|
Spatiotemporal Characterization of Wetlands in the Ethiopian Highlands using Landsat, SRTM and MODIS data
Alemayehu
Midekisa, South Dakota State University, alemayehu.midekisa@sdstate.edu
(Presenter)
Michael
Wimberly, South Dakota State University, michael.wimberly@sdstate.edu
Gabriel
Senay, US Geological Survey Earth Resources Observation and Science (EROS) Center, senay@usgs.gov
Tropical wetlands provide important ecosystem services but also serve as habitats for malaria-transmitting mosquitoes. Wetland maps are critical tools for management decisions including vector control strategies. In this study we will address the following three research questions focusing on the Amhara region of Ethiopia: (1). Will integration of Landsat TM/ETM+ and SRTM topographic derivatives improve the mapping accuracy? (2). Are there differences in intra-annual (seasonal) variability of biophysical land surface characteristics over various lad use land cover types? (3). Are there temporal trends of biophysical land surface characteristics of wetlands? To address these questions, we developed a wetland map of the study region using a random forests decision tree model. We tested three models with (1) topographical indices from 30 m SRTM data, (2) individual reflectance bands and multispectral wetness indices from Landsat TM/ETM+ imagery, and (3) combined spectral and topographic data. Wetland inventory data were obtained by photo interpretation of high-resolution imagery in Google Earth and ground surveys. Integrating reflectance data, multispectral wetness indices, and topographic derivatives provided the most accurate wetland maps. We explored seasonal and interannual variability of Actual Evapotranspiration (ETa), Enhanced Vegetation Index (EVI), and Land Surface Temperature (LST) over a ten year period (2001-2010) for three land use/land cover classes (cropland, permanent wetland and seasonal wetland “musk”) using MODIS time series data. We observed differences in land surface phenology after the major rainy season across three land use/land cover classes: permanent wetlands, croplands, and seasonal wetlands. Our findings also indicated that over ten years (2001-2010), there was a declining trend in ETa for both permanent and seasonal wetlands while cropland showed an increasing trend. Information about the spatial extent and temporal dynamics of wetlands in Amhara region can help to monitor malaria risk because wetlands near human population are potential reservoirs of mosquitoes.
Presentation:
2013_Poster_Midekisa_65_27.pdf (3695k)
Presentation Type: Poster
Session: Poster Session 2-A
(Wed 11:00 AM)
Associated Project(s):
Poster Location ID: 65
|