Assessment of Tree Density, Tree Cover, Species Diversity and Biomass in Semi-arid Human Dominated Landscape Using Large Area Inventory and Remote Sensing Data

Abstract

It is essential to develop methods for estimating species diversity and biomass in human-dominated mosaic landscapes to minimise uncertainty. The level of detail provided by very high-resolution satellite imagery enables the precise mapping and monitoring of individual trees and tree patches in trees outside forests. This work is the first of its kind and attempts to estimate tree density, tree cover, species diversity, and biomass from a comprehensive survey and very high-resolution remote-sensing data. This research compared the census of the entire tree population over a 900-ha site (local landscape) and a 15,142-ha site (regional landscape) in the Sri Sathya Sai district of Andhra Pradesh. This study mapped 47,054 tree individuals that cover a land area of 1.64% in a regional landscape. The most dominant species based on the importance value index are Tamarindus indica, Mangifera indica, Cocos nucifera, Prosopis juliflora, and Pongamia pinnata. Estimated tree density indicates about 3 trees per hectare in regional and local landscapes, respectively. Among the 42 inventoried tree species, 22 were wild. Analysis shows evergreen trees are dominating over deciduous trees with 88% of tree density. Alpha diversity of the local landscape reaching up to H′ = 1.93. The findings show that the maximum above-ground biomass is 40.61 tonnes/ha at one site, while it is relatively low at the remaining sites. Since no sampling is involved, the estimates derived from census data are not subject to sampling error, leading to high precision in the results. The spatial approach used in the study combines field-based data collection with the advantages of remote sensing technology to provide a detailed assessment of tree resources in rural landscapes.