Spatio-temporal variability of Turbidity over Ukai Reservoir, by using RS2/R2A LISS III Satellite datasets

Abstract

Turbidity is an optical determination of water clarity. It is one of the most important optically active water parameter to assess the water quality through the remote sensing observations. Turbidity measurements come from suspension of sediment such as silt or clay, inorganic materials, or organic matter such as algae, plankton and decaying material. Turbidity and total suspended matter often overlap each other. However, it is not a direct measurement of the total suspended materials in water. Instead, as a measure of relative clarity, turbidity is often used to indicate changes in the total suspended solids concentration in water without providing an exact measurement of solids. Through remote sensing we can monitor the turbidity in large water bodies, rives, coastal areas etc. An algorithm has been developed to estimate the turbidity (in NTU: Nephelometric Turbidity Unit) over inland waters (Ukai reservoir) using empirical relationship between normalized Green and Red bands (NDTI : Normalized Difference Turbidity Index) of Resourcesat-2 and Resourcsat-2A Linear Imaging Self Scanning-III (RS2 and R2A LISS-III) dataset. Derived algorithm shows a strong coefficient of determination (R2 = 0.97) with the in-situ turbidity measurements. The field measurements were carried out over Ukai reservoir on 27-28th March 2018, where synchronous in situ water leaving reflectance and turbidity were measured. Model was derived between in situ measured turbidity and NDTI derived from spectral reflectance of band 2 (Green) and band 3 (Red) of RS2 and R2A LISS-III. The model was applied to derive the turbidity maps of Ukai reservoir for pre-monsoon (March, April and May months) season during the period 2012 to 2018. Overall turbidity ranges from 1.47-20 NTU during the field data collection of pre-monsoon season and overall scene derived turbidity ranges are between 2 – 33 NTU. The highest observed turbidity value was more than fourteen times greater than the lowest value that shows the natural variability within the reservoir for the same season. Remotely sensed data sets can increase the abilities of water resources researchers and decision making persons to monitor waterbodies more effectively and frequently.