CALCULATION OF EVAPOTRANSPIRATION FOR A WATER BUDGET USING LANDSAT DATA AND AERIAL PHOTOGRAPHS: A COMPARATIVE STUDY

Effective integration of landuse/landcover (LULC) data provides viable information to any spatial modeling approach of the study of the environment, including calculations of hydrologic budgets. Use of moderate resolution satellite imagery (Landsat) provides a reliable source of spatio-temporal data for a watershed to conduct long-term studies, whereas aerial photographs provide a higher resolution but infrequent data source. The objective of this research was to compare effects of resolution of data, both spatial and temporal in the water budget calculation. This was done by utilizing satellite digital image processing techniques to derive landuse classes at different resolutions (Landsat 30m and Southwest Florida Water Management District [SWFWMD] photo-interpreted aerials) to be incorporated into a simple spreadsheet-based hydrologic water budget model. The study area is located in selected sub-watersheds of Peace River Watershed in South Florida, USA. The two sub-watersheds are chosen for their diverse LULC characteristics and are called Joshua Creek and Payne Creek sub-watersheds. The Joshua Creek is located in DeSoto County, and the Payne Creek basin is located in Polk, Hardee and Hillsborough Counties. Joshua Creek encompasses ~120 square miles and has agricultural and rural land usage, while Payne Creek includes ~122 square miles and has phosphate mining as well as rural and agricultural landuse. Landsat 5 TM images were obtained for summer and winter of 1985-2000 (path 16, row 41) to characterize winter and summer LULC for the entire duration of the study. Due to infrequent availability of aerial coverages from SWFWMD (compared to Landsat), only the corresponding years of SWFWMD LULC were used in our spreadsheet to determine the effects of resolution on the water budget. These LULC data were then used to calculate evapotranspiration (ET) losses as well as to quantify water input from the irrigation sources. The ET estimates were obtained for each LULC type based on the seasonal literature value of ET and the total area coverage of a given LULC type. These results were then input into the spreadsheet-based water budget model to determine ET losses from the differing land cover types. The LULC were further reclassified to quantify irrigation input based on the agricultural crop types. Average seasonal irrigation rates were used for each crop and were multiplied to their respective area coverage to estimate the irrigation input. These resultant ET and irrigation input values were then incorporated in a spreadsheet model to calculate water budget for winter, summer and annual (water year) for the 16 years. The final water budget model results were then validated against actual US Geological Survey (USGS) gagging data for stream flow.