Study Area & Data

Introduction

The Wildlife Conservation Society in conjunction with Clark University undertook this project to guide preservation efforts by analyzing the habitat trends for Chilean Flamingos (Phoenicopterus chilensis). This work represents the first known endeavor to incorporate solely remotely sensed data on a regional area to identify and measure the essential characteristics of flamingo habitats. These important flamingo habitats were prioritized based on their size, health, habitat size, proximity to anthropogenic threat, and ability to withstand the effects of climate change, most notably the El Niño/Southern Oscillation (ENSO).

The objective of this research is to answer the following questions:

1) Which Habitat Lakes will prove to be resilient given the potential impacts of climate change.

2) How have Flamingo habitats changed in their seasonality? 3) How stable are the lakes? 4) How healthy are the lake systems? 5) What is the expected human threat? 6) How much habitat space do these lakes provide? 7) What is climate changes expected impact on these areas?

Methodology In order to find the most persistent lakes for the flamingos, six factors were taken into consideration. 1. Lake Stability Lake area was calculated monthly to evaluate the lake stability. If the sizes

of a lake didn’t change much, the lake was ranked higher. If the sizes of a lake varied a lot, the lake would be considered less suitable for the flamingo.

2. Habitat Size Flamingo Habitats are shallows up to 50cm in depth. Shallows with an ideal

depth for the Chilean Flamingo were measured monthly using the Near Infrared reflectance. Lakes with large, persistent shallows are considered better habitats.

3. Predation Risk Anthropogenic activity is one of the greatest threats to flamingo

populations. As a proxy, the distance from the lakes to the nearest towns or cities was calculated.

4. Lake Health Normalized Difference Vegetation Index (NDVI) on the shoreline was used

as a proxy for determining lake health. 5. Seasonal Changes Evaluate the seasonal changes of the precipitation data with Earth Trend

Modeler in IDRISI. The percentage of change were calculated between 1982 to 2010.

6. Long Term Stability Principal Components Analysis was run on Sea Surface Temperature

anomalies using NOAA's Optimum Interpolation version 2 (SST) to determine a correlation between ENSO and precipitation anomalies using the Global Precipitation Climatology Project (GPCP) dataset.

Habitat Conservation Prioritization for Phoenicopterus Chilensis

Lake Stability

Predation Risk

Sainan Zhang, Devin Dworkin, and Kevin Butler Graduate School of IDCE

Clark University, Worcester, MA.

Chart of Ranking Lakes

Discussions

References: · Baldasarre and Arengo 2000. A review of the ecology and conservation of Caribbean Flamingos in Yucatan, Mexico. Waterbirds 23 (Special Publication): 71-79

· Boyle et al 2004. Landsat TM inventory and assessment of waterbird habitat in the southern altiplano of South America. Wetlands Ecology and Management 12: 563-573

· Bucher 1992. Population and Conservation Status of Flamingos in Mar Chiquita, Cordoba, Argentina. Colonial Waterbirds 15(2) 179 – 184. · Caziani et al 2001. Waterbird Richness in Altiplano Wetlands of Northwestern Argentina. Waterbirds 24(1): 103-117. · Caziani et al 2007. Seasonal Distribution, Abundance, and Nesting of Puna, Andean, and Chilean Flamingos. The Condor 109(2): 276-287

· Cezilly et al 1995. Interannual variation in Greater Flamingo breeding success in relation to water levels. Ecology 76(1) 20-26

· Fan et al 2011. Shallow water depth retrieval from space-borne SAR imagery. Journal of Oceanography 67(4) 405-413

· Hurlbert, Loayza and Moreno 1986. Fish-flamingo-plankton interactions in the Peruvian Andes. Limnology and Oceanography 31(3) 457 - 468

· Mascitti 2001. Habitat Changes in Laguna de Pozuelos, Jujuy, Argentina: Implications for South American Flamingo Populations. Waterbirds 24(1): 16-21. · Mascittiand Bonaventura 2002. Patterns of Abundance, Distribution and Habitat Use of Flamingos in the High Andes, South America. Waterbirds 25(3): 358-365 · Mascitti and Castanera 2006. Foraging Depth of Flamingos in Single-species and Mixed species Flocks at Laguna de Pozuelos, Argentina. Waterbirds 29(3): 328-334. · Meehl et al, 1999. Trends in Extreme Weather and Climate Events: Issues Related to Modeling Extremes in Projections of Future Climate Change. Bulletin of the American Meteorological Society. · Sathyendranath et al 2001. Remote sensing of phytoplankton pigments: a comparison of empirical and theoretical approaches. International Journal of Remote Sensing 22 (2 and 3) 249-273 · Ouma and Tateishi 2006. A water index for rapid mapping of shoreline changes of five East African Rift Valley lakes: and empirical analysis using Landsat TM and ETM+ data. International Journal of Remote Sensing 27 (15) 3153-3181. · Prigent et al 2007. Global inundation dynamics inferred from multiple satellite observations, 1993 – 2000. Journal of Geophysical Research, 112 · Vuille 1999. Atmospheric circulation over the Bolivain Altiplano during dry and wet periods and extreme phases of the Southern Oscillation. International Journal of Climatology 19:1579-1600.

Note the 0 score for both Mar Chiquita and for Lago Titicaca, which have populated centers in close proximity. A conservation effort based on ecosystems services has potential. Lagunas Pululos and Celeste have 0 scores for Lake Stability, meaning they have dried up at least once in the past ten years. Additionally, when these lakes do inundate, there is little competition for food. We assume that there is an abundance of food because of the response in NDVI along the shores, making them a prime feeding ground for the Chilean Flamingo. Based on this assumption, lakes with scores of 0 in the area index should be the focus of an increased monitoring effort, which would give a clearer picture of how flamingos respond to drought and flood events.

The study region consists of the Andean altiplano which extends from southern Peru, into Bolivia and northern Argentinian lowlands, representing 88% of the known flamingo habitats (Caziani, 2007). A list of coordinates for wetlands habitats was provided by the WCS. Of the approximately 430 wetlands, 21 were selected for further study.

Data Source Range (values) Resolution Range (time)

MODIS NDVI GLCF -1 – 1 250m 2000-2011

MODIS Land Cover GLCF 0-100 1km 2000-2011

MODIS NIR GLCF 0-255 250m 2000-2011

S. America Wetlands Land Cover Map American Museum Natural History 0/1 Vector 2000

World Cities ArcGIS Qual Vector 2008

Sea Surface Temperature NOAA OI v2 K 1 km 1982-2010

Precipitation GPCP mm/day 2.5° 1982- 2010

Data

1. Las Encandenadas 4 2. Las Encandenadas 5 3. M1 4. Sur de Maggiolo

6340 6360 6380 6400 6420 6440 6460 6480 6500 6520 6540 6560

J M S J M S J M S J M S J M S J M S J M S J M S J M S J M S J M S

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Squ

are

d

Kilo

met

ers

Mar Chiquita y Banados Rio Dulce

0

5

10

15

20

25

J M S J M S J M S J M S J M S J M S J M S J M S J M S J M S J M S

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Squ

are

d K

ilom

eter

s

Laguna Celeste

Area of wetlands was measured monthly. Because the goal is to find persistent lakes, the range of the fluctuation in area was divided by the maximum area.

Squ

are

d

Kilo

met

ers

Flamingo feed in the shallows of lakes and wetlands. Shallows were measured monthly and classified into 4 groups based on the percentage of months that a given area within the lake is at an ideal depth for the Chilean Flamingo.

Proximity to populated centers was based on the Euclidean distance from the closest towns to each lake, and was standardized by dividing each of the distance values by the longest distance.

Laguna Loriscota Lago Titicaca

Vegetation along the shores was measured using the Normalized Difference Vegetation Index. The chart shows an example of lake Lago Titicaca and Laguna Celeste (left) and the mean values of NDVI (right) for all the lakes.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

J M S J M S J M S J M S J M S J M S J M S J M S J M S J M S J M S

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

NDVI 2001 - 2011

Lake Titicaca Laguna Celeste

Lake Health

Habitat Size

Long Term Stability Seasonal Changes

For each lake the difference between the two curves was calculated and standardized based on the 1982 precipitation rates to see to what degree the seasonality has changed. The index scores were calculated with the area difference divided by the area of earlier year. The following are four examples among all the lakes.

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

Jan

Feb

Mar

Ap

r

May

Jun

Jul

Au

g

Sep

Oct

No

v

Dec

Salinas de Ambargasta

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Jan

Feb

Mar

Ap

r

May

Jun

Jul

Au

g

Sep

Oct

No

v

Dec

Laguna Negro Francisco

1982

2010

-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Jan

Feb

Mar

Ap

r

May

Jun

Jul

Au

g

Sep

Oct

No

v

Dec

Hedionda Norte

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

NDVI Mean Lake Name Dis(km)

Laguna Carucue 4.5

Llancanelo 28

Lago Titicaca 0

Lago Poopo (Lago Uru Uru ) 4.8

Lago Loriscota 34

Laguna Colorada 29

Laguna Celeste 23

Laguna Pululos 35

Pozuelos 0

Laguna Negro Fransisco 22

Salinas Grandes 0

Salinas de Ambargasta 11

Mar Chiquita y Banados Rio Dulce 0

Argentina Group 1.6

To undesrtand how El Nino affects each lake, anomalies in Sea Surface Temperature were correlated with anomalies in Precipitation from 1982 – 2010.

Id Lake Name R

1 Lago Loriscota 0.1043

2 Lago Titicaca 0.0609

3 Lago Uru Uru and Lago Poopo 0.0357

4 Laguna Colorada 0.1042

5 Laguna Celeste 0.0463

6 Laguna Pululos 0.0765

7 Pozuelos 0.0464

8 Laguna Negro Fransisco 0.0304

9 Salinas Grandes 0.0353

10 Salinas de Ambargasta 0.0507

11 Mar Chiquita y Banados Rio Dulce 0.1035

12 Llancanelo 0.1152

13 Laguna Carucue 0.0828

14 Argentina Group 0.1202

Ranking Lakes with six factors *Scores are on a scale of 0 – 10, with 10 as ideal.

Name Lake Stability Habitat

Size Predation Risk Lake Health

Seasonal Changes

Long Term Stability

Final Score

Mar Chiquita, Banados Rio Dulce

9.80 10 0 8.23 9.37 8.96 46.37

Laguna Loriscota 6.00 0.07 9.79 5.84 8.38 8.95 39.06

Lago UruUru and Poopo 4.87 7.734 1.38 6.23 8.53 9.64 38.41

Lago Titicaca 9.51 1.37 0 8.40 9.17 9.39 37.86

Llancanelo 8.27 0.28 7.926 4.32 6.93 8.84 36.59

Salinas de Ambargasta 6.27 0.17 3.22 6.16 9.40 9.49 34.73

Laguna Colorada 7.41 0.36 8.32 1.76 5.55 8.95 32.38

Argentina Group 6.02 0.53 0.46 8.28 8.22 8.79 32.33

Pozuelos 8.05 0.52 0 3.48 8.88 9.53 30.48

Salinas Grandes 5.18 1.40 0 6.03 7.85 9.64 30.12

Laguna Caruce 3.72 0.01 1.29 6.77 8.96 9.17 29.93

Laguna Pululos 0 0.01 10 2.01 7.47 9.23 28.73

Laguna Negro Fransico 4.45 0.13 6.42 1.83 5.99 9.69 28.54

Laguna Celeste 0 0.004 6.62 1.30 8.82 9.53 26.29

0

10

20

30

40

50

60

Final Score

All the six factors were given the same weights and added up to get the final scores ranging from 0 to 60 with 60 as ideal. 8000

10000

12000

14000

16000

18000

20000

J A J O J A J O J A J O J A J O J A J O J A J O J A J O J A J O J A J O J A J O J A J O

2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Lake Titicaca Area Lake Lago Loriscota Area

Lake Laguna Carucue Area Lake Llancanelo Area

Lago Uru Uru and Lago Poopo Salinas de Ambargasta

Mar Chiquita y Banados Rio Dulce Argentina Group

Pozuelos Laguna Celeste

Laguna Colorada Laguna Pululos

Laguna Negro Fransisco Salinas Grandes

0

0.5

1

1.5

2

2.5

3

Jan

Feb

Mar

Ap

r

May

Jun

Jul

Au

g

Sep

Oct

No

v

Dec

Llancanelo