SPATIO-TEMPORAL EUTROPHICATION ASSESSMENT OF KALLAR KAHAR LAKE USING REMOTE SENSING DATA

Abstract

Assessment and mapping of the spatio-temporal variations in water quality is a problem for the concerned agencies. It's timely monitoring and assessment is a key for various conservation strategies as well as best management. This study was undertaken to determine trophic state of KallarKahar Lake through the application of remote sensing data by retrieving spatial variability in water transparency depths and chlorophyll-a (Chl-a) concentration, both major water quality parameters. Statistical models were developed on the basis of correlation between Landsat ETM+ image data and water transparency depths and Chl-a concentration, with prevailing mapping technique. Surface water samples were collected from 21 accessible locations within the lake. Transparency depths were measured in field and samples were analyzed in water testing laboratory for Chl-a content determination. Image classification was performed to extract water only pixels, and the conversion of respective DN to water leaving radiance values. These values were used in multi-variate regression analysis to find the correlation between dependent (transparency depth, Chl-a) and independent (different band combinations) variables. Resultant equations were used for spatial modeling of dependent variables after determine the significant positive correlation (R2 = 0.707 for Chl-a and R2 = 0.608 for transparency depths). The resultant surfaces were compared with spatial geo-statistical interpolated surfaces and found that Landsat image based model performed well with better accuracy and less RMSE (4.73 for Chl-a and 1.8 for transparency depth using remote sensing methodology that increases to 8.94 for Chl-a and 2.94 for transparency depth using geo-statistical interpolation technique). Two previous Landsat ETM+ imageries of same season were then analyzed for temporal assessment and steady increase in Chl-a content was observed that constitute towards the eutrophic state of lake. The results show that remote sensing data based analysis is time and cost effective methodology, that doesn't require frequent field measurements for regular monitoring of lake for conservation, restoration and proper management.