Selection of the appropriate data source for Agriculture ... · Selection of the appropriate data...

Selection of the appropriate data source

for Agriculture Risk Modelling

Mozambique Agro Risk - Barue District

Presented By: Pushpendra Johari

www.rmsi.comwww.rmsi.com

Agri. Risk Model Components

Hazard Module – Hazard related information

Vulnerability Module – Crop wise and region specific curves

Exposure Module – Crop production and acreage data

Insurance Module – Policy Information, Indices – by policy, by

Geographic areas, Perform accumulation control, Portfolio

analysis ,Cession limits, Liability limits

# 2


Hazard Module – Hazard related information

Primary parameters - Rainfall and Temperature

Available from local Met Agencies

Available from local Agriculture departments

Satellite based

– TRMM (Tropical Radar

– NOAA

Synthetic – Santa Clara University

? WHICH ONE TO USE

# 3


Case Study

Mozambique Agro Risk - Barue District

# 4

www.rmsi.com

General Information about Barue

Area:

– 5,800 km^2 (equivalent to

160km * 40km)

– 0.73% of country area

– 60% of districts are smaller

Maize production:

Year 2009

Season 2008-09

Planted area Barue 18,825

Ha Country total 1,613,405

Percentage 1.2%

Production Barue 28,237

Tons Country total 1,842,993

Percentage 1.5%

Yield Barue 1.50

Ton/ha Country range 0.4 to 2.0

# 5


Rainfall Stations

No rainfall stations located in the

district

Two stations (Luenha II, and

Mazoi I) located north

Two stations (Sussundenga, and

Chimoio) located south.

Luenha II has about 17% missing

data and Mazoi I has about 51 %

missing. These two stations are

from DNA. Ignored.

Sussundenga and Chimoio (From

INAM/IAM) have continuous data

for nearly 40 years. Used in this

study.

# 6

www.rmsi.com

Stations Sussundenga and Chimoio

41 years of daily rainfall (1969-2009)

– Chimoio (1960-2009)

Distance from Barue centroid

– Sussundenga: 111.3 km

– Chimoio: 127.3 km

– Weighted by the inverse of distance^2

– They are essentially averaged.

Seasonal rainfall: sum of November to April (6 months)

# 7

www.rmsi.com

Seasonal Rainfall Comparison

0.0

500.0

1000.0

1500.0

2000.0

2500.0

1965 1970 1975 1980 1985 1990 1995 2000 2005 2010

Seasonal total rainfall (mm)

Chimoio

Sussandenga

Centroid of Barue

# 8


SPI- Index

The standard normal distribution has a mean of zero and a standard deviation of one. SPI is a representation of the number of standard deviations from the mean at which an event occurs. The unit of the SPI can thus be considered to be standard deviation or a Z score.

# 9


GeoWRSI

is a geo-spatial, stand-alone

application for estimation of WRSI

(Water Requirements Satisfaction

Index )

is implemented by the USGS for the

FEWSNET Activity.

runs a crop-specific water balance

model for a selected region in the

world, using raster data inputs

Includes satellite based Dekadal

rainfall from 1995 onwards.

produces a range of outputs which can

either be used qualitatively to assess

and monitor crop conditions during the

crop growing season, or can be

regressed with yields to produce yield

estimation models and yield estimates.

(http://hollywood.geog.ucsb.edu/wb/geowrsi.php)

# 10

http://hollywood.geog.ucsb.edu/wb/geowrsi.php

www.rmsi.com

Year Classification: Normal, Dry & Wet

Deviation from Normal

– + 25 % from - Wet year,

– - 25% from Normal - Dry Year

– SPI greater than 0.5 - Wet year

– SPI less than - 0.5 - Dry Year

Summary rainfall (mm per season)

Mean 1032.6

75%Mean 774.5

125%Mean 1290.8

# 11

www.rmsi.com

Year Classification

Season

Barue

Seasonal RF

Barue

Seasonal

SPI

Classificaion

Deviation

Classification

SPI

1969/1970 911.1 -0.21 N N

1970/1971 795.3 -0.56 N D

1971/1972 1175.6 0.48 N N

1972/1973 510.5 -1.60 D D

1973/1974 1677.2 1.56 W W

1974/1975 1202.1 0.55 N W

1975/1976 1403.8 1.01 W W

1976/1977 1138.0 0.39 N N

1977/1978 1113.1 0.33 N N

1978/1979 801.6 -0.54 N D

1979/1980 923.3 -0.18 N N

1980/1981 1381.4 0.96 W W

1981/1982 1158.3 0.44 N N

1982/1983 709.4 -0.84 D D

1983/1984 787.7 -0.59 N D

1984/1985 1438.7 1.08 W W

1985/1986 1300.9 0.78 W W

1986/1987 766.3 -0.65 D D

1987/1988 1404.9 1.01 W W

1988/1989 1391.8 0.98 W W

Season

Barue

Seasonal

RF

Barue

Seasonal

SPI

Classificaio

n Deviation

Classification

SPI

1989/1990 1041.3 0.14 N N

1990/1991 562.8 -1.38 D D

1991/1992 389.7 -2.17 D D

1992/1993 950.0 -0.10 N N

1993/1994 852.3 -0.39 N N

1994/1995 921.9 -0.18 N N

1995/1996 502.2 -1.64 D D

1996/1997 1486.5 1.18 W W

1997/1998 382.2 -2.21 D D

1998/1999 1380.7 0.96 W W

1999/2000 1239.4 0.64 N W

2000/2001 1504.1 1.22 W W

2001/2002 1135.2 0.38 N N

2002/2003 685.1 -0.93 D D

2003/2004 1038.7 0.14 N N

2004/2005 516.1 -1.58 D D

2005/2006 1015.1 0.08 N N

2006/2007 681.3 -0.94 D D

2007/2008 1664.1 1.54 W W

2008/2009 1365.9 0.92 W W

# 12

www.rmsi.com

Rainfall Stations SPI

Combination of both stations

– SPI derivation

• Seasonal

• 6 month ISPI (Nov through March)

• 3 month ISPI (Jan, Feb, March)

Chimoio station

– SPI derivation

• Seasonal



Drought Nov Dec Jan Feb Mar Apr Total

Weights for 3 months 0.00 0.00 0.55 0.30 0.15 0.00 1.00


Flood Nov Dec Jan Feb Mar Apr Total



# 13

www.rmsi.com

Water Requirement Satisfaction Index

Dekadal (10 days) rainfall at 10km grid (1995-2011) (WRSI)

– SPI derivation

• Seasonal



WRSI analysis

– Dekadal values

– 16th Dekad after planting WRSI values correlate best with

production

# 14

www.rmsi.com

Planted Area and Production

Nine years of Maize crop data

-

5,000

10,000

15,000

20,000

25,000

30,000

-

5,000

10,000

15,000

20,000

25,000

30,000

2000 2002 2004 2006 2008 2010

Pla

nte

d a

rea

(ha)

Pro

du

ctio

n (

ton

)Production

Planted Area

Season

Barue

Production(t)

Barue Planted Area

(ha)

2000/2001 13,618 14,487

2001/2002 13,969 13,969

2002/2003 17,421 14,517

2003/2004 21,326 15,233

2004/2005 9,212 12,877

2005/2006 20,452 15,732

2006/2007 20,689 15,915

2007/2008 24,134 16,090

2008/2009 28,237 18,825

# 15

www.rmsi.com

Maize Yield (Ton/ha)

0.60

0.70

0.80

0.90

1.00

1.10

1.20

1.30

1.40

1.50

1.60

2000 2002 2004 2006 2008 2010

Barue Maize Yield (Ton/HA)

Season

Barue Yield

(t/ha) Pot Yield

85% Pot

Yield

2000/2001 0.94 1.50 1.28

2001/2002 1.00 1.50 1.28

2002/2003 1.20 1.50 1.28

2003/2004 1.40 1.50 1.28

2004/2005 0.72 1.50 1.28

2005/2006 1.30 1.50 1.28

2006/2007 1.30 1.50 1.28

2007/2008 1.50 1.50 1.28

2008/2009 1.50 1.50 1.28

Barue Seasonal_SPI

1.22

0.38

-0.93

0.14

-1.58

0.08

-0.94

1.54

0.92

# 16

www.rmsi.com

Chimoio Potential Yield vs. 85% Potential

# 17

www.rmsi.com

Seasonal SPI vs. ISPI (6 months)

y = 0.305x - 0.048R² = 0.167

-60%

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

-3.00 -2.00 -1.00 0.00 1.00 2.00 3.00

MD

R (

85

%P

otY

ield

)

Barue ISPI (7)

Barue ISPI (7) Vs MDR (85%Potential Yield)

y = -0.194x2 - 0.055x + 0.030R² = 0.699

-60%

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

-3.00 -2.00 -1.00 0.00 1.00 2.00 3.00

MD

R (

85

%P

otY

ield

)

Barue Seasonal SPI

Barue Seasonal SPI Vs MDR (85%Potential Yield)

y = -0.450x - 0.095R² = 0.408

-60%

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

-3.00 -2.00 -1.00 0.00 1.00 2.00 3.00

MD

R (

85

%P

otY

ield

)

Seasonal SPI

Chimoio ISPI Vs MDR (85% Potential Yield)

# 18

www.rmsi.com

SPI WRSI PPT – Seasonal & ISPI (6)

y = -0.272x2 - 0.120x + 0.127R² = 0.485

-60%

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

-3.00 -2.00 -1.00 0.00 1.00 2.00 3.00

MD

R (

85

%P

otY

ield

)

Seasonal SPI

Barue WRSI Seasonal SPI Vs MDR 85% Potential Yield)

y = 0.590x2 + 1.077x - 0.050R² = 0.444

y = 0.994x - 0.040R² = 0.441

-60%

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

30%

-1.00 -0.50 0.00 0.50 1.00

MD

R (

85

%P

otY

ield

)

Seasonal SPI

Barue WRSI IISPI (ONJFM) Vs MDR (85% Potential Yield)

y = 0.447x - 0.096R² = 0.386

-60%

-50%

-40%

-30%

-20%

-10%

0%

-1.00 -0.50 0.00 0.50 1.00

MD

R

ISPI(JFM))

Drought - ISPI(JFM) Vs MDR (Potential Yield)

y = 0.853x - 0.484R² = 0.333

-60%

-50%

-40%

-30%

-20%

-10%

0%

-1.00 -0.50 0.00 0.50 1.00

MD

R

ISPI(NDJFMA))

Flood - ISPI(NDJFMA) Vs MDR (Potential Yield)

# 19

www.rmsi.com

WRSI Variation from 2nd to 19th Dekad

y = -0.028x + 2.522R² = 0.295

-60%

-50%

-40%

-30%

-20%

-10%

0%

10%

60 70 80 90 100

MD

R (

Po

t Y

ield

)

WRSI

Yield MDR Vs WRSI 2nd Dekad After Planting

y = -0.011x + 0.835R² = 0.358

-60%

-50%

-40%

-30%

-20%

-10%

0%

60 70 80 90 100

MD

R (

Po

t Y

ield

)

WRSI

Yield MDR Vs WRSI 8th Dekad After Planting

y = 0.003x - 0.571R² = 0.058

-60%

-50%

-40%

-30%

-20%

-10%

0%

60 70 80 90 100

MD

R (

Po

t Y

ield

)

WRSI

Yield MDR Vs WRSI 13th Dekad After Planting

y = 0.012x - 1.106R² = 0.511

-50%

-40%

-30%

-20%

-10%

0%

10%

20%

60 70 80 90 100

MD

R (P

ot Y

ield

)

WRSI

Yield (85%)MDR Vs WRSI 16th Dekad After Planting

# 20

www.rmsi.com

Reconstruction of Historical Events (Expected mode)

AAL Historical: 1,926 tons

AAL = 4,982 AAL = 1,803

-

2,000

4,000

6,000

8,000

10,000

12,000

- 5,000 10,000 15,000 20,000

HIS

TOR

ICA

L

2 stations SPI

SPI Seasonal 2 Stations

Series1

Series2

-

2,000

4,000

6,000

8,000

10,000

12,000

- 5,000 10,000 15,000

HIS

TORI

CAL

2 stations SPI

ISPI 6 month 2 Stations

Series1

Series2

AAL = 2,423 AAL = 2,545

-

2,000

4,000

6,000

8,000

10,000

12,000

0 5000 10000 15000

HIS

TOR

ICA

L

Chimoio SPI

SPI Seasonal Chimoio

Series1

Series2

-

2,000

4,000

6,000

8,000

10,000

12,000

0 5000 10000 15000

HIS

TORI

CAL

Chimoio SPI

ISPI 6 month Chimoio

Series1

Series2

# 21

www.rmsi.com

Reconstruction of Historical Events (Expected mode)

AAL Historical: 1,926

AAL = 2,364 AAL = 2,177

AAL = 3,284 AAL = 1,518

0.0

2,000.0

4,000.0

6,000.0

8,000.0

10,000.0

12,000.0

- 2,000 4,000 6,000 8,000 10,000 12,000

HIS

TORI

CAL

MODELED PPT

WRSI 2001-2009

0.0

2,000.0

4,000.0

6,000.0

8,000.0

10,000.0

12,000.0

- 2,000 4,000 6,000 8,000 10,000 12,000

HIS

TORI

CAL

MODELED PPT

WRSI Removing wet and normal years

-

2,000

4,000

6,000

8,000

10,000

12,000

- 2,000 4,000 6,000 8,000 10,000 12,000

HIS

TORI

CAL

MODELED PPT

ISPI Seasonal based on PPT (WRSI)

Series1

Series2

-

2,000

4,000

6,000

8,000

10,000

12,000

- 2,000 4,000 6,000 8,000 10,000 12,000

HIS

TORI

CAL

MODELED PPT

ISPI 6 month based on PPT (WRSI)

Series1

Series2

# 22

www.rmsi.com

Conclusions

A significant amount of uncertainty is associated with the

correlation weather parameter-MDR

– The seasonal parameters behave better than the weighted

parameters – counter intuitive

– Seasonal SPI based on WRSI PPT and Seasonal SPI based on

station Chimoio reconstruct best the last 9 years of production

The overall risk can be estimated with a reasonable of confidence

The loss in a specific year may be significantly miscalculated

Pursue WRSI and WRSI PPT

# 23

www.rmsi.com

Thank you !

24

Selection of the appropriate data source for Agriculture ... · Selection of the appropriate data...

Documents

Transcript of Selection of the appropriate data source for Agriculture ... · Selection of the appropriate data...