Diurnal differences in vegetation dielectric constant as a measure of water stress

Abstract

Currently, vegetation is considered a barrier to soil moisture retrieval by both passive and active remote sensing missions. Microwave emission and backscattering of vegetation is driven by the vegetation dielectric constant, which is a function of vegetation water content. The latter is a measure of root zone water availability. Understanding the variation in dielectric properties of vegetation will contribute to soil moisture retrieval using microwaves in vegetated areas. This study presents a unique dataset of the diurnal pattern of the leaf dielectric properties, which was linked to vegetation water content and water stress. Using a microstrip line sensor, in-vivo dielectric property measurements were conducted on three maize leaves (leaf 8, 10 and 12) from 8 to 19 October 2012. A correlation was found between the resonant frequency of the microstrip line and the leaf water content of maize. This showed that a decrease of leaf water content during the day led to an increase of the resonant frequency. Water stress was quantified by calculating the evaporation deficit and by measuring the soil water tension at 30cm and 50cm depth. It was found that the diurnal difference in resonant frequency of the sensor at leaf 8 increased in similar fashion as the soil tension and evaporation deficit, which indicates a correlation between water stress and vegetation dielectric properties. The upper leaves 10 and 12 responded differently to increased water stress. The diurnal difference in resonant frequency of the sensor at leaf 10 and 12 decreased or was non-existent. The dielectric measurements revealed the complex reaction of vegetation to water stress and pointed out many opportunities for further research. The water-cloud model was used to demonstrate the impact of changing water content at different frequencies and polarizations. For L-,C-,X-,Ku- and Ka-band the sensitivity of radar backscatter to soil moisture and vegetation water content was modeled. This showed that at L-band, for low volumetric soil moisture (