MULTIPARAMETER RELATION WITH SOIL ORGANIC CARBON STOCK AT INDIAN WETLANDS

Abstract

Carbon Storage can be of two types i.e., 1) Organic Carbon storage and 2) Inorganic Carbon Storage. Research studies depicts that in arid, semiarid area and coastal area with high alkalinity and salinity in soil is of great importance to store Carbon in inorganic form. However, Soil Organic Carbon (SOC) is a widely accepted indicator to predict the soil health and carbon storage in soil. Soil Organic Carbon stock (tonnes/hector) data (0 – 30 m) from the latest release of SoilGrids (May 2020) has been used in this study to correlate the SOC stock with multiple indices derived from Landsat 8 band dataset. The study areas selected here are East Calcutta Wetland (Ramsar site no. 1208) & Sundarban Wetland (site no. 2370) which are inscribed in a single Landsat 8 tile (WRS Path/Row: 138/45) and are of international importance for rich biodiversity and climate change. Several studies by numerous scholars have been proved that SOC stock has a deep correlation with NDVI and crop phenology. Numerous physical, chemical, and biological parameters regulate the carbon cycle and this complex interaction is hard to predict with a few laboratories analysis and can be cost-extensive or site accessibility for sampling is often denied for numerous bio-geophysical constraints. However, spatiotemporal changes of parameters derived from satellite data can be a good option to specify the sampling area with more certainty and cross validation for lab analysis. Built-up area, cloud cover and presence of surface water plays important role to regulate the prediction of subsurface SOC stock. Hence Multiparameter (NDVI, NDBI, NDWI, LST & SMI) analysis, variation, and best possible correlation (r, Pearson correlation coefficient) is established here which can be used for further research progress in these study areas. Sampling in the study areas is done by Google Earth Engine to generate 1000 and 19000 random sampling points at the site no. 1208 and 2370, consequently. ArcGIS 10.8.2 software is used in this analysis to estimate the band statistics and interpolation of SOC data input, data extraction, resampling, and raster calculation for derived indices (NDVI, NDBI, NDWI, SMI) and LST. Predictability of SOC stock using Ordinary Kriging method with spherical variogram has been established in this study through the variation trend of covariates like Root Mean Square (RMSE) and R2 are shown here where extracted point dataset is used as measured and interpolated point dataset are used as predicted value in SOC stock analysis. Linear Regression method is also to compare statistical outcome between the two study areas. Monthly variation of the derived multiparameter with SOC stock data is iv the key concern throughout this work. Excel and Tableau software are used to analysis the data in this study. The results showing the variation of statistical metrices and linear regression has done to fit the line with a minimum R2 values for both extracted and interpolated SOC data points at both study sites. In this analysis, the monthly variation of multiparameter in case of EKW is higher than Sundarban Wetlands which is mainly due to anthropogenic activity at Ramsar site 1208. Though both the wetlands are of different types, but they have a good on an average amount of SOC stock as per the data collection. The importance of wetland in environmental balance is the key moto which has tried to be established in through this study.