Modeling chlorophyll-a and turbidity concentrations in river Ganga (India) using Landsat-8 OLI imagery

Rivers, one of the most complex ecosystems are highly dynamic and vary spatially as well as temporally. Chlorophyll-a (Chl-a) is considered one of the primary indicators of water quality and a measure of river productivity, while turbidity in rivers is a measure of suspended organic matter. Monitoring of river water quality is quite challenging, demand tremendous efforts and resources. Numerous algorithms have been developed in the recent years for estimating environmental parameters such as chlorophyll-a and turbidity from remote sensing imagery. However, most of these algorithms were focused on the lentic ecosystems. There is a paucity of algorithms for rivers from which water quality variables can be estimated using remotely sensed imagery. The primary objective of our study is to develop algorithms based on Landsat 8 OLI imagery and in-situ observations for estimating of Chl-a and turbidity in the Upper Ganga river, India. Band reflectance images from multispectral Landsat-8 OLI pertaining to May and October 2016, and May 2017 were used for model development and validation along with near synchronous ground truth data. Algorithms based on Band 3 (R-2=0.73) proved to be the best applicable algorithm for estimating chlorophyll-a. The best algorithm for estimating turbidity was found to be log (B4/B5) (R-2=0.69) based on band combinations (individual band reflectance, band ratio, logarithmically transformed band reflectance and ratios) tested. The developed algorithms were used to generate maps showing the spatiotemporal variability of chlorophyll-a and turbidity concentration in the Upper Ganga river (Brijghat to Narora) which is also a Ramsar site.