Estimation of Mangrove Carbon Stocks by Applying Remote Sensing and GIS Techniques

Abstract

This paper presents a methodological framework for predicting C stock in Avicennia marina stands in the Thane creek of Mumbai. This methodology combines ground-based (GB) and remote sensing (RS) approaches for C stock estimation. RS based approach use Normalized differential vegetation index (NDVI), Light use efficiency (LUE) and Photo-synthetically Active Radiation (PAR) as the most important parameters for C stock estimation. The sensitivity of NDVI values to aerosols, water vapor and ozone was removed using the 6S radiative transfer code. The difference in NDVI values before and after atmospheric correction was assessed using student’s t-test and was found to be statistically significant. The total carbon stock of the area was observed to be about 39.7188 t/ha. The bias estimation between C stock calculated using allometry and RS approaches confirmed the prediction accuracy and validated both the techniques (R 2 = 0.964 and bias = 0.915 %). The paper, thus, reports a statistically robust framework, which is a combination of the RS and GB approaches, and can be used for estimating the biomass and carbon stock of any ecosystem. This framework is especially effective where forest inventory data is unavailable, the site is geographically inaccessible or harvesting of mangroves or other trees is prohibited.