Leaf Chlorophyll Concentration Estimation Using Absorption Spectroscopy of AVIRIS-NG for a Mangrove Forest in India
摘要
Chlorophyll concentration is one of the important biochemical properties of vegetation as it relates to photosynthetic activity and health. The amount of chlorophyll in a vegetation canopy indicates the physiological status or the health condition. Compared to other terrestrial ecosystems, mangroves are highly productive, so there is a need for a better understanding of the dynamics of carbon sequestration by monitoring their health and nutrition status for ecological conservation and restoration processes. In spite of many ecosystem services, limited research has been conducted concerning mangrove chlorophyll assessment due to the challenges of field sampling. The majority of the chlorophyll assessments in mangroves are being executed with the help of remote sensing data-derived vegetation indices (VIs). However, they are site or species-specific, which prohibits a universal adaptation. Our study quantifies leaf chlorophyll concentration (LCC) distribution using the Airborne Visible InfraRed Imaging Spectrometer—Next Generation (AVIRIS-NG) hyperspectral imagery and field observed dataset for the Bhitarkanika National Park (BNP), a mangrove ecosystem of India. This study aims to predict the LCC utilizing absorption features such as absorption band depth (ABD) as a predictor variable. This was calculated using continuum removal techniques and further predicted using machine learning (Random Forest, RF). This study identifies the red-edge region (676–722 nm) as the prominent part of the electromagnetic spectrum that is useful for predicting LCC. Our model achieved an acceptable accuracy (R2 = 0.82, RMSE = 0.34) and comparable validation statistics (R2 = 0.44, RMSE = 0.38), despite on-field logistic constraints in LCC measurements. This study demonstrated a protocol for a rapid estimate of biochemical variables using (AVIRIS-NG) hyperspectral imagery.