LINEAR REGRESSION PREDICTION MODELS FOR TURBIDITY IN KENTUCKY LAKE UTILIZING LANDSAT 7 ETM+ AND KLMP IN SITU DATA

Kentucky Lake was formed in 1944. Since 1988, it has been a focus of water quality monitoring by the Center for Reservoir Research’s Kentucky Lake Monitoring Program (KLMP). Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) data coincides with in situ data taken from the corresponding date. The goals of this research are to find if there is a correlation between band reflectance and turbidity and the best regression models for turbidity using Landsat 7 ETM+ imagery data along with in situ data collected concurrently by the KLMP in 2000 and 2001. All of the reflective bands as well as both thermal bands were evaluated with the exception of the panchromatic. Maximum R-square regression models and correlations were created using the Statistical Analysis Software (SAS). Bands determined to be significant for model use by the maximum R-square regression analysis were consistent. The correlation analysis showed the thermal bands (Band 6LO and Band 6HI) and Band 3 to have high correlations with turbidity. The results of the best regression models suggest that it may be possible to accurately predict turbidity, for Kentucky Lake, using Landsat 7 ETM+ data.