Pollution Research Paper

WATER CLARITY ASSESSMENT FOR A REGIONAL LAKE USING LANDSAT OPERATIONAL LAND IMAGER DATA: A CASE OF NALBAN LAKE IN EAST KOLKATA WETLAND, INDIA

PULAK PRITI PATRA, SOURABH KUMAR DUBEY, MALANCHA ROY, RAMAN KUMAR TRIVEDI, SANJEEV KUMAR SAHU AND BASUDEV MANDAL

Abstract

Landsat operational land imager (OLI) data and consequent laboratory measurements were used to predict water clarity for an inland lake within the East Kolkata Wetland, India (a Ramsar site). Total suspended sediment (TSS) and turbidity was considered as responsible parameters for assessment of lake clarity. The most suitable band ratio was identified by performing Pearson correlation analysis between water clarity concentrations and possible OLI band and band ratios from the ‘study points’. The OLI 4 band (636-673 nm) showed the best ‘r’ value, 0.96 and 0.89 in case of TSS and turbidity respectively. The two separate prediction models (using non-transformed and logarithmically transformed water clarity data) was developed by applying regression analysis between the band reflectance value of OLI4 and water clarity concentrations of the study points. The band reflectance value of the ‘validation points’ was given as input in the prediction model and model predicted dataset was considered as predicted water clarity parameters. The model predicted dataset exhibit lower standard error of estimates (SEE) with contemporaneous in situ measurements. The validation of the multi-temporal competence of the best models indicated that it is feasible to apply the linear regression model using OLI 4 band to estimate water clarity concentrations across the seasons in Nalban Lake without any in situ data. The water clarity mapping of the lake was then developed using the predicted dataset. This empirical study showed that Landsat 8 OLI imagery could be effectively applied for the mapping of TSS and turbidity for inland lakes.