982 0.131 814.181 8.33E-14

Turbidity -32.166+599.642*reflec_8 0.889 0.881 8.448 112.602 4.46E-08

6th TSS 4.014-407.597*reflec_1+328.357*reflec_5 0.948 0.931 2.205 54.995 0.000138

95

The overall conclusion from the previous analysis is that for each mission

one can note that there are common bands suitable for predicting both

optical and non-optical water quality parameters for each mission. For

example, the retrieval models for both turbidity and transparency for

second mission contain reflec_2, reflec_5 and reflec_6, where models of

non-optical parameters have best correlation with reflec_5 and reflec_6.

The same can be found in other mission like fifth mission.

5.3.3. Validation of retrieval model

Water quality in Lake Nasser varied along the year due to the effect of

flood, especially in the southern part which influenced by suspended

sediment. Variation in water quality parameters resulted in different

regression models for each water quality parameter depending on the

season of measurement, as discussed in the previous sections.

Regression models validation will be based on selecting a model for each

water quality parameter; one per each season. Missions will be

categorized into groups depending on season. Each model will be

validated using all collected data for the same group; statistical

performance measures, such as MNB, NMSE, FB, MG, VG, Fac2 will be

used to select the suitable regression model for each season.

Field missions were conducted in three seasons at the start of rising

period, at end of flood season, and during the falling period. Figure ( 3-8)

illustrated the timing of each mission with respect to the change in water

level upstream AHD. There are four missions were conducted just after

the end of flood season, first three missions and sixth mission. So that,

the shared measured parameters will be validated using the proposed

method. For other two seasons, it is found that only one mission was

conducted per season, hence, it is not possible to validate its regression

models. So that, retrieval models for these seasons need to be validated in

order to have a graph for time series change for each mission.

Tables ( 5-9) and ( 5-10) show validation results for each retrieval model,

different statistical performance measures are calculated for each season.

96

The selection of the suitable retrieval model will be based on the criteria

used before in validating Case 2 water quality processors. Validation

process here will account for R-Squared value as additional criteria in

selecting the suitable retrieval model. Bold and underlined values means

that it comply with criteria of good model.

TSS were measured during three aforementioned seasons, during rising

and falling seasons it was measured only once while during end of flood

period it was measured three times. The statistical performance measures

indicate that retrieval model for the third mission, conducted during

November 2007, gives good comply with the criteria of good model for

both NMSE and FB. Also, it has the greatest value of R-Squared.

Chl-a was measured during two missions, in November 2007 and May

2009, as indicated before, but only one mission, during 2007, gives a

correlation with reflec_8, and the R-squared value for this model is

0.431.

Turbidity was measured twice during end of flood season, validation

results indicate that regression model for November 2003 is suitable to

get turbidity during this season where all statistical performance

measures have good comply with the criteria of good model except for

VG and has a greater R-squared.

Validation of retrieval models indicate that model for November 2006 is

the suitable model to get transparency during end of flood, all statistical

performance measures give values within limits of good model, in

addition, its R-squared value is the greater between other missions.

Total phosphorus (TP) was validated during end of flood season, the

results show that retrieval model of 2006 mission is suitable for

estimating TP as all statistical performance measures comply with criteria

of good model. Also, Silica (SiO2) validation results show that model for

2007 mission is suitable to estimate silica during the end of flood.

97

All statistical performance measures calculated for TDS comply with

limits for good model, although model of 2007 mission has the highest

R-squared value, but it will be excluded as the other models show a

relation with reflec_6 as a common band. So that, the suitable retrieval

model for estimating TDS is 2003 model.

98

Table (‎5-9) Retrieval Models Validation for different seasons (TSS, Chl-a, Turbidity and transperancy)

Parameter Mission Retrieval Model R-Sq MNB NMSE FB MG VG Fac2

TSS 2003 -3.954+270.496*reflec_5 0.913 80.908 0.633 -0.139 0.729 1.944 0.649

2007 1.914+258.003*reflec_6 0.969 111.571 0.463 -0.235 0.596 1.958 0.702

2011 4.014-407.597*reflec_1+328.357*reflec_5 0.948 -8.031 1.383 0.325 1.494 2.809 0.632

2009 0.954+120.622*reflec_5 0.578 - - - - - -

2010 -33.274+632.144*reflec_8 0.881 - - - - - -

Chl-a 2007 9.81-76.269*reflec_8 0.431 - - - - - -

Turbidity 2003 -11.412+316.079*reflec_5 0.917 46.585 0.242 -0.174 0.938 1.933 0.818

2006 8.189-365.413*reflec_2+334.423*reflec_6 0.913 65.850 0.240 0.031 0.795 1.789 0.591

2010 -32.166+599.642*reflec_8 0.889 - - - - - -

Transparency 2003 0.207+436.153*reflect_4-346.301*reflec_5 0.889 -238.301 5.915 1.129 - - 0.500

2006 2.45+50.634*reflec_2-38.832*reflec_5 0.983 9.430 0.099 0.061 0.978 1.123 0.938

2007 3.297+99.144*reflec_2-71.577*reflec_5 0.978 35.766 0.201 -0.279 0.845 1.352 0.844

2009 2.336-35.723*reflec_3+19.756*reflec_10 0.893 - - - - - -

2010 5.409-39.844*reflec_5 0.984 - - - - - -

99

Table (‎5-10) Retrieval Models Validation for different seasons (TP, SiO2, and TDS)

Parameter Mission Retrieval Model R-Sq MNB NMSE FB MG VG Fac2

TP 2003 -0.056+10.066*reflec_4-9.331*reflec_9 0.823 -15.677 0.558 0.276 1.685 9.320 0.848

2006 -0.0124+3.275*reflec_5 0.878 -0.509 0.063 0.054 1.072 1.166 0.909

2007 -0.019+8.246*reflec_2 0.986 107.866 0.451 -0.482 0.564 1.901 0.576

2009 0.0321+3.121*reflec_4-1.144*reflec_8 0.94 - - - - - -

2010 0.052+0.437*reflec_8 0.55 - - - - - -

SiO2 2003 2.082+225.371*reflec_8 0.47 -10.776 0.264 0.136 1.288 1.474 0.758

2006 6.542+233.258*reflec_5-155.372*reflec_6 0.796 10.719 0.038 -0.072 0.925 1.054 1.000

2007 11.769+102.322*reflec_4-85.277*reflec_9 0.847 17.603 0.049 -0.096 0.879 1.081 0.970

2009 9.295-180.949*reflec_1 0.45 - - - - - -

TDS 2003 164.488-395.906*reflec_6 0.808 -2.867 0.008 0.031 1.034 1.009 1.000

2006 164.505-150.108*reflec_6 0.446 5.687 0.006 -0.052 0.948 1.007 1.000

2007 154.985+2038.49*reflec_2-1701.5*reflec_4 0.923 4.901 0.013 -0.048 0.958 1.013 1.000

2010 134.257+176.93*reflec_6 0.814 - - - - - -

100

5.4. Time series analysis

The selected retrieval models will be used in mapping of the spatial

distribution for each water quality parameter for period of about ten years

form December, 2002 until March, 2012. An atmospheric correction will

be performed to MERIS image in order to get surface reflectance of each

band. MPT will be used as illustrated before to automate all process.

5.4.1. Image selection criteria

Total number of images used in time series analysis is 913 images. To

use image for time series analysis for water quality parameters, we

should assure that the cloud coverage over Lake Nasser is not so large,

also to assure that there is no sun glint found over the water surface. Sun

glint occurs in remote sensing image when sun light is reflected by water

surface towards the sensor. The sun glint can contribute in increasing the

water-leaving radiance from sub-surface features (Kay et al., 2009).

5.4.1.1. Cloud coverage

Clouds over Lake Nasser was calculated using Cloud Probability

Processor, it uses a clear sky conservative cloud detection algorithm

which is based on artificial neural nets developed by Rene Preusker. The

cloud probability algorithm has been developed and implemented by Free

University Berlin and Brockmann Consult. It is found as a plugin in

BEAM. The output product contains one raster indicating the cloud

probability for each pixel. The three flags indicate pixels which are

cloudy (probability > 80%), cloud free (probability < 20%) or where it is

uncertain (20% < probability < 80%).

Cloud probability processor is applied to the total number of 913 MERIS

images used in time series analysis using MPT. The output of cloud

probability is reprojected and converted to GEOTIFF file, and then water

pixels are extracted in separate product, in order to calculate the

percentage of each category of cloud probability over the surface of lake

in different seasons.

101

Figure ( 5-11) provides a summary of cloud analysis results; it indicates

the percentage of cloud free, cloud uncertain and cloudy pixels against

the date of each image. Criteria for excluding image from time series

analysis will be based on cloud free percentage, if the value of cloud free

percentage is less than 95%, the image will be excluded. So that, the total

number of excluded MERIS image is 185 images.

5.4.1.2. Sun glint

On the other hand, the excluded images due to sun glint will be

performed manually by extracting a true image and visually excluding

images that affected by sun glint. Total number of images excluded due

to sun glint is 284 images.

The total number of excluded images due to both clouds and sun glint is

469 images out of 913 images, distributed along the total period used for

time series analysis. A further analysis of the excluded images can be

done in order to find the best times for future water quality sampling to

be coupled with remote sensing images. Figure ( 5-12) is a bar chart

shows the percentage of excluded images due to clouds and sun glint

categorized by month. The minimum percentage of excluded images due

to cloud coverage is found during period from September to November,

the percentage of ranges from zero in October while during September

and November are 2.67% and 8.86% respectively. Also the period from

February to April also has relatively low percentage of excluded images

as the percentages are 14.86%, 10.96% and 24.62% respectively. The

maximum percentage of excluded images due to cloud coverage is

observed during May, August, and January, the percentages are 43.06%,

39.76 and 35.62% during the three periods respectively.

The excluded images percentage due to sun glint ranges from 18.46% to

14.29%, the percentage of exclusion is uniform during months, as

indicated in Figure ( 5-12).

102

a- Cloud free %

b- Cloud Uncertain%

c- Cloudy%

Figure (‎5-11) Cloud probability summary surface of Lake Nasser

0

20

40

60

80

100

120

1-Nov-02 9-Jan-05 20-Mar-07 28-May-09 6-Aug-11

Clo

ud

Fre

e (%

)

Date

95% Cloud Free %

-10

0

10

20

30

40

50


Clo

ud

Un

certa

in (

%)

Date

Cloud Uncertain % 5%

-20

0

20

40

60

80

100


Clo

ud

y (

%)

Date

Cloudy % 5%

103

Figure (‎5-12) Percentage of excluded images due to clouds and sun

glint

Hence, it can be concluded previous analysis that in May the cloud

probability is high, and then it decreases during June, while it starts to

increase during the next two month until August, followed by period that

has minimum cloud probability from September to November while it

increased again until it reaches to January, during February and March

there are another low cloud probability period.

5.4.2. Flood seasons

In order to create a time series for the different water quality parameters,

the dates of each season flood will be identified using US AHD

hydrograph. Figure ( 5-13) illustrates how to define start and end of

different seasons. The rising season begins with the start of flood and the

water levels starts to significantly increase until it reaches the maximum

water level, at this stage the end of flood season is started where it

characterized by a slowly decrease in water level, in this period the water

level may increase for period of two or three days. While the falling

period starts as illustrated in the figure where characterized by a

significant decrease in water level until the next flood. Based on this

definitions start and end of each flood season are identified for the period

from 2002 to 2012 as shown in Table ( 5-11).

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

% E

xcl

ued

im

ages

Month

Total

Clouds

Sunglint

104

MERIS images are seasonally categorized; different regression models

are applied to MERIS images using MPT in order to get the

concentrations of each water quality parameter.

Figure (‎5-13) Percentage of excluded images due to clouds and sun

glint

Table (‎5-11) Start/End of Flood Seasons

Rising End of Flood Falling

Start End Start End Start End

2-Nov-2002 14-Jan-2003 16-Jan-2003 25-Jul-2003

26-Jul-2003 31-Oct-2003 1-Nov-2003 15-Jan-2004 16-Jan-2004 24-Jul-2004

25-Jul-2004 8-Nov-2004 9-Nov-2004 9-Feb-2005 10-Feb-2005 25-Jul-2005



9-Jul-2007 19-Oct-2007 20-Oct-2007 18-Feb-2008 19-Feb-2008 27-Jul-2008

28-Jul-2008 8-Oct-2008 9-Oct-2008 10-Feb-2009 11-Feb-2009 4-Aug-2009

5-Aug-2009 28-Sep-2009 29-Sep-2009 30-Jan-2010 31-Jan-2010 1-Aug-2010

2-Aug-2010 10-Oct-2010 11-Oct-2010 13-Nov-2010 14-Nov-2010 3-Aug-2011

4-Aug-2011 16-Oct-2011 17-Oct-2011 16-Jan-2012 17-Jan-2012 5-Aug-2012

105

5.4.3. Time series of Water quality parameters

The time series for different water quality parameters will be used the

average monthly values for both curves and spatial distributions maps. So

that, after calculating the concentrations, images will be grouped by

months and average values were calculated for each group, a python code

are written to automate the calculation of average.

The primary purpose when planning to create time series is to track the

change in water quality for surface layer of Lake Nasser all over the year

during the period from 2002 to 2012, Figure ( 5-14) shows the monthly

average discharge for TSS as calculated using seasonal regression model

found during this research, in addition to the measured TSS value. The

comparison between the average monthly calculated concentrations and

measured concentrations indicate that there is a good agreement between

them during different months except for months of rising period of the

flood as the measured concentration during August, 2010 at Arkeen cross

section is 84 mg/l while the calculated average value is about 35.7 mg/l,

this big error in TSS will be reflected on the estimation of other values.

So that, it is decided to only perform time series for the season which

have a validated regression models, i.e. during the end of flood season as

the invalidated model may give us unrealistic values. October, November

and December was selected to represent the end of flood season

Figure (‎5-14) Monthly average TSS at Arkeen Cross section as

calculated from MERIS Images

106

Figure ( 5-15) shows a sample of the spatial distribution maps of monthly

average concentration of TSS, Transparency, and TP as samples of

concentration maps. Time series curves for TSS, Transparency TP, SiO2,

and TDS are drawn for October, November and December at location of

monitoring sections as indicated in Figure ( 5-16) to Figure ( 5-20). The

measure values are identified in the figures to assure that the obtained

time series indicate values in the same range as the measured one.

It is noted that TSS values at the southern part of the lake increases until

it reaches its maximum values in 2005, after that TSS started to decrease

until 2009. This pattern of change is noted in sections from Arkeen to

Ebream, this part is influenced by the suspended sediment transported by

flood. Generally the pattern of change can be similar to a sin curve. Also

it is noted that there is no specific pattern of change for the northern part

of the lake, starting from Krosko to High dam, this may be due to the

small concentrations.

Transparency in the southern part of the lake is not showing a specific

pattern and the estimated values are greater than the measured values,

while in the northern part it indicates a change in range of 0.5 m but it

gives under estimation when compared with the measured transparency.

This can be due to that transparency measurements depend on the vision

of secchi disk which may differ from person to other.

Silica concertation for section found in the northern part from High Dam

until reaches Krosko cross section changes in range of 0.2 mg/l. it is

noted that for cross sections from Ebreem to Arkeen, the values of Silica

is started to decrease until 2005 and started to increase again until 2009.

The changing pattern of Total Phosphorus is the same as that of TSS,

while for Total dissolved Solids it is the same as that of Silica.

107

10/2005 11/2005 12/2005

a- TSS

10/2005 11/2005 12/2005

b- Transparency

10/2005 11/2005 12/2005

c- TP

Figure (‎5-15) Spatial distribution maps of monthly average

concentration of TSS, Transparency, and TP (mg/l)

Legend

0.95

- 5

5.01

- 10

10.0

1 - 1

5

15.0

1 - 2

0

20.0

1 - 2

5

25.0

1 - 3

0

30.0

1 - 3

5

35.0

1 - 4

0

40.0

1 - 4

5

Legend

0.95

- 1

1.01

- 1.

2

1.21

- 1.

4

1.41

- 1.

6

1.61

- 1.

8

1.81

- 2

2.01

- 2.

2

2.21

- 2.

4

2.41

- 2.

6

2.61

- 2.

8

2.81

- 3

3.01

- 40

.6

Legend

0 - 0

.1

0.11

- 0.

2

0.21

- 0.

3

0.31

- 0.

4

0.41

- 1

108

Figure (‎5-16) TSS during Period from October to December (mg/l)

109

Figure (‎5-17) Transparency during Period from October to

December (mg/l)

110

Figure (‎5-18) Silica during Period from October to December (mg/l)

111

Figure (‎5-19) Total Phosphorus during Period from October to

December (mg/l)

112

Figure (‎5-20) TDS during Period from October to December (mg/l)

113

CHAPTER (6)

CONCLUSION AND RECOMMENDATIONS

Lake Nasser is considered the strategic storage of water in Egypt, so that,

the preservation and continues monitoring of its water quality is

considered very important. Satellite-based remote sensing (RS) has

become a useful tool for coastal and inland waters

(management/preservation). This research introduces a detailed study on

estimating water quality parameters in Lake Nasser using remote sensing

techniques. Six field missions are used to fulfill the objectives of this

research, missions were conducted during period from 2003 to 2011, and

different water quality parameters were measured at a specific monitoring

location along the lake. MERIS images are used during this research

Match-up images are identified for each field mission. the research has

two main objectives, the first is to validate MERIS Case 2 water quality

processors, such as Case2 Regional (C2R), Eutrophic Lake and boreal

Lake to estimate Total suspended Solids (TSS) and Chlorophyll-a (Chl-a)

in Lake Nasser, some statistical performance measures are used in

validation for each individual mission, mission is grouped by season to

evaluate the processor for different seasons. The second objective is to

create a regression models for optical and non-optical water quality

parameters with different MERIS bands reflectance. Time series for

different water quality parameters are drawn especially for season that

has a validated regression models. In addition to the analysis of cloud

coverage and sun glint over Lake Nasser surface.

6.1. Conclusion

The conclusion of the current research can be summarized as follows:

1- Validation MERIS Case 2 water quality processors

The drawn conclusions resulted from validation of Case 2 water quality

processors, can be pointed out as follows:

114

A- Validation of Case 2 water quality processors to estimate TSS

- The statistical performance measure and overall ranking of TSS

estimation for first, third, and six missions which conducted

during end of flood season indicates that the most appropriate

processor that can be used in this period is eutrophic lake

processor without using ICOL.

- During falling period which represented by fourth mission it is

found that C2R and eutrophic lake processor without ICOL can

be used to estimate TSS values as the overall rank value equal to

one.

- The evaluation of water quality processor to estimate TSS during

rising period, fifth mission, did not indicate a specific processor to

estimate TSS as all processors failed to estimate TSS when its

value is greater than 40 mg/l.

B- Validation of Case 2 water quality processors to estimate Chl-a

The statistical performance measure and overall ranking of Chl-a third

and fourth missions indicates that eutrophic lake processor with using

ICOL process yields good results but it cannot be concluded that the

processor is valid to estimate Chl-a concentration for all year.

2- Regression analysis between MERIS reflectance and water quality

parameters

The Pearson correlation between measured parameters indicates a very

good correlation between optical water quality parameters and other non-

optical parameters which enables us to estimate different parameters

using remote sensing techniques.

The conclusion resulted from the regression between band reflectance

extracted from MERIS images and water quality parameters can be

summarized as follows:

A- Retrieval models for Total Suspended Solids (TSS)

- Different band reflectance can be used to estimate TSS in

different season. reflec_5 or reflec_6 can be used during the end

115

of flood and falling seasons either individually or with other band

such as reflec_1. While reflec_8 has good correlation with TSS

during rising season (flood season), but more repetitive

measurements is need during this period to get more reliable

regression model.

- Validation of regression models in the end of flood season

indicated that that retrieval model for the third mission, conducted

during November 2007, can be used to estimate TSS as it

indicated good comply with the criteria of good model.

B- Retrieval models for Chl-a

- reflec_8 can be used as a predictor of Chl-a during the third

mission, but with low R-squared value (0.431).

- Stepwise regression was failed to get an equation for fourth

mission.

C- Retrieval models for Turbidity indicate very good R-squared value,

but it did not show a correlation with a common band for all models

constructed to estimate turbidity. Validation results indicate that

regression model for first mission is suitable to get turbidity during

this season.

D- Retrieval models for Transparency

- Retrieval models for different missions indicate that reflec_5 can

be used as a predictor for transparency; it can be used alone or

with other band reflectances.

- Validation of retrieval models indicate that model for November

2006 is the suitable model to get transparency during the end of

flood.

E- Retrieval models for Total Phosphorus (TP)

- TP can be correlated to different band reflectance depending on

the season and mission.

- The validation indicates that regression model found for 2006

mission can be used for the end of flood season, and it is noted

that reflec_5 is suitable to predict TP.

116

F- Retrieval models for Silica (SiO2)

- Silica can be correlated to different band reflectance depending on

the season and mission.

- Validation results show that model for 2007 mission is suitable to

estimate silica at the end of flood.

G- Retrieval models for Total dissolved Solids (TDS)

- Results show that reflec_6 is commonly correlated to TDS.

- All regression models for the end of flood show good comply

with the criteria of good model. Model of 2006 mission was select

to estimate TDS.

The analysis and validation results shows that common bands (5 and 6)

are used to predict both optical especially TSS and non-optical water

quality parameters.

3- Time Series

It is concluded from time series curves of the three months (October,

November and December) which represent the end of flood season that

value of TSS concentration started to increase until it reaches its

maximum value in 2005 for the southern part of the lake. The change in

TP is the same as TSS. The change pattern is similar to sin curve. An

opposite behavior is found for both Silica and TDS as they decreased in

2005. Transparency in the southern part of the lake is not showing a

specific pattern and indicates values greater than the measured values.

4- Cloud and sun glint analysis

It concluded that period from September to November has minimum

cloud probability and it considered the best time for coupling between

field measurements and remote sensing. The second period that has low

cloud probability is during January and February.

117

6.2. Recommendations

The following represents the recommendation for the future researches

that can be done by the author or by others:

1- A complete field program, oriented only for water quality

monitoring, should be conducted in different seasons with repetitive

measurements for the same season with preforming some optical

measurements.

2- Creating a bio-optical model that can be used for the case of Lake

Nasser.

3- Creating a new processor that can be used for Lake Nasser.

4- Coupling between water quality modelling and remote sensing in

order to calibrate and predict different water quality parameters.

5- Maximize the benefits of all MERIS images used through this

research by using it in other applications such as evaporation or land

use/land cover maps.

6- Integration between MERIS processing Tool with other used tools to

create a new tool that can be easily used to have a near real time

status for water quality in the lake using remote sensing images.

119

REFERENCES

Abdel-Latif, M. M., & Yacoub, M. (2011). Effect of Change of

Discharges at Dongola Station due to Sedimentation on the Water

Losses from Nasser Lake. Nile Basin Water Science & Engineering

Journal, Vol.4, issue 1. Pp 86-98

Aguib A., Smith S., and EL-Moattassem, M. (1992), Application of

Satellite Remote Sensing Techniques for Monitoring Sedimentation

Patterns in Lake Nasser. The International Conference Nile 2000,

Cairo Egypt.

Ahuja S., and Kumar A. (1996), Evaluation of MESOPUFF-II SOx

Transport and Deposition in the Great Lakes Region, AWMA

Speciality Conference on Atmospheric Deposition to the Great

Lakes, VIP-72, pp. 283-299, Oct. 28-30.

Allan, M.G, Hicks, B.J., and Brabyn, L. (2007). Remote sensing of the

Rotorua lakes for water quality. CBER Contract Report No. 51,

client report prepared for Environment Bay of Plenty. Hamilton,

New Zealand: Centre for Biodiversity and Ecology Research,

Department of Biological Sciences, School of Science and

Engineering, The University of Waikato.

APHA (American Public Health Association); American Water Works

Association (2012); Water Environment Federation. Standard

Methods for the Examination of Water and Wastewater, 22nd ed.;

American Water Works Assn: New York, NY, USA, 2012.

Azab, A. M. (2012). Integrating GIS, Remote Sensing, and Mathematical

Modelling For Surface Water Quality Management In Irrigated

Watersheds. Delft University of Technology. (Ph.D. Thesis).

Brockmann (2013). Remote Sensors. Retrieved April 10, 2014, from

http://www.brockmann-consult.de/cms/web/beam/

120

Bukata R. P. (2005). Satellite Monitoring of Inland and Coastal Water

Quality: Retrospection, Introspection, Future Directions. CRC

Press.

CEOS (2016). Measurement Timelines. Retrieved January 10, 2016,

from http://database.eohandbook.com/timeline/timeline.aspx

Chang, N. B., Imen, S., & Vannah, B. (2015). Remote Sensing for

Monitoring Surface Water Quality Status and Ecosystem State in

Relation to the Nutrient Cycle: A 40-Year Perspective. Critical

Reviews in Environmental Science and Technology, 45(2), 101–

166. http://doi.org/10.1080/10643389.2013.829981

Colwell, R. N., Brewer, W., Landis, G., Langley, P. Morgan, J., Rinker,

J., Robinson, J. M., and Sorem, A. L. (1963). Basic Matter and

Energy Relationships Involved in Remote Reconnaissance.

Photogrammetric Engineering, 29:761-799.

Doerffer, R., and Schiller H. (2008). MERIS Regional Coastal and Lake

Case 2 Water Project - Atmospheric Correction ATBD,GKSS

Research Center, Geesthacht, Version 1.0 18.

Ebaid, H. M. I., and Ismail, S. S. (2010). Lake Nasser evaporation

reduction study. Journal of Advanced Research, 1(4), 315–322.

http://doi.org/http://dx.doi.org/10.1016/j.jare.2010.09.002

El Saadi, A. M., Yousry, M. M., & Jahin, H. S. (2014). Statistical

estimation of Rosetta branch water quality using multi-spectral

data. Water Science, 28(1), 18–30.

El Sammany, M. S. (2002). Design of Lake Nasser Environmental

Monitoring System. Cairo University. (Ph.D. Thesis).

EL-Moattassem, M, and Abdel-Aziz, T.M. (1988), A Study of the

Characteristics of Sediment Transport in Aswan High Dam

Reservoir, Report No. 117, Cairo, Egypt.

ESA (2006). MERIS Product Handbook.

121

Fahmy, H. (2001). Modification and re-calibration of the simulation

model of Lake Nasser. Journal of International Water Research

Association (IWRA) 26: 129–135.

Farag, H. (2011), "Using Earth Observation (EO) Technique for

Monitoring Water Quality in Lake Burullus", Nile Water Science

and Engineering Journal.

Fox, D. G. (1984). Uncertainty in Air Quality Modeling. Bulletin of the

American Meteorological Society, 65(1), 27–36.

Gordon, H., and A. Morel (1983). Remote Assessment of Ocean Color

for Interpretation of Satellite Visible Imagery: A Review. Lecture

Notes on Coastal and Estuarine Studies, Vol. 4, Springer Verlag,

New York, 114 pp.

Hanna S. R. (1989). Confidence limits for air quality model evaluations,

as estimated by bootstap and jackknife resampling methods.

Atmospheric Environment 23, 1385-1398.

Hanna, S. R., Chang, J. C., & Strimaitis, D. G. (1993). Hazardous gas

model evaluation with field observations. Atmospheric

Environment. Part A. General Topics, 27(15), 2265–2285.

Hassan, M. (2013). Evaporation estimation for Lake Nasser based on

remote sensing technology. Ain Shams Engineering Journal, 4(4),

593–604. http://doi.org/10.1016/j.asej.2013.01.004

Hassan, M., Fahmy, A. (2005), Remote sensing as a tool for water quality

modeling of Lake Nasser. Faculty of Engineering Shoubra,

Engineering Research Journal, No.2, pp. 123-128.

Hinkle DE, Wiersma W, Jurs SG. (2003). Applied Statistics for the

Behavioral Sciences. 5th ed. Boston: Houghton Mifflin.

ICOLD (1984). “World Register of Dams,” International Commission of

Large Dams, Paris.

122

Kay, S., Hedley, J. D., & Lavender, S. (2009). Sun glint correction of

high and low spatial resolution images of aquatic scenes: A review

of methods for visible and near-infrared wavelengths. Remote

Sensing, 1(4), 697–730. http://doi.org/10.3390/rs1040697

Kirk, J. T. O. (1994). Light and Photosynthesis in Aquatic Ecosystems,

Cambridge University Press, Cambridge.

Koponen S., Ruiz Verdu A., Heege T., Heblinski J., Sorensen K., Kallio

K., et al., (2008). Development of MERIS lake water algorithms:

Validation report, Helsinki University of Technology Version 1.01,

Helsinki, Finland.

Kumar A., Luo J., and Bennett G. (1993). Statistical Evaluation of Lower

Flammability Distance (LFD) using Four Hazardous Release

Models, Process Safety Progress, 12(1), pp. 1-11, 1993.

Liliesand, T.M. and Kiefer, R. (1993). Remote Sensing and Image

Interpretation. Third Edition John Villey, New York.

Liqin Qu (2014). Remote Sensing Suspended Sediment Concentration in

the Yellow River. University of Connecticut. ( Ph.D. Thesis).

Miloud Chikr El-Mezouar (2012). Fusion d'images en télédétection

satellitaire. Environmental Engineering. INSA de Rennes;

Université Djillali Liabes de Sidi Bel Abbès.

Mitsch, W.J. (1973). Remote sensing of water quality-a state of the art

report. Florida Water Resources Center Report No. 21, Gainesville,

FL. 14 pp.

Mobley, C.D. (1995). The Optical Properties of Water. In: Bass, M.

(Ed.). Handbook of Optics, I, McGraw-Hill, New York, NY.

Morel, A. (1980). In-water and remote measurements of ocean color.

Boundary-Layer Meteorology, 18:177-201.

Morel, A. and Gordon, H.R. (1980); Report of the working group on

water color; Boundary layer meteorology, 18:p 343-355.

123

Morel, A. and Maritorena, S. (2001). Bio-optical properties of oceanic

waters: a reappraisal. Journal of Geophysical Research,

106(C4):7163-7180.

Morel, A. and Prieur L. (1977). Analysis of variations in ocean color.

Limnol. Oceanogr., 22(4), 709-722.

Mostafa, M. M. and Soussa, H. K. (2006). Monitoring of Lake Nasser

using remote sensing and GIS techniques. (ISPRS Mid-term

Symposium Proceeding. May 2006, Enschede).

NASA. (2015). Remote Sensors. Retrieved January 10, 2015, from

https://earthdata.nasa.gov/user-resources/remote-sensors

Patryla L., Galeriua D., (2011). Statistical Performances Measures –

Models Comparison [PowerPoint Slides]. Retrieved from

https://www-ns.iaea.org/downloads/rw/projects/emras/emras-

two/first-technical-meeting/sixth-working-group-meeting/working-

group-presentations/workgroup-7-presentations/presentation-6th-

wg7-statistical-performances.pdf.

Preisendorfer, R. W. (1976). Hydrologic optics. Washington: U.S.

Department of Commerce.

Rahman, H. and Dedieu G., (1994). SMAC: A simplified method for the

atmospheric of satellite measurements in the solar spectrum.

International Journal of Remote Sensing, 15, 123-143.

Ritchie, J. C. and Cooper, M. C. (1991) “An Algorithm for Estimating

Surface Suspended Sediment Concentrations with Landsat MSS

Digital Data” Water Resources Bulletin, Vol. 27, No. 3, pp.373-

379.

Sadek, M. F., Shahin, M. M. & Stigter, C. J. (1997). Evaporation from

the reservoir of the High Aswan Dam, Egypt: a new comparison of

relevant methods with limited data. Journal of Theoretical and

Applied Climatology l56: 57–66.

124

Said, R. (1993). The river Nile: geology, hydrology, and utilization.

Pergamon, Oxford, pp. 320.

Santer, R. and Zagolski, F. (2009). Improved contrast between ocean and

land- MERIS algorithm theoretical basis document,Version 1.1, 6

January 2009. Universite du Littoral Cote d’Opale, Wimereux,

France.

Shafik, N. M. (2004). Study of Evaporation Losses in Lake Nasser. Ain

Shams University. (Ph.D. Thesis).

Shahin, M. (2002). Hydrology and water resources of Africa. Kluwer

Academic Publishers, Dordrecht ; London, pp. 659.

Springuel, I. and Ali, O. (2005). The River Nile Basin. In: L.H. Fraser

and P.A. Keddy (Editors), The world's largest wetlands : ecology

and conservation. Cambridge University Press, Cambridge ; New

York.

Tantirimudalige, M. (2002). Basics of Remote Sensing. Lecture Notes.

Open University of Sri Lanka.

Yang, X.Q. (2003). Manual on Sediment Management and Measurement.

World Meteorological Organization, Operational Hydrology Report

No. 47, WMO-No. 948, Secretariat of the World Meteorological

Organization, Geneva.

Zaghloul S. S., Pacini N., Schwaiger K., Henry De Villeneuve P. (2012).

Towards a Lake Nasser management plan: results of a pilot test on

integrated water resources management. International water

technology journal, Vol. 1, pp. 249-258.

125

APPENDIX (A)

MERIS IMAGES CHARACTERISTICS AND

WATER QUALITY ANALYSIS

A.1. MERIS image characteristics

The Medium Resolution Imaging Spectrometer (MERIS) is one of the

main instruments on board Envisat platform lunched by the European

Space Agency (ESA)'s in March 2002. An array of nine instruments,

used in earth observation, was carried out by Envisat to collect

information about the Earth (land, water, ice, and atmosphere) using a

variety of measurement principles. A tenth instrument, DORIS, provided

guidance and control, see Figure ( A-1).

MESIS is a passive sensor that depends on sun solar radiation as its

source of energy. The observation of Earth using MERIS images is

performed simultaneously in 15 spectral bands (ESA, 2006).

Figure (‎A-1) Envisat Satellite Configuration and Payload

Instruments (ESA, 2006)

Name Description

ASAR Advanced Synthetic Aperture

Radar

MERIS Medium-Spectral Resolution

Imaging Spectrometer

AATSR Advanced Along Track Scanning

Radiometer

RA-2 Radar Altimeter

MWR Microwave Radiometer

GOMOS Global Ozone Monitoring by

Occultation of Stars

MIPAS Michelson Interferometer for

Passive Atmospheric Sounding

SCIAMACHY

Scanning Imaging Absorption

Spectrometer for Atmospheric

Chartography

DORIS Doppler Orbitography and Radio-

positioning Integrated by Satellite

LRR Laser Retro-Reflector

126

A.1.1. MERIS product processing levels

The Envisat products are delivered to researches in three processing

levels; Level 1B which contains data about Top Of Atmosphere (TOA)

radiance, Level 2 are produced by processing from Level 1B images to

get geophysical measurements, and Level 3 products which are synthesis

of more than one MERIS products (and possibly external data) to display

geophysical measurements for a time period.

The output or products of MERIS are image packages of several

parameters such as radiances at the top of the atmosphere (TOA) in

different bands. Different products are created according to the

parameters plotted and their resolution. In Table ( A-1), all the MERIS

products are listed with their respective ID’s and the working mode of the

instrument.

A.1.2. Resolutions

Spatial, Spectral and Temporal resolution of MERIS image are described

below.

- Spatial Resolution

MERIS was provided in three main spatial resolutions; Full-Resolution

(FR), Reduced Resolution (RR) and Low Resolution (LR). Pixel in an FR

image represents an area of 260 m × 290 m and in an RR image an area

of 1,040 m × 1,160 m, and in LR an area of 4,160 m × 4,640 m.

The instrument always records information with its full resolution. The

Reduced Resolution then can be generated using onboard averaging,

while LR images can be generated by averaging RR data in a ground

processor.

- Spectral Resolution

The MERIS sensor acquire an image contains fifteen (15) spectral bands

within the spectral range from 390 nm to 1040 nm, this range lies in the

visible and near infrared wavelengths of the electromagnetic spectrum.,

127

Table ( A-2) illustrates the spectral characteristics of each band and the

application used for each band.

Table (‎A-1) MERIS products description (ESA, 2006).

Product ID Product Name Application

MER_RR__0P Reduced Resolution Level 0

Not generally available to

users

MER_FR__0P Full Resolution Level 0

MER_CA__0P Calibration Level 0

MER_RV__0P Reduced Field of View Level 0

MER_RR__1P Reduced Resolution Level 1 Serve as the basis for level

2 processing

MER_FR__1P Full Resolution Level 1

Application in atmospheric

modeling, land use

monitoring, ocean color

monitoring, vegetation

indices, and others

MER_RR__2P Reduced Resolution

Geophysical

Ocean, land or atmosphere

characterization at 1040 by

1160 m pixel spatial

resolution

MER_FR__2P Full Resolution Geophysical

Climatology, meteorology,

environmental monitoring,

etc.

MER_LRC_2P

Extracted Cloud Thickness and

Water Vapour for

Meteorological Users

Intended only for

meteorological

applications

MER_RRC_2P Extracted Cloud Thickness and

Water Vapour

Intended for

meteorological

applications

MER_RRV_2P Extracted Vegetation Indices Intended for near real time

land monitoring

MER_RR__BP Browse Product

Support queries to a

MERIS archive for land,

sea, ice or cloud features,

to be viewed from a

remote user terminal

128

Table (‎A-2) MERIS spectral bands and applications (ESA, 2006).

No

.

Band center

(nm)

Band width

(nm) Applications

1 412.5 10 Yellow substance and detrital pigments

2 442.5 10 Chlorophyll absorption maximum

3 490 10 Chlorophyll and other pigments

4 510 10 Suspended sediment, red tides

5 560 10 Chlorophyll absorption minimum

6 620 10 Suspended sediment

7 665 10 Chlorophyll absorption & fluorescence

reference

8 681.25 7.5 Chlorophyll fluorescence peak

9 708.75 10 Fluorescence reference, atmosphere

corrections

10 753.75 7.5 Vegetation, cloud, O2 absoption band

reference

11 760.625 3.75 O2 R- branch absorption band

12 778.75 15 Atmosphere corrections

13 865 20 Atmosphere corrections

14 885 10 Vegetation, water vapour reference

15 900 10 Water vapour

- Temporal Resolution

MERIS field-of-view angle is 68.5° and its swath width is about 1150

km, which allows covering the Earth every three days, this can be useful

to track changes in oceanographic, land, and in atmospheric

investigations. Figure ( A-2) shows the global coverage of MERIS in both

winter and summer.

129

Figure (‎A-2) MERIS Global coverage (ESA, 2006)

A.1.3. MERIS Case 2 water quality processors

Estimating water quality parameters, as discussed before, is one of the

main objectives of MERIS mission. So that, many water processors were

developed for that reason. All of these processors were developed as a

plugins for BEAM software.

Algorithms developed to process MERIS images and are reliable far from

land and on open oceans (These type of water are called Case 1 Waters).

However, in coastal and lake waters (Case 2 Waters) where high

concentrations of reflective particles can be found in the water surface,

traditional algorithms fail. Alternative and more sophisticated algorithms

have been developed to deal with these complications; three of these

processors will be described, analyzed and compared in a Case 2 Waters

scenario:

- Coastal Case 2 Regional Water Processor (C2R).

- Boreal Lakes Water Processor.

- Eutrophic Lakes Water Processor.

The three have been developed by the Gesellschaft fur

Kernenergieverwertung in Schiffbau und Schiffahrt mbH (GKSS). All

processors use neural networks and are based on the architecture of the

C2R processor developed by (Doerffer R., 2008)

130

The algorithm of these processors can be conceptually divided in two

main parts: the atmospheric correction algorithm and the water algorithm.

The first part, determines the surface reflectance spectrum RLw() using

the top of atmosphere radiance spectrum RLtoa() acquired by MERIS;

the second part, uses the water leaving radiances as input and computes

different products that provide information about the quality of the water.

A.2. Water Quality Analysis

A.2.1. Water sampling

At any site water samples were taken from five depths (50) cm from

surface, 25%, 50%, 65% and 80% from total depth) using water sampler

of different depths. Field measurements were determined in all water

samples for all sites.

A.2.2. Preservation

Sample bottles should be cleaned before use by soaking it in detergent,

followed by rinsing with tap water several times until free of detergent,

rinsed with 5% nitric acid and then thoroughly with distilled-deionized

water. Sample bottles were cleaned and sterilised according to the

standard procedures for microbiological analysis.

The collected water samples were preserved in glass and plastic clean

bottles for the parameter needed to be measured and preserved in icebox

to retard any chemical and biological changes.

A.2.3. Types of analysis

Field measurements (on site analysis)

Certain parameters can vary while transporting samples to the laboratory

and it is always required to determine the following parameters in the

field (pH, temperature, dissolved oxygen, turbidity and conductivity) by

using portable field equipment moreover water depths, Transparency,

flow velocity and discharges have been also measured.

131

A.2.4. Laboratory measurements

Sensitive parameters such as, nutrients (orthophosphate, total

phosphorous, nitrate, nitrite and ammonia), fecal coliform (FC), Silicate,

Chlorophyll a, alkalinity, sulphate, major ions, surfactants, total

suspended solids (TSS), total dissolved solids (TDS) and total hardness

were analyzed within 24 hours of sampling. The other parameters

including sodium, potassium and heavy metals were measured within one

month. Physicochemical analyses were performed according to the

standard methods for examination of water and wastewater suggested by

American Public Health Association (APHA, 2012).

A.2.5. Reagents

The standards, reagents, solvents, indicators and buffer solutions were

prepared following recommended procedures (APHA, 2012). All the

chemicals used in the analysis were of analytical grade.

A.2.6. Equipment

Field equipment

- Different depths water sampler.

- Secchi Disk.

- Portable conductivity meter, used for measurement of

conductivity, salinity and temperature (WTW, LF 197).

- Portable pH meter with electrode for measurement of pH and

temperature (WTW, pH 197).

- Portable dissolved oxygen meter with oxygen probe used for

measurement of dissolved oxygen (WTW, Oxi 197).

- Portable spectrophotometer (DR 2400 HACH) used for field

measurements of phosphorous, COD, nitrite, detergents.

- YSI 6-Series Multi-parameter Water Quality Monitors 650 MDS

used for measuring pH, DO, Conductivity, Temperature and

Depth.

132

A.2.7. Laboratory equipment

- Orion research ionanalyzer EA 940 for measuring cations, anions

and gases by using specific ion selective electrode.

- In-Spectra Analyzer used for measuring TOC, Nitrate, TSS and

Surfactant.

A.2.8. General equipment

- Incubator for Fecal test.

- Automatic autoclave for sterilization and digestion.

- Hot plates with magnetic stirrer.

- Electronic analytical balance.

- Oven.

133

APPENDIX (B)

MERIS PROCESSING TOOL (MPT) AND

EXCEL CUSTOM FUNCTIONS

B.1. Introduction

This appendix contains the programming details and visual basic code of

MERIS Processing tool, which create in order to automate the MERIS

images processing to save time need for manual processing. Also, it

contains visual basic code for excel custom function created to calculate

statistical performance measures used to evaluate water quality

processors and regression models.

B.2. MERIS PROCESSING TOOL (MPT)

This tool was developed in Excel using Visual Basic for Application

(VBA) to automate processing of MERIS images, to help in processing

of large number of MERIS images, to save time needed for processing

and to reduce the mistakes which may found in the manual process. The

advantage of using VBA in excel is that Excel is found in all computers,

so there is no need to have additional programs; also it is easy to convert

it to a stand-alone program if needed.

B.2.1. Main Idea of MPT

The main idea of MPT is based on using the BEAM Graph Processing

Framework (GPF), a processor is as a software module which uses one or

more input product to create an output product using a set of processing

parameters. The GPF allows constructing directed acyclic graphs (DAG)

of processing nodes. A node in the graph refers to a GPF operator, which

makes a certain algorithm to be executed. The main role of the node is to

configure the operation by identifying the operator’s source nodes and

providing values for different processing parameters. Figure ( B-1) shows

a sample of a processing graph comprising five nodes.

134

GPF operators can be executed in batch mode using Graph processing

Tool (GPT), the operators can be used stand-alone or combined as a

directed acyclic graph (DAG). The concept of using GPT is to use the

BEAM from command line; Processing graphs are represented using

XML. GPT command can be used as a single command or using a set of

commands using batch file to handle more than one processing operation

as well as handling more than one file. The role of MPT is to write and

runs the final batch and XML files in easy way.

Figure (‎B-1) Sample of A processing graph

B.2.2. MPT description

The flow chart of MPT programming is shown in Figure ( B-2), a detailed

programming code for the tool was represented below. MPT was found

in MS Excel workbook contains the VBA Code for MPT, Once you open

this workbook was opened, the shown window in Figure ( B-3a) will

appears, it is a simple window and easy to use. The first step in using

MPT is to select the directory in which MERIS image files were found,

the default output directory is a directory in the same location of the MPT

workbook. Also it is important to select the type of MERIS image format

even it in Envisat MERIS (*.N1) format or BEAM-DIMAP product files

135

(*.dim) and to select the format of final output image a description of the

processing operations for MERIS images are:

Figure (‎B-2) Flow chart of MPT programming

a. Subsetting of the MERIS products in order to avoid processing

area larger than required to decrease time needed for processing,

the boundaries of the region of interest (ROI) were defined in

MPT by two ways, by entering the coordinates of lower left

corner and upper right corner coordinates of ROI bounding box or

by loading a Google Earth (*.KML) file contains a rectangular

polygon that defines the ROI. If the image is already subsetted to

the ROI, this step can be skipped. MERIS image can be corrected

against adjacency effect using ICOL process at this step, as there

are two option, to preform subset with ICOL processor or to

preform ICOL processor only as show in Figure ( B-3b)

b. MPT have an option to reproject the final BEAM-DIMAP product

files, to geo-locate the final product in its original coordinates.

136

c. After subsetting and/or ICOL processor operation were selected,

there are two processes to preform; water processor operation or

to band math operation.

d. Validating the use of a three water processors (Case 2 Regional,

Boreal Lakes, Eutrophic Lakes) for use in Lake Nasser is one of

the main objectives this research. So that, MPT give the ability to

select one or all of water processors to be applied to MERIS

image in the same batch file. A parameter XML file, which stores

the settings, is need for each processor, and it can easily obtained

directly from BEAM, MPT take this parameter file and then

generates another output XML graph file that used in processor

operation.

e. The other process can be performed in MPT is band math, in

order to perform a mathematical operation of the level 1B MERIS

image, a simplified method for atmospheric correction (SMAC),

should be performed to the image to calculate the water leaving

reflectance for 15 bands. A request XML file contains the

parameters used in SMAC process is required. SMAC processor

can be skipped in case the images are already atmospherically

corrected. Applying the band math operations can be done by

check “Apply Band Math” box, a window will popups to enter the

required information, Figure ( B-3c) for each mathematical

expression, an XML graph file containing the all added

expression will be written to be used later by MPT.

f. Cloud probability processor can be applied using MPT. This

allows us to classify each pixel as Cloudy, Cloud uncertain or

Cloud free.

137

a.

b.

c.

Figure (‎B-3) MPT interface

138

B.2.3. MPT Code

This section illustrates the programming code for MTP, the code will be

shown for every object found in the tool and categorized as follows:

- Code for Main MPT window.

- Code for Subset region definition and ICOL processor window.

- Code for Band Math window.

The shown code is for buttons, check boxes and option buttons.

1- Code for Main MPT window

Dim CHL_CPfile

Dim TSM_CPfile

Dim c2r_xml

Dim eut_xml

Dim bor_xml

Dim FUBMath_xml

Dim SmacRfile

Private Sub bmath_Change()

If bmath.Value = -1 Then

fndloc.Form_bm.Enabled = True

fndloc.smac.Enabled = True

ElseIf bmath.Value = 0 Then

fndloc.Form_bm.Enabled = False

fndloc.Form_bm.Value = 0

fndloc.smac.Enabled = False

fndloc.smac.Value = 0

End If

End Sub

Private Sub c2r_Change()

Dim P_file_cont As String * 6000

quot = """"

Dim PF_Typ As String * 12

If c2r.Value = -1 Then

5:

fndloc.c2rer = ""

P_file = Application.GetOpenFilename("Case 2

Regional Parameters (*.xml),*.xml")

Open P_file For Binary As #2

Get #2, 1, P_file_cont

Get #2, 1, PF_Typ

Close #2

If P_file = flase Then

c2r.Value = 0

GoTo 20

End If

139

If PF_Typ <> "<parameters>" Then

MsgBox("Please Select Proper Parameter file")

fndloc.c2rer = ""

GoTo 5

End If

c2r_xml = Application.ActiveWorkbook.Path &

"\Processing Parameters\" & "Case2R.xml"

Open c2r_xml For Output As #1

Print #1, "<graph id="; quot; "Case2RGraph"; quot; ">"

Print #1, " <version>1.0</version>"

Print #1, " <node id="; quot; "case2r"; quot; ">"

Print #1, "

<operator>Meris.Case2Regional</operator>"

Print #1, " <sources>"

Print #1, " <source>${source}</source>"

Print #1, " </sources>"

Print #1, Left(P_file_cont, FileLen(P_file))

Print #1, " </node>"

If fndloc.reproj.Value = -1 Then

Print #1, " <node id="; """reprojNode"""; ">"

Print #1, " <operator>Reproject</operator>"


Print #1, " <source>case2r</source>"


Print #1, " <parameters>"

Print #1, " <crs>EPSG:63266405</crs>"

Print #1, "

<resampling>Nearest</resampling>"

Print #1, "

<includeTiePointGrids>false</includeTiePointGrids>"

Print #1, " </parameters>"


End If

Print #1, "</graph>"

Close #1

ElseIf c2r.Value = 0 Then

End If

20:

End Sub

Private Sub Clouds_Click()

Dim objFSO, Infolder, fold, f1, fc

quot = """"

Dim SMAC_F_cont As String * 6000

If Clouds.Value = -1 Then

CloudsRfile = Application.GetOpenFilename("Clouds

Request File (*.xml),*.xml")

Const ForReading = 1

Const ForWriting = 2

objFSO = CreateObject("Scripting.FileSystemObject")

140

Infolder = Left(fndloc.SDname, Len(fndloc.SDname) -

1)

fold = objFSO.GetFolder(Infolder)

fc = fold.Files

For Each f1 In fc

If f1.Type = "BEAM-DIMAP file, BEAM's standard

EO data format" Or f1.Type = "Envisat data product in N1

format" Then

If f1.Type = "BEAM-DIMAP file, BEAM's

standard EO data format" Then

exlen = 4

extName = ".dim"

ElseIf f1.Type = "Envisat data product in

N1 format" Then

exlen = 3

extName = ".N1"

End If

In_Clouds_File =

objFSO.OpenTextFile(CloudsRfile, ForReading, True)

Out_Clouds_File =

objFSO.OpenTextFile(Application.ActiveWorkbook.Path &

"\Processing Parameters\Clouds\" & Left(f1.Name,

Len(f1.Name) - exlen) & ".xml", ForWriting, True)

Do While Not In_Clouds_File.AtEndofStream

myLine = In_Clouds_File.ReadLine

If Subset.Value = -1 Then

InValue = " <InputProduct

file=" & quot & Left(fndloc.TDname, Len(fndloc.TDname) - 1)

& "\Subset\" & Left(f1.Name, Len(f1.Name) - exlen) &

"_Subset.dim" & quot & " />"

ElseIf Subset.Value = 0 Then


file=" & quot & Left(f1.Path, Len(f1.Path) - exlen) &

extName & quot & " />"

End If

OutValue = " <OutputProduct


& "\Clouds\" & Left(f1.Name, Len(f1.Name) - exlen) &

"_Clouds.dim" & quot & " format=" & quot & "BEAM-DIMAP" &

quot & " />"

If InStr(myLine, "InputProduct") Then

myLine = InValue & whatever

ElseIf InStr(myLine, "OutputProduct")

Then

myLine = OutValue & whatever

End If

Out_Clouds_File.WriteLine(myLine)

Loop

End If

141

Next

End If

End Sub

Private Sub eut_l_Change()


quot = """"


If eut_l.Value = -1 Then

6:

fndloc.c2rer = ""

P_file = Application.GetOpenFilename("Eutrophic

lakes Parameters (*.xml),*.xml")



Get #2, 1, PF_Typ

Close #2


eut_l.Value = 0

GoTo 21

End If



fndloc.c2rer = ""

GoTo 6

End If

eut_xml = Application.ActiveWorkbook.Path &

"\Processing Parameters\" & "Eut_Lakes.xml"

Open eut_xml For Output As #1

Print #1, "<graph id="; quot; "Eut_Lakes"; quot; ">"


Print #1, " <node id="; quot; "Eutrophic"; quot; ">"

Print #1, " <operator>Meris.Lakes</operator>"










Print #1, " <source>Eutrophic</source>"




Print #1, "


Print #1, "


142



End If


Close #1

ElseIf c2r.Value = 0 Then

End If

21:

End Sub

Private Sub bor_l_Change()


quot = """"


If bor_l.Value = -1 Then

7:

fndloc.c2rer = ""

P_file = Application.GetOpenFilename("Boreal lakes

Parameters (*.xml),*.xml")



Get #2, 1, PF_Typ

Close #2


bor_l.Value = 0

GoTo 21

End If



fndloc.c2rer = ""

GoTo 7

End If

bor_xml = Application.ActiveWorkbook.Path &

"\Processing Parameters\" & "Bor_Lakes.xml"

Open bor_xml For Output As #1

Print #1, "<graph id="; quot; "Eut_Lakes"; quot; ">"


Print #1, " <node id="; quot; "Eutrophic"; quot; ">"

Print #1, " <operator>Meris.Lakes</operator>"










Print #1, " <source>Eutrophic</source>"


143



Print #1, "


Print #1, "




End If


Close #1

ElseIf bor_l.Value = 0 Then

End If

21:

End Sub

Private Sub exchl_Change()

If exchl.Value = -1 And out_typ.Value <> "GeoTIFF

product format (*.tif)" Then

CHL_CPfile = Application.GetOpenFilename("CHL Color

Palette (*.cpd),*.cpd")

End If

End Sub

Private Sub extsm_Click()

If extsm.Value = -1 And out_typ.Value <> "GeoTIFF

product format (*.tif)" Then

TSM_CPfile = Application.GetOpenFilename("TSM Color

Palette (*.cpd),*.cpd")

End If

End Sub

Private Sub fndlocext_Click()

fndloc.Hide()

End Sub

Private Sub Form_bm_Click()

If Form_bm.Value = -1 Then

FBmath.Show()

ElseIf Form_bm.Value = 0 Then

End If

End Sub

Private Sub fub_Change()


quot = """"

Dim FUB_F_cont As String * 6000

If fub.Value = -1 Then

8:

fubRfile = Application.GetOpenFilename("FUB\WEW


If fubRfile = flase Then

fub.Value = 0

GoTo 22

144

End If





1)


fc = fold.Files

For Each f1 In fc



format" Then



exlen = 4

extName = ".dim"


N1 format" Then

exlen = 3

extName = ".N1"

End If

In_FUB_File = objFSO.OpenTextFile(fubRfile,

ForReading, True)

Out_FUB_File =


"\Processing Parameters\FUB\" & Left(f1.Name, Len(f1.Name)

- exlen) & ".xml", ForWriting, True)

Do While Not In_FUB_File.AtEndofStream

myLine = In_FUB_File.ReadLine










End If



& "\FUB\001 Main\" & Left(f1.Name, Len(f1.Name) - exlen) &

"_FUB.dim" & quot & " format=" & quot & "BEAM-DIMAP" & quot

& " />"




Then


145

End If

Out_FUB_File.WriteLine(myLine)

Loop

End If

Next

FUBMath_xml = Application.ActiveWorkbook.Path &

"\Processing Parameters\FUB\FUBExpBands.xml"

V_Express = "(reflec_2 > 0) and not

(l1_flags.INVALID or l1_flags.BRIGHT or bright)"

If Regdef.icol.Value = -1 Then

V_Express = "(reflec_2 > 0) and not

(l1_flags.INVALID or l1_flags.BRIGHT)"

End If

Open FUBMath_xml For Output As #1

Print #1, "<graph id="; """BandmathId"""; ">"


Print #1, " <node id="; """BandmathNode"""; ">"

Print #1, " <operator>BandMaths</operator>"





Print #1, " <targetBands>"

Print #1, " <targetBand>"

Print #1, " <name>"; "algal_2";

"</name>"

Print #1, " <expression>";

V_Express; " ? (algal_2) : NaN</expression>"

Print #1, " <description>";

"Chlorophyll 2 content"; "</description>"

Print #1, " <type>float32</type>"

Print #1, " <unit>";

"log10(g/m^3)"; "</unit>"

Print #1, "

<noDataValue>0</noDataValue>"

Print #1, " </targetBand>"


Print #1, " <name>"; "yellow_subs";

"</name>"


V_Express; " ? (yellow_subs) : NaN</expression>"

Print #1, " <description>"; "Yellow

substance"; "</description>"



"log10(g/m^3)"; "</unit>"

Print #1, "



146


Print #1, " <name>"; "total_susp";

"</name>"


V_Express; " ? (total_susp) : NaN</expression>"

Print #1, " <description>"; "Total

suspended matter"; "</description>"



"log10(g/m^3)"; "</unit>"

Print #1, "



Print #1, " </targetBands>"







Print #1, "

<source>BandmathNode</source>"




Print #1, "


Print #1, "




End If


Close #1

End If

22:

End Sub

Private Sub lakes_Change()

If lakes.Value = -1 Then

fndloc.c2r.Enabled = True

fndloc.eut_l.Enabled = True

fndloc.bor_l.Enabled = True

fndloc.fub.Enabled = True

fndloc.extsm.Enabled = True

fndloc.exchl.Enabled = True

ElseIf lakes.Value = 0 Then

fndloc.c2r.Enabled = False

fndloc.c2r.Value = 0

fndloc.eut_l.Enabled = False

147

fndloc.eut_l.Value = 0

fndloc.bor_l.Enabled = False

fndloc.bor_l.Value = 0

fndloc.fub.Enabled = False

fndloc.fub.Value = 0

fndloc.extsm.Enabled = False

fndloc.extsm.Value = 0

fndloc.exchl.Enabled = False

fndloc.exchl.Value = 0

End If

End Sub

Private Sub Run_Click()

Dim Ex_num As Integer

Mbat = Application.ActiveWorkbook.Path & "\Processing

Parameters\" & "Main Batch.bat"

quot = """"

reg_xml = Application.ActiveWorkbook.Path &

"\Processing Parameters\" & Regdef.XML_fn & ".xml"

bmath_xml = Application.ActiveWorkbook.Path &

"\Processing Parameters\Band Math.xml"

Open Mbat For Output As #1

Print #1, "@echo off"

Print #1, "setlocal ENABLEDELAYEDEXPANSION"

Print #1,

"::::::::::::::::::::::::::::::::::::::::::::"

Print #1, ":: User Configuration"

Print #1, "set gptPath=" & """C:\Program Files\beam-

4.10.3\bin\gpt.bat"""

If extsm.Value = -1 Or exchl.Value = -1 Or

Form_bm.Value = -1 Then

Print #1, "set PconvPath=" & """C:\Program

Files\beam-4.10.3\bin\pconvert.bat"""

If extsm.Value = -1 Then

Print #1, "set TSM_CP="; TSM_CPfile

'D:\Phd\GPT\MERIS Processing Tool\Input

Parameters\TSM_colours.cpd"

End If

If exchl.Value = -1 Then

Print #1, "set CHL_CP="; CHL_CPfile

'D:\Phd\GPT\MERIS Processing Tool\Input

Parameters\TSM_colours.cpd"

End If

End If

Print #1, "set sourceDirectory=" & quot &

Left(fndloc.SDname, Len(fndloc.SDname) - 1) & quot


Print #1,

"::::::::::::::::::::::::::::::::::::::::::::"

Print #1, ":: Subset commandline handling"

148

Print #1, "rem first parameter is a path to the

subset graph xml"

Print #1, "set subset_graphXml=" & quot & reg_xml &

quot

Print #1, "set sourceDirectory=" & quot &

Left(fndloc.SDname, Len(fndloc.SDname) - 1) & quot

Print #1, "set subset_TDirectory=" &

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\Subset"

Print #1, "set subset_Suffix=Subset"

End If

If lakes.Value = -1 Then

Print #1,

"::::::::::::::::::::::::::::::::::::::::::::"

Print #1, ":: Water Processors commandline

handling"


Print #1, "set C2R_graphXml=" & quot & c2r_xml

& quot

Print #1, "set C2R_TDirectory=" &

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\Case 2

Regional"


'Print #1, "md "; quot; "!TSM_TDirectory!";

quot

ElseIf exchl.Value = 0 Then

'Print #1, "md "; quot; "!CHL_TDirectory!";

quot

End If

Print #1, "set C2R_Suffix=C2R"

End If


Print #1, "set EUT_graphXml=" & quot & eut_xml

& quot

Print #1, "set EUT_TDirectory=" &

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\Eutrophic

lakes"

Print #1, "set EUT_Suffix=EUT"

End If


Print #1, "set BOR_graphXml=" & quot & bor_xml

& quot

Print #1, "set BOR_TDirectory=" &

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\Boreal

lakes"

Print #1, "set BOR_Suffix=BOR"

End If


Print #1, "set FUB_Path=" & """C:\Program

Files\beam-4.10.3\bin\fub.bat"""

149

Print #1, "set FUB_XMLDirectory=";

Application.ActiveWorkbook.Path; "\Processing

Parameters\FUB"

Print #1, "set FUB_TDirectory=";

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\FUB"

Print #1, "set FUB_BM_XML=" & quot &

FUBMath_xml & quot

End If

End If


Print #1,

"::::::::::::::::::::::::::::::::::::::::::::"

Print #1, ":: Band Math and SMAC commandline

handling"

If smac.Value = -1 Then

Print #1, "set SMAC_Path=" & """C:\Program

Files\beam-4.10.3\bin\meris-smac.bat"""

Print #1, "set SMAC_XMLDirectory=";


Parameters\SMAC"

Print #1, "set SMAC_TDirectory=";

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\SMAC"

End If


Print #1, "set Bmath_Xml=" & quot & bmath_xml &

quot

Print #1, "set Bmath_TDirectory=" &

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\Band Math"

Print #1, "set Bmath_Suffix=Bmath"

End If

End If


Print #1,

"::::::::::::::::::::::::::::::::::::::::::::"

Print #1, ":: Clouds Propability commandline

handling"

Print #1, "set Clouds_Path=" & """C:\Program

Files\beam-4.10.3\bin\meris-cloud.bat"""

Print #1, "set Clouds_XMLDirectory=";


Parameters\Clouds"

Print #1, "set Clouds_TDirectory=";

Left(fndloc.TDname, Len(fndloc.TDname) - 1) & "\Clouds"

End If

Print #1,

"::::::::::::::::::::::::::::::::::::::::::::"

Print #1, ":: Main processing"

Img_Intyp = typ.Value

If typ.Value = "ENVISAT MERIS (*.N1)" Then

150

Img_Intyp = "N1"

ElseIf typ.Value = "BEAM-DIMAP product files (*.dim)"

Then

Img_Intyp = "dim"

Else

If MsgBox("MERIS Input file type not selected,

Select (*.N1) as a default value", vbOKCancel) = vbOK Then

Img_Intyp = "N1"

End If

End If

If out_typ.Value = "GeoTIFF product format (*.tif)"

Then

out_format = "tifp"

TSM_Color_p = ""

CHL_Color_p = ""

ElseIf out_typ.Value = "JPEG image format (*.jpg)" Then

out_format = "jpg"

TSM_Color_p = " -c " & """!TSM_CP!"""

CHL_Color_p = " -c " & """!CHL_CP!"""

ElseIf out_typ.Value = "Portal Network Graphics

(*.png)" Then

out_format = "png"



ElseIf out_typ.Value = "Geotif Image Format (*.tif)"

Then

out_format = "tif"



ElseIf out_typ.Value = "Microsoft Bitmap image (*.bmp)"

Then

out_format = "bmp"



End If

Print #1, "for %%F in (%sourceDirectory%\*." &

Img_Intyp & ") do ("


Print #1, "set subset_SFile=%%~fF"

Print #1, "set

subset_TFile=%subset_TDirectory%\%%~nF_%subset_Suffix%.dim"

Print #1, "set subset_procCmd=%gptPath%

%subset_graphXml% -Ssource=" & """!subset_SFile!""" & " -t

" & """!subset_TFile!"""

Print #1, "echo !subset_procCmd!"

Print #1, "call !subset_procCmd!"

End If



151

Print #1, "set C2R_SFile=!subset_TFile!"


Print #1, "set C2R_SFile=%%~fF"

End If

Print #1, "set C2R_TFile=%C2R_TDirectory%\001

Main\%%~nF_%C2R_Suffix%.dim"

Print #1, "set C2R_procCmd=%gptPath% %C2R_graphXml%

-Ssource=" & """!C2R_SFile!""" & " -t " & """!C2R_TFile!"""

Print #1, "echo !C2R_procCmd!"

Print #1, "call !C2R_procCmd!"


Print #1, "set

TSM_TDirectory=!C2R_TDirectory!\TSM"

Print #1, "md "; quot; "!TSM_TDirectory!"; quot

Print #1, "set TSMPconv_Cmd=%pconvPath% -f ";

out_format; " -b 35 " & """!C2R_TFile!""" & " -o " &

"""!TSM_TDirectory!""" & TSM_Color_p

Print #1, "echo !TSMPconv_Cmd!"

Print #1, "call !TSMPconv_Cmd!"

End If


Print #1, "set

CHL_TDirectory=!C2R_TDirectory!\CHL"

Print #1, "md "; quot; "!CHL_TDirectory!"; quot

Print #1, "set CHLPconv_Cmd=%pconvPath% -f ";

out_format; " -b 36 " & """!C2R_TFile!""" & " -o " &

"""!CHL_TDirectory!""" & CHL_Color_p

Print #1, "echo !CHLPconv_Cmd!"

Print #1, "call !CHLPconv_Cmd!"

End If

End If



Print #1, "set EUT_SFile=!subset_TFile!"


Print #1, "set EUT_SFile=%%~fF"

End If

Print #1, "set EUT_TFile=%EUT_TDirectory%\001

Main\%%~nF_%EUT_Suffix%.dim"

Print #1, "set EUT_procCmd=%gptPath% %EUT_graphXml%

-Ssource=" & """!EUT_SFile!""" & " -t " & """!EUT_TFile!"""

Print #1, "echo !EUT_procCmd!"

Print #1, "call !EUT_procCmd!"


Print #1, "set

TSM_TDirectory=!EUT_TDirectory!\TSM"


152


out_format; " -b 35 " & """!EUT_TFile!""" & " -o " &




End If


Print #1, "set

CHL_TDirectory=!EUT_TDirectory!\CHL"



out_format; " -b 36 " & """!EUT_TFile!""" & " -o " &




End If

End If



Print #1, "set BOR_SFile=!subset_TFile!"


Print #1, "set BOR_SFile=%%~fF"

End If

Print #1, "set BOR_TFile=%BOR_TDirectory%\001

Main\%%~nF_%BOR_Suffix%.dim"

Print #1, "set BOR_procCmd=%gptPath% %BOR_graphXml%

-Ssource=" & """!BOR_SFile!""" & " -t " & """!BOR_TFile!"""

Print #1, "echo !BOR_procCmd!"

Print #1, "call !BOR_procCmd!"


Print #1, "set

TSM_TDirectory=!BOR_TDirectory!\TSM"



out_format; " -b 35 " & """!BOR_TFile!""" & " -o " &




End If


Print #1, "set

CHL_TDirectory=!BOR_TDirectory!\CHL"


Print #1, "Set CHLPconv_Cmd=%pconvPath% -f ";

out_format; " -b 36 " & """!BOR_TFile!""" & " -o " &




End If

153

End If


Print #1, "Set

FUB_XML=%FUB_XMLDirectory%\%%~nF.xml"

Print #1, "Set FUB_procCmd=%FUB_Path% "; quot;

"!FUB_XML!"; quot

Print #1, "echo !FUB_procCmd!"

Print #1, "call !FUB_procCmd!"

Print #1, "set FUB_BM_SFile=%FUB_TDirectory%\001

Main\%%~nF_FUB.dim"

Print #1, "set FUB_BM_TFile=%FUB_TDirectory%\001

Main\%%~nF_FUB_BM.dim"

Print #1, "set FUB_BM_procCmd=%gptPath%

%FUB_BM_XML% -Ssource=" & """!FUB_BM_SFile!""" & " -t " &

"""!FUB_BM_TFile!"""

Print #1, "echo !FUB_BM_procCmd!"

Print #1, "call !FUB_BM_procCmd!"


Print #1, "set

TSM_TDirectory=!FUB_TDirectory!\TSM"



out_format; " -b 3 " & """!FUB_BM_TFile!""" & " -o " &




End If


Print #1, "set

CHL_TDirectory=!FUB_TDirectory!\CHL"



out_format; " -b 1 " & """!FUB_BM_TFile!""" & " -o " &




End If

End If



Print #1, "set

SMAC_XML=%SMAC_XMLDirectory%\%%~nF.xml"

Print #1, "echo !SMAC_XML!"

Print #1, "Set SMAC_procCmd=%SMAC_Path% ";

quot; "!SMAC_XML!"; quot

Print #1, "echo !SMAC_procCmd!"

Print #1, "call !SMAC_procCmd!"

End If


154



Print #1, "set

Bmath_SFile=%SMAC_TDirectory%\%%~nF_Smac.dim"

Print #1, "set

Bmath_TFile=%Bmath_TDirectory%\001

Main\%%~nF_%Bmath_Suffix%.dim"

Print #1, "set Bmath_procCmd=%gptPath%

%Bmath_Xml% -Ssource=" & """!Bmath_SFile!""" & " -t " &

"""!Bmath_TFile!"""

Print #1, "echo !Bmath_procCmd!"

Print #1, "call !Bmath_procCmd!"

ElseIf smac.Value = 0 Then

Print #1, "set

Bmath_SFile=!subset_TFile!"

Print #1, "set





"""!Bmath_TFile!"""



End If



Print #1, "set

Bmath_SFile=%SMAC_TDirectory%\%%~nF_Smac.dim"

Print #1, "set





"""!Bmath_TFile!"""



ElseIf smac.Value = 0 Then

Print #1, "set Bmath_SFile=%%~fF"

Print #1, "set


Main\%%~nF%Bmath_Suffix%_.dim"



"""!Bmath_TFile!"""



End If

End If

For Ex_loop = 1 To Application.Cells(100, 100)

155

Ex_Var = Application.Cells(100 + Ex_loop,

101) & "_TDirectory"

Print #1, "set "; Ex_Var;

"=%Bmath_TDirectory%\"; Application.Cells(100 + Ex_loop,

101)

Print #1, "md "; quot; "!"; Ex_Var; "!";

quot

Ex_Cmd = Application.Cells(100 + Ex_loop,

101) & "Pconv_Cmd"

Print #1, "set "; Ex_Cmd; "=%pconvPath% -f

"; out_format; " -b "; Application.Cells(100 + Ex_loop,

100); """!Bmath_TFile!""" & " -o " & """!"; Ex_Var; "!""" &

TSM_Color_p

Print #1, "echo !"; Ex_Cmd; "!"

Print #1, "call !"; Ex_Cmd; "!"

Next Ex_loop

End If

End If


Print #1, "set

Clouds_XML=%Clouds_XMLDirectory%\%%~nF.xml"

Print #1, "echo !Clouds_XML!"

Print #1, "Set Clouds_procCmd=%Clouds_Path% ";

quot; "!Clouds_XML!"; quot

Print #1, "echo !Clouds_procCmd!"

Print #1, "call !Clouds_procCmd!"

End If

Print #1, ")"

Print #1, "echo msgbox "; quot; "Process finished!";

quot; ", vbInformation, "; quot; "MERIS Processing Tool";

quot; " > %tmp%\tmp.vbs"

Print #1, "cscript /nologo %tmp%\tmp.vbs"

Print #1, "del %tmp%\tmp.vbs"

Print #1, "pause"

Close #1

Shell(Mbat, vbMaximizedFocus)

fndloc.Hide()

End Sub

Private Sub S_Dir_Click()

in_file = Application.GetOpenFilename("All Supported

Products (*.dim; *.N1),*.dim;*.N1,BEAM-DIMAP product files

(*.dim), *.dim , ENVISAT MERIS (*.N1),*.N1")

If in_file = flase Then GoTo 10

fndloc.SDname = Left(in_file, Len(in_file) -

Len(Dir(in_file))) 'Left((in_file))

10:

End Sub

Private Sub smac_Change()


156

quot = """"

Dim SMAC_F_cont As String * 6000


SmacRfile = Application.GetOpenFilename("SMAC






1)


fc = fold.Files

For Each f1 In fc



format" Then



exlen = 4

extName = ".dim"


N1 format" Then

exlen = 3

extName = ".N1"

End If

In_SMAC_File =

objFSO.OpenTextFile(SmacRfile, ForReading, True)

Out_SMAC_File =


"\Processing Parameters\SMAC\" & Left(f1.Name, Len(f1.Name)

- exlen) & ".xml", ForWriting, True)

Do While Not In_SMAC_File.AtEndofStream

myLine = In_SMAC_File.ReadLine










End If



& "\SMAC\" & Left(f1.Name, Len(f1.Name) - exlen) &

"_Smac.dim" & quot & " format=" & quot & "BEAM-DIMAP" &

quot & " />"


157



Then


End If

Out_SMAC_File.WriteLine(myLine)

Loop

End If

Next

End If

End Sub

Private Sub Subset_Change()


Regdef.Show()

'fndloc.Reproj.Enabled = True


fndloc.RegCoord = ""

'fndloc.Reproj.Enabled = False

End If

End Sub

Private Sub T_Dir_Click()

out_file = Application.GetOpenFilename("All Files

(*.*),*.*")

If out_file = flase Then GoTo 20

fndloc.TDname = Left(out_file, Len(out_file) -

Len(Dir(out_file))) 'Left((out_file))

20:

End Sub

Private Sub UserForm_Initialize()

fndloc.TDname = Application.ActiveWorkbook.Path &

"\Output\"

typ.AddItem("ENVISAT MERIS (*.N1)")

typ.AddItem("BEAM-DIMAP product files (*.dim)")

out_typ.AddItem("GeoTIFF product format (*.tif)")

out_typ.AddItem("JPEG image format (*.jpg)")

out_typ.AddItem("Portal Network Graphics (*.png)")

out_typ.AddItem("Geotif Image Format (*.tif)")

out_typ.AddItem("Microsoft Bitmap image (*.bmp)")

fndloc.c2r.Enabled = False

fndloc.eut_l.Enabled = False

fndloc.bor_l.Enabled = False

fndloc.fub.Enabled = False

fndloc.extsm.Enabled = False

fndloc.exchl.Enabled = False

fndloc.Form_bm.Enabled = False

fndloc.smac.Enabled = False

bmath.Value = 0

c2r.Value = 0

End Sub

158

2- Code for Subset region defination and ICOL processor

window

Dim icol_P_file

Dim icolP_file_cont

Private Sub ClrFld2_Click()

Regdef.ll_lat = ""

Regdef.ll_long = ""

Regdef.ur_lat = ""

Regdef.ur_long = ""

End Sub

Private Sub icol_Change()

Dim icol_P_file_cont As String * 6000

quot = """"


If icol.Value = -1 Then

5:

icol_P_file = Application.GetOpenFilename("ICOL

Parameters (*.xml),*.xml")

Open icol_P_file For Binary As #2

Get #2, 1, icol_P_file_cont

Get #2, 1, PF_Typ

Close #2

If icol_P_file = flase Then

icol.Value = 0

GoTo 20

End If



GoTo 5

End If

Close #1

fndloc.bmath.Enabled = False

ElseIf icol.Value = 0 Then

'fndloc.extsm.Enabled = False

'fndloc.exchl.Enabled = False

fndloc.bmath.Enabled = True

End If

20:

icolP_file_cont = icol_P_file_cont

End Sub

Private Sub ReadKML_Click()

Dim xmlDom As MSXML2.DOMDocument

Dim xmlPlaceMark As MSXML2.IXMLDOMNode

Dim xmlPolygon As MSXML2.IXMLDOMNode

Dim xmlCoord As MSXML2.IXMLDOMNode

Dim sName As String

159

Dim vaSpace As Object, vaComma As Object

Dim i As Long, j As Long

xmlDom = New MSXML2.DOMDocument

Kmlfile = Application.GetOpenFilename("SMAC Request

File (*.kml),*.kml")

xmlDom.Load(Kmlfile)

For i = 0 To

xmlDom.ChildNodes(1).ChildNodes(0).ChildNodes.Length - 1

'Polygone

If

xmlDom.ChildNodes(1).ChildNodes(0).ChildNodes.Item(i).nodeN

ame = "Placemark" Then

xmlPlaceMark =

xmlDom.ChildNodes(1).ChildNodes(0).ChildNodes.Item(i)

xmlPolygon = xmlPlaceMark.ChildNodes(2)

If xmlPolygon.nodeName = "Polygon" Then

xmlCoord =

xmlPolygon.ChildNodes(1).ChildNodes(0).ChildNodes(0)

'MsgBox xmlCoord.nodeName

With Sheet1.Cells(Sheet1.Rows.Count,

1).End(xlUp).Offset(1, 0)

vaSpace =

Split(xmlCoord.ChildNodes(0).Text, " ")

For j = LBound(vaSpace) To

UBound(vaSpace)

vaComma = Split(vaSpace(j), ",")

.Offset(j + 1, 1).Value =

vaComma(0)

Regdef.ll_lat = vaComma(1)

.Offset(j + 1, 2).Value =

vaComma(1)

Regdef.ll_long = vaComma(0)

Next j

End With

End If

'Path

xmlLine = xmlPlaceMark.ChildNodes(2)

If xmlLine.nodeName = "LineString" Then

'MsgBox xmlLine.nodeName

xmlCoord = xmlLine.ChildNodes(1)

'MsgBox xmlCoord.nodeName

With Sheet1.Cells(Sheet1.Rows.Count,

1).End(xlUp).Offset(1, 0)

vaSpace =

Split(xmlCoord.ChildNodes(0).Text, " ")

For j = LBound(vaSpace) To

UBound(vaSpace)

If j = 0 Then

160

vaComma = Split(vaSpace(j),

",")

'.Offset(j + 1, 1).Value =

vaComma(1)

Regdef.ll_lat = vaComma(1)


vaComma(0)

Regdef.ll_long = vaComma(0)

End If

If j = 1 Then

vaComma = Split(vaSpace(j),

",")


vaComma(1)

Regdef.ur_lat = vaComma(1)


vaComma(0)

Regdef.ur_long = vaComma(0)

End If

Next j

End With

End If

End If

Next i

End Sub

Private Sub regset_Click()

quot = """"

reg_xml = Application.ActiveWorkbook.Path &

"\Processing Parameters\" & Regdef.XML_fn & ".xml"

'MsgBox Reg_xml

If Regdef.ll_lat = "" Or Regdef.ll_long = "" Or

Regdef.ur_lat = "" Or Regdef.ur_long = "" Then

MsgBox("Please Define Region coordinate")

Else

URlat = Regdef.ur_lat ': Application.Cells(11, 3) =

URlat

URlong = Regdef.ur_long ': Application.Cells(11, 2)

= URlong

ULlat = Regdef.ur_lat ': Application.Cells(12, 3) =

ULlat

ULlong = Regdef.ll_long ': Application.Cells(12, 2)

= ULlong

LLlat = Regdef.ll_lat ': Application.Cells(13, 3) =

LLlat

LLlong = Regdef.ll_long ': Application.Cells(13, 2)

= LLlong

LRlat = Regdef.ll_lat ': Application.Cells(14, 3) =

LRlat: Application.Cells(15, 3) = URlat

161

LRlong = Regdef.ur_long ': Application.Cells(14, 2)

= LRlong: Application.Cells(15, 2) = URlong

If Regdef.SubIcol.Value = -1 Then

Open reg_xml For Output As #1

Print #1, "<graph id=" & """SubSetId""" & ">"


Print #1, " <node id=" & """SubSetNode""" & ">"

Print #1, " <operator>Subset</operator>"





Print #1, " <geoRegion>POLYGON((" & URlong

& " " & URlat & "," & ULlong & " " & ULlat & "," & LLlong &

" " & LLlat & "," & LRlong & " " & LRlat & "," & URlong & "

" & URlat & "))</geoRegion>" 'Coordinates

Print #1, "

<copyMetadata>True</copyMetadata>"


Print #1, "</node>"

fndloc.RegCoord = "Only Subset was select"

If icol.Value = -1 Then

Print #1, " <node id="; quot; "ICOL"; quot;

">"

Print #1, " <operator>icol.Meris</operator>"


Print #1, "

<sourceProduct>SubSetNode</sourceProduct>"


Print #1, Left(icolP_file_cont,

FileLen(icol_P_file))


fndloc.RegCoord = "Subset And ICOL

Processor were selected"

End If


Close #1

End If

If Regdef.ICOLonly.Value = -1 Then

Open reg_xml For Output As #1

Print #1, "<graph id=" & """SubSetId""" & ">"


Print #1, " <node id="; quot; "ICOL"; quot; ">"

Print #1, " <operator>icol.Meris</operator>"


Print #1, "

<sourceProduct>${source}</sourceProduct>"


162

Print #1, Left(icolP_file_cont,

FileLen(icol_P_file))



Close #1

fndloc.RegCoord = "Only ICOL Processor was

selected"

End If

End If

Regdef.Hide()

End Sub

Private Sub SubIcol_Change()

If Regdef.SubIcol.Value = -1 Then

Regdef.Frame1.Enabled = True

Regdef.Frame2.Enabled = True

Regdef.ur_lat.Enabled = True

Regdef.ur_long.Enabled = True

Regdef.ll_lat.Enabled = True

Regdef.ll_long.Enabled = True

Regdef.ReadKML.Enabled = True

Regdef.Label1.Enabled = True




End If

End Sub

Private Sub ICOLonly_Change()

If Regdef.ICOLonly.Value = -1 Then

Regdef.Frame1.Enabled = False

Regdef.Frame2.Enabled = False

Regdef.ur_lat.Enabled = False

Regdef.ur_long.Enabled = False

Regdef.ll_lat.Enabled = False

Regdef.ll_long.Enabled = False

Regdef.ReadKML.Enabled = False

Regdef.Label1.Enabled = False




End If

End Sub

Private Sub UserForm_Initialize()

Regdef.ll_lat = 21.483242

Regdef.ll_long = 30.366176

Regdef.ur_lat = 23.965172

Regdef.ur_long = 33.48221

Regdef.XML_fn = "Subset Region"

End Sub

163

3- Code for Band Math window

Dim Ex_num As Integer

Public Sub Bm_add_Click()

bmath_xml = Application.ActiveWorkbook.Path & "\Processing

Parameters\Band Math.xml"

Open bmath_xml For Append As #1


Print #1, " <name>"; FBmath.Ex_Name;

"</name>"


FBmath.Bm_Ex; "</expression>"

Print #1, " <description>";

FBmath.Ex_Des; "</description>"


Print #1, " <unit>"; FBmath.Ex_Un;

"</unit>"

Print #1, "



add_Ex = MsgBox("Add Another Expression", vbYesNo)

Ex_num = Ex_num + 1

Application.Cells(100 + Ex_num, 100) = Ex_num

Application.Cells(100 + Ex_num, 101) =

FBmath.Ex_Name.Value

If add_Ex = 6 Then

FBmath.Ex_Name.Value = ""

FBmath.Bm_Ex.Value = ""

FBmath.Ex_Des.Value = ""

FBmath.Ex_Un.Value = ""

End If

Close #1

Application.Cells(100, 100) = Ex_num

End Sub

Private Sub Bm_OK_Click()



If Ex_num > 0 Then

Open bmath_xml For Append As #1

Print #1, " </targetBands>"







Print #1, "

<source>BandmathNode</source>"


164


Print #1, " <crs>AUTO:42001</crs>"

Print #1, "


Print #1, "

<includeTiePointGrids>false</includeTiePointGri

ds>"



End If


Close #1

Else

Open bmath_xml For Output As #1

Close #1

End If

FBmath.Hide()

End Sub

Private Sub UserForm_Activate()

Ex_num = 0



Open bmath_xml For Output As #1

Print #1, "<graph id="; """BandmathId"""; ">"


Print #1, " <node id="; """BandmathNode"""; ">"

Print #1, " <operator>BandMaths</operator>"





Print #1, " <targetBands>"

Close #1

End Sub

165

B. 1. Excel Custom functions

Function MNB(ByVal obs As Range, ByVal comp As Range)

MNB = 0

nRows = obs.Rows.Count

For i = 1 To nRows

If obs(i, 1) = 0 Or comp(i, 1) = 0 Then GoTo 10

MNB = MNB + 100 * (comp(i, 1) - obs(i, 1)) / obs(i,

1)

n = n + 1

10:

Next i

Result:

MNB = MNB / n

End Function

Function ModBias(ByVal obs As Range, ByVal comp As Range)

ModBias = 0

n = 0


For i = 1 To nRows


ModBias = ModBias + (comp(i, 1) - obs(i, 1))

10:

Next i

Result:

ModBias = ModBias

End Function

Function FracBias(ByVal obs As Range, ByVal comp As Range)

obs_Sum = 0

comp_Sum = 0

n = 0


For i = 1 To nRows


obs_Sum = obs_Sum + obs(i, 1)

comp_Sum = comp_Sum + comp(i, 1)

n = n + 1

10:

Next i

obs_avg = obs_Sum / n

comp_avg = comp_Sum / n

Result:

FracBias = 2 * ((obs_avg - comp_avg) / (obs_avg +

comp_avg))

End Function

Function NMSE(ByVal obs As Range, ByVal comp As Range)

DifSqr = 0

obs_avg = 0

comp_avg = 0

166

n = 0


For i = 1 To nRows


DifSqr_Sum = DifSqr_Sum + ((comp(i, 1) - obs(i, 1))

^ 2)

obs_Sum = obs_Sum + obs(i, 1)

comp_Sum = comp_Sum + comp(i, 1)

n = n + 1

10:

Next i

DifSqr_avg = DifSqr_Sum / n

obs_avg = obs_Sum / n

comp_avg = comp_Sum / n

Result:

NMSE = DifSqr_avg / (obs_avg * comp_avg)

End Function

Function GMBias(ByVal obs As Range, ByVal comp As Range)

lnobs_Sum = 0

lncomp_Sum = 0


For i = 1 To nRows


lnobs_Sum = lnobs_Sum +

Application.WorksheetFunction.Ln(obs(i, 1))

lncomp_Sum = lncomp_Sum +

Application.WorksheetFunction.Ln(comp(i, 1))

n = n + 1

10:

Next i

Result:

lnobs_avg = lnobs_Sum / n

lncomp_avg = lncomp_Sum / n

GMBias = Exp(lnobs_avg - lncomp_avg)

End Function

Function GMVariance(ByVal obs As Range, ByVal comp As

Range)

GMVar1 = 0


For i = 1 To nRows


lnobs = Application.WorksheetFunction.Ln(obs(i, 1))

lncomp = Application.WorksheetFunction.Ln(comp(i,

1))

GMVar1 = GMVar1 + ((lnobs - lncomp) ^ 2)

n = n + 1

10:

Next i

Result:

167

GMVariance = Exp(GMVar1 / n)

End Function

Function fac_2(ByVal obs As Range, ByVal comp As Range)


For i = 1 To nRows


comp_obs_R = comp(i, 1) / obs(i, 1)

If comp_obs_R >= 0.5 And comp_obs_R <= 2 Then

m = m + 1

End If

n = n + 1

10:

Next i

Result:

fac_2 = m / n

End Function

168

Figure (‎B-4) ArcMap Model Builder used to cloud coverage percentage

169

Figure (‎B-5) ArcMap Model Builder used to Extract Lake Nasser Surface

171

APPENDIX (C)

Stepwise Regression charts

This appendix contains Matlab output of stepwise regression, the output

was represent as a graph for each parameter showing coefficient value for

each independent variable in means of error bar and t-stat and p-value.

The intercept of the regression model and R-squared value and RMSE are

presented as well. The variables with blue color are that included in the

selected regression model.

172

TDS

Transparency

SiO2

TP Turbidity

TSS

Figure (‎C-1) Stepwise Regression Analysis Results For 2003

173

TDS

TP

SiO2

Transparency

Turbidity


174

Chl-a

TSS

TP

Transparency

SiO2

TDS


175

TSS

SiO2

TP

Transparency


176

TSS

TDS TP

Transparency

Turbidity


177

TSS


جامعة عين شمس

كلية الهندسة

قسم الري والهيدروليكا

بحيرة ناصر باستخدام تقنية التوزيع المكاني لمعامالت نوعية المياه في

االستشعار من بعد

مهندس

محمد علي حامد غريب (2003) جامعة بنها – كلية الهندسة بشيرا -مدنيةالهندسة البكالوريوس

(2010) جامعة عين شمس -كلية الهندسة - الهندسة المدنية في ماجستير

رسالة مقدمة كجزء من متطلبات الحصول على

ري وهيدروليكا –في الهندسة المدنية الفلسفة دكتوراهدرجة

شراف البحراوي نبيه علي .د أ.

الهيدروليكاأستاذ

كليــــــــــة الهندســـــــــة - والهيــــدروليكا قسم الري

جامعــــــة عيـــن شمــــس

د. محسن محمود يسري معهد بحوث النيل - أمين عام

المركز القومي لبحوث المياه

والريوزارة الموارد المائية

عطية محمود كريمة .د أ. بحوث الموارد المائية دمدير معه



جمهورية مصر العربية – القاهرة

2016

كليــة الهندســة


نوعية المياه في بحيرة ناصر باستخدام تقنية التوزيع المكاني لمعامالت


إعداد

محمد علي حامد غريبالمهندس / (2003بكالوريوس الهندسة المدنية )

(2010ري وهيدروليكا ) –ماجستير في الهندسة المدنية

رسالة مقدمة كجزء من المتطلبات للحصول علي درجة الدكتوراه

في الهندسة المدنية )ري وهيدروليكا(

لجنة الحكم

التوقيع

……………… سعد عزيز مدحت .د أ. معهد بحوث النيل مدير



……………… إيمان العزيزي .د أ. الهيـــدروليكاأستاذ

ةالهندســكليــة -الري والهيـــدروليكا قسم

جامعـــــة عيـــن شمــــس

……………… علي نبيه البحراوي .د أ. أستاذ الهيـــدروليكا

كليــة الهندســة -قسم الري والهيـــدروليكا

جامعـــــة عيـــن شمــــس

……………… كريمة محمود عطية .د أ. بحوث الموارد المائية معهد مدير



2016/----/----التاريخ

كلية الهندسة


: محمد علي حامد غريب الطــــالب اسم

التوزيع المكاني لمعامالت نوعية المياه في بحيرة ناصر : عنوان الرسالة

باستخدام تقنية االستشعار من بعد

دكتوراه الفلسفة : الدرجـــة

شرافلجنة اإل

قسم الري والهيــــدروليكا دروليكاـــالهيأستاذ البحراوي نبيه علي .د أ.

جامعــــــة عيـــن شمــــس -ـة الهندســـــــــةـكليــــ

معهد بحوث الموارد المائية مدير عطية محمودد. كريمة أ.

والريوزارة الموارد المائية - المركز القومي لبحوث المياه

معهد بحوث النيل -أمين عام د. محسن محمود يسري

والريوزارة الموارد المائية - المركز القومي لبحوث المياه

2016تاريخ البحث : / /

2016أجيزت الرسالة بتاريخ / / :الدراسات العليا

:ختم اإلجازة

موافقة مجلس الجامعة الكليةموافقة مجلس

/ /2016 / /2016

هداءإ

إلى زوجتي العزيزة إلى ابنائي علي واحمد

إلى والدي العزيز إلى والدتي العزيزة

إلى أخي أحمد وأسرته

الرسالة بمقدم التعريف

محمد علي حامد غريب: االســــــــــم

6/12/1980: تاريــــخ الميــالد

القليوبية -شبرا الخيمة ثان -بهتيم : محــــل الميــالد

: بكالوريوس الهندسة المدنية الدرجة الجامعية االولـي

مدنية: هندسة التــخـــصـــص

بنهاجامعة –كلية الهندسة بشبرا : االوليالجهة المانحة الدرجة الجامعية

2003: تاريخ المنــــــح

لمدنية: ماجستير الهندسة ا الدرجة الجامعية الثانية

ري وهيدروليكا: التــخـــصـــص

جامعة عين شمس -كلية الهندسة : الجهة المانحة الدرجة الجامعية الثانية

2010 ليويو: تاريخ المنــــــح

.المركز القومي لبحوث المياه –معهد بحوث النيل – باحث مساعد: الوظيفــــــــة

2016التاريخ .../...../

جـــامعــة عيـــن شمس

كليـــــــة الهنـــــدســـــة


محمد علي حامد غريب اسم مقدم الرسالة:

المكاني لمعامالت نوعية المياه في بحيرة ناصر باستخدام تقنية التوزيع عنوان الرسالة:


مستلخص الرسالةيهدف هذا البحث الى استخدام االستشعار من بعد في تقدير معامالت نوعية المياه ببحيرة ناصر،

رحالت تم استخدام نتائج ستة حيث تعتبر بحيرة ناصر المخزون االستراتيجي للمياه بمصر.

للتحقق من (Envisat/MERIS)تم قياسها في فترات مختلفة، مع استخدام صور ،حقلية

Case 2 Regional (C2R), Eutrophic lake, and Boreal)استخدام بعض االدوات مثل

Lake)باستخدام ين الناتج عن االنعكاسات االرضيةاب، سواء بعد تصحيح الت(Improved

Contrast between ocean and Land – ICOL) ،لحساب المعامالت وذلك او بدونه

الضوئية )المواد العالقة والكلوروفيل(. تم التحقق من هذه االدوات باستخدام بعض المقاييس

اوضحت نتائج البحث انه بالنسبة للمواد العالقة يمكن استخدام االحصائية. بالنسبة للمواد العالقة،

كما يمكن استخدام في فترة نهاية الفيضان، (ICOL)بدون استخدام (Eutrophic Lake)اداة

لتقدير المواد العالقة في فترة (ICOL)بدون استخدام (Eutrophic Lake)و (C2R)كال من

عدم مالئمة كل االدوات في تقدير تركيز المواد العالقةالتصرفات المنخفضة، كما اظهر البحث

مع استخدام (Eutrophic Lake) لكلوروفيل، يمكن استخدامفي فترة بداية الفيضان .اما ا

(ICOL) تم .فترة نهاية الفيضانو التصرفات المنخفضةفي تقدير تركيز الكلوروفيل في فترة

والغير ضوئية في فترات الفيضان المختلفة ةاستنتاج معادالت خاصة لتقدير المعامالت الضوئي

يمكن 6و 5الغالف الجوي ووجد ان البندين باستخدام الصور بعد تصحيحها من تأثير

استخدامها في تقدير المعامالت المختلفة. تم التحقق من المعادالت في فترة انتهاء الفيضان

صورة لعمل سالسل زمنية )منحنيات وخرائط( بعد استبعاد الصور المتأثرة 913وتطبيقها على

وبعدها 2005ة وإجمالي الفسفور حتي عام بالسحب. حيث تم مالحظة ازدياد تركيز المواد العالق

بدأ التركيز في االنخفاض، وبالنسبة للسليكا والمواد الصلبة الذائبة فوجد انهما يتغيران بشكل

في التغير شكل معين. وقد وجد ان الفترة من عمعاكس. وبالنسبة لشفافية المياه فوجد انها ال تتب

لية وجود سحب وبالتالي تعتبر مناسبة إلجراء سبتمبر الى نوفمبر هي االقل من حيث احتما

قياسات ضوئية بالبحيرة والتي تفيد في عمل نموذج ضوئي خاص بالبحيرة.

البحراوي نبيه علي .د أ. المشرفون:

ا.د. كريمة محمود عطية

د. محسن محمود يسري

باللغة العربية ملخص الرسالة

عنوان الرسالة

المكاني لمعامالت نوعية المياه في بحيرة ناصر باستخدام تقنية االستشعار من التوزيع

مقدمة من المهندس/ محمد علي حامد غريب بعد

استخدام الطرق التقليدية في قياس نوعية المياه تعتمد على اخذ عينات عند نقاط محددة وبذلك

حالة يكون من الصعب معرفة فهي تعبر فقط عن التغييرات عند نفس نقط القياس، وفي هذه ال

هذا باالضافة الى ارتفاع تكاليف المأموريات الحقلية مما التغيرات في باقي المسطح المائي،

االستشعار تقنية لذا فإن استخدام يصعب مع تكرار هذه المأموريات في اوقات مختلفة من العام.

في حالة البحيرات ذات المساحات معرفة المعامالت المختلفة لنوعية المياه يكون مفيدا لمن بعد

السطحية الكبيرة مثل بحيرة ناصر التي تعتبر من اكبر البحيرات الصناعية في العالم، كما تعتبر

.المخزن الرئيسي للمياه العذبة بمصر

تنحصر مشاكل دراسة نوعية المياه في ارتفاع تكاليف القياسات الحقلية المتكررة باإلضافة الى

بقا عن عدم امكانية دراسة التغير في المسطح المائي بالكامل، ما تم ذكرة سا

تي:وتتلخص اهداف البحث في اآل

( في Case 2 water quality processorsالتحقق من استخدام بعض االدوات ) .1

Case 2تقدير المواد العالقة والكلوروفيل ببحيرة ناصر. ومن هذه االدوات )

Regional (C2R), Eutrophic lake, and Boreal Lake وهي عبارة عن )

ادوات تم تجربتها في مناطق مختلفة من العالم.

نوعية المياه واالنعكاسات السطحية المختلفة لمعامالت الن ايجاد معادالت تربط بي .2

( بعد تصحيحها من اثر الغالف الجوي.MERISالمستنتجة من صور )

فة لنوعية المياه.عمل سالسل زمنية للتغير في المعامالت المختل .3

يمكن حساب نسبة السحاب واالنعكاس الشمسي فوق البحيرة في االوقات المختلفة حيث .4

إلجراء القياسات الحقلية المستقبلية لربطها ةالمناسباالستفادة منها في تحديد االشهر

ببيانات االقمار الصناعية

وقد إشتملت الدراسة على األبواب األتية:

المقدمة الباب االول:

بالتغير في تركيز المواد العالقةبمشكلة البحث وعالقة التعريف يشمل على تقديم يهدف الى

محتويات كل باب. واستعراضكما يحتوي أيضا على أهداف الرسالة معامالت نوعية المياه.

الدراسات السابقة الباب الثاني:

التعريف باالستشعار من بعد وتاريخه يحتوي هذا الباب على الدراسات السابقة في مجال البحث و

في مجال نوعية المياه على مستوى العالم، كما يتضمن على ملخص االبحاث وعن استخدامه

التي أُستخدم فيها االستشعار من بعد في بحيرة ناصر في مجاالت عدة مثل البخر. كما يعطي

ايضا اسباب اختيار نقطة البحث.

والبيانات المستخدمة اسةخصائص منطقة الدر الباب الثالث:

يحتوي هذا الباب على التعريف ببحيرة ناصر من حيث موقعها الجغرافي واهميتها، كما يتضمن

اقبة البحيرة المستخدم حاليا. باإلضافة الى الخصائص الفزيائية توصيف لبرنامج مر

المتاحة للدراسة توصيف البيانات . كما يحتوي على والهيدرولوجية وخصائص الترسيب للبحيرة

.سواء كانت بيانات حقلية او بيانات اقمار صناعية

منهجية البحث الباب الرابع:

Case)البحث سواء المستخدمة في التحقق من استخدامالمتبعة بمنهجية العلى هذا البابيحتوي

2 water quality processors) بكل معامل من ةاو المستخدمة في ايجاد معادالت خاص

. كما يحتوي ايضا ، باإلضافة الى توصيف االدوات المستخدمة بالبحثت نوعية المياهمعامال

على توصيف للمقاييس االحصائية المستخدمة في تقييم النتائج.

تحليل النتائج الباب الخامس:

ناتجة عن التحقق من استخدام ادوات سواء البحث نتائج ليحتوي هذا الباب عل تحليل تفصيلي

(Case 2 Water quality processor) لتقدير كال من المواد العالقة والكلورفيل مع ايضاح

كما يحتوي على المعادالت التي تم .االدوات المالئمة لالستخدام في الفترات المختلفة في البحيرة

بين معامالت نوعية (Stepwise regression)استنتاجها باستخدام طريقة االنحدار المتدرج

مع تحديد معادلة واحدة لكل معامل لكل فترة من (MERIS)النعكاسات السطحية لصورالمياه وا

على تحليل مفصل عن نسبة السحاب فترات الفيضان باستخدام المقاييس االحصائية. كما يحتوي

االنعكاس الشمسي على سطح البحيرة حيث تفيد النسبة في استبعاد بعض الصور المستخدمة في و

ل الزمني للمعامالت المختلفة ويمكن االستفادة منها ايضا في تحديد افضل عمل منحني التسلس

يحتوي على االشهر إلجراء القياسات الحقلية المستقبلية لربطها ببيانات االقمار الصناعية. و

التي توضح تغير المعامالت المختلفة. منحنيات التسلسل الزمني

التوصياتالخالصة و الباب السادس:

من نتائج باالضافة الى التوصيات التي يجب ان البحث أهم ما توصل إليه ىلباب عليحتوي هذا ا

تؤخذ في االعتبار في االبحاث المستقبلية.

: كما يلي كما تحتوى الرسالة على ثالثة مالحق

من حيث دقة الصورة (MERIS)خصائص الصور المستخدمة يحتوي الملحق على ملحق )أ(:

ومستويات المعالجة الخاصة بها واالستخدامات المختلفة لها. كما يحتوي على شرح مفصل

خالل البحث. باالضافة الى توضيح طريقة (Case 2 water quality)لالدوات المستخدمة

.(APHA, 2012)أخذ وتحليل العينات من الطبيعة طبقا للمواصفات العالمية

الذي تم برمجته (MERIS Processing Tool)علي عرض لبرنامج يحتوي ملحق )ب(:

لتسهيل عملية معالجة الصور (Visual Basic) خالل هذا البحث باستخدام لغة الفيجوال بيسك

(MERIS) التحقيق واستنباط المعادالت او في مرحلة عمل التسلسل الزمني مرحلة سواء في

توي لب معالجة عدد كبير من صور االقمار الصناعية. كما يحللمعامالت نوعية المياه حيث تتط

الى بعض االدوات التي باإلضافةعلى الكود المستخدم في حساب المقاييس االحصائية المختلفة.

وذلك ايضا لتسهيل عملية معالجة بيانات صور (ArcMap)تم تطويرها باستخدام برنامج

االقمار الصناعية

بواسطة برنامج المتدرج اشكال توضح نتائج استخدام طريقة االنحدار يحتوي على ملحق )جـ(:

.(Matlab) مات الب

AIN SHAMS UNIVERSITY FACULTY OF ENGINEERING ... Thesis...AIN SHAMS UNIVERSITY FACULTY OF ENGINEERING...

Documents

Transcript of AIN SHAMS UNIVERSITY FACULTY OF ENGINEERING ... Thesis...AIN SHAMS UNIVERSITY FACULTY OF ENGINEERING...