Imaging Spectroscopy of Carotenoid Content of Himalayan Broadleaved and Needle-Leaved Canopies from Space
摘要
The Himalayan forests are experiencing several environmental stresses such as prolonged dry winters, summer droughts, human-induced fires, insect outbreaks and regional climate warming. Carotenoid pigments play a key role in photoprotection of plants and serve as a measure for early detection of environmental stress. The detection of carotenoid signals using remote sensing is challenging due to the overriding reflectance of the chlorophyll pigments in the healthy vegetation. Present study evaluates the potential of PRISMA hyperspectral satellite imagery to estimate and subsequently map canopy carotenoid content (Ccar) of two extensive, broadleaved and needle-leaved forest forming tree species viz., Shorea robusta (sal) and Pinus roxburghii (chir pine), respectively of the Western Himalaya. The study uses a genetic algorithm based partial least square regression (GA-PLSR) to identify the optimal wavebands and spectral indices sensitive to carotenoids based on the in-situ measurements and PRISMA derived variables. Carotenoids exhibited strong absorption peaks in the blue and green regions of the visible spectrum (441 nm, 449 nm, 478 nm, 515 nm, 530 nm, 571 nm), irrespective of the species. The Carotenoid Reflectance Index (CRI550) was found to be the most sensitive index for both the species (sal: R2 0.73, RMSE 0.14 g m− 2; chir pine: R2 0.63, RMSE 0.20 g m− 2). Among the PLSR models, the model based on reflectance bands performed the best with R2 > 0.8 and RMSE of 0.12 g m− 2 and 0.10 g m− 2 for sal and chir pine, respectively. Higher carotenoid content observed for needle-leaved compared to broadleaved species indicates its ability to cope with harsh winters. This is the first reporting of estimation of canopy carotenoid of the forests using in-situ and hyperspectral satellite remote sensing data.