Estimation of vegetation stress in the mangrove forest using AVIRIS-NG airborne hyperspectral data

Abstract

This study assessed the potential and demonstrated the applicability of hyperspectral images from AVIRIS-NG data in mapping the health condition of the Lothian Island estuarine mangrove forest with rich floral diversity for the years 2016 and 2018. Nine vegetation indices covering every aspect of the health status were considered and a weighted overlay analysis was performed to generate health maps from AVIRIS-NG data. Discriminant Normalized Vegetation Index (DNVI) was produced from Sentinel-2 data for both the years 2016 and 2018 to validate the results from AVIRIS-NG data. The analysis showed considerable negative trends in plant pigment contents indicating poor vegetation health and increased stress in the entire mangrove forest of the Lothian Island. Furthermore, red-edge analysis at four specific points helped in validating our health maps generated from both AVIRIS-NG and Sentinel-2 data. Nearly 56% of the total area exhibits evidence of increasing stress within the span of the study years. 35% of the area has shown no change in the existing stress conditions in 2018. Both new plantation and species-specific resilience to stress appear to be the key factors for the sustenance of the existing mangrove population in the Lothian Island of the Indian Sundarbans.