Total soil carbon modelling along the altitudinal gradients in Eastern Himalaya, Arunachal Pradesh

To bridge the data gap of carbon reservoirs in the Arunachal Himalaya, estimation of carbon stock is of utmost importance. Despite being a major carbon reservoir, the high-altitude forests are lacking in such data mainly due to inaccessibility and rough terrain. The present study aimed to model the total soil carbon along the altitudinal gradient in the Eastern Himalayan region to understand the variation in soil carbon and other important soil variables in the ecosystem. The model offers predicting the readily available soil physico-chemical properties to predict total soil carbon in undisturbed forest ecosystems of Eaglenest Wildlife Sanctuary, Arunachal Pradesh, India. Soil samples were collected using stratified random sampling and analysed following standard methodologies. The XLSTAT application was used for partial least squares regression modelling. The results indicated an average annual total soil carbon content of 4.79±0.36% in tropical, 4.32±0.42% in subtropical and 3.88±0.35% in temperate zone. Notably, there was a significant decrease in microbial biomass carbon and soil inorganic carbon with increasing altitude. Highly accurate partial least squares regression prediction models were developed, with R² values ranging from 0.87 to 0.96, root mean squares error values from 0.29% to 0.59% and mean squared error values from 0.08% to 0.35%. These models will serve as valuable tools for assessing soil carbon stocks across different elevations, particularly in inaccessible areas. The study highlights the effectiveness of partial least squares regression models in predicting total soil carbon along altitudinal gradients and underscores the need to better understand ecosystem responses to environmental change. This information can be utilized by policymakers to gain insights into the important implications for REDD + + reporting, policy making and other relevant applications.