REMOTE SENSING AND PHYSIOLOGICAL BIOMARKERS FOR DROUGHT STRESS IN FOREST TREES AND CROPS

Abstract

An increased risk to the productivity of forests and farms throughout the world, drought stress is proving to be an even greater risk hence necessitating more methods of detecting and gauging them as rapidly and precisely as possible.  As evidence by this study, a combination of remote sensing and physiological biomarkers can be used to locate and quantify drought stress in trees and crops growing in forests.  The model identified early indications of drought in multiple spatial scales and made use of hyperspectral image, thermal remote sensing and satellite based vegetation indices (NDVI and NDWI).  Meanwhile, the remotely observed drought indicators were confirmed with the help of ground-truthing the physiological data of plants, such as chlorophyll fluorescence, leaf water content, and stomatal conductance.  To guess how severe a drought will be we grouped the spectral information through highly developed machine-learning approaches--particularly convolutional neural networks. It had very accurate results (R 2 = 0.87) with low values of RMSE.  The scheme of the technique workflow is depicted in Figure 1 and consists of the pieces of data fusion, preprocessing, spectrum analysis, and validation.  The findings indicated that strong relationships existed between remotely sensed indices and physiological responses in case of drought.  The model was also able to identify the drought-tolerating and drought-intolerating plant species correctly which is valuable data under adaptive water management and precision agriculture.  The given study advances the science of drought detection by offering a scaleable, non-invasive and data-driven technique in integration of remote sensing technologies with physiological science.  The findings have far reaching consequences on climate change resilience of ecosystems, the enhancement of irrigation practices, and climate change resilience of ecosystems.