Improvements in lake water budget computations using Landsat data

Abstract

A supervised multispectral classification was performed on Landsat data for Lake Okeechobee's extensive littoral zone to provide two types of information. First, the acreage of a given plant species as measured by satellite was combined with a more accurate transpiration rate to give a better estimate of evapotranspiration from the littoral zone. Second, the surface area coupled by plant communities was used to develop a better estimate of the water surface as a function of lake stage. Based on this information, more detailed representations of evapotranspiration and total water surface (and hence total lake volume) were provided to the water balance budget model for lake volume predictions. The model results based on information derived from satellite demonstrated a 94 percent reduction in cumulative lake stage error and a 70 percent reduction in the maximum deviation of the lake stage.