Using the vegetation signal to determine the root zone soil moisture from radar remote sensing

Two outstanding challenges persist in measuring soil moisture using microwave remote sensing. Firstly, even L-band instruments are only capable of measuring surface soil moisture (top 0-5cm), while root zone soil moisture is the true variable of interest because it represents water available for transpiration. Secondly, even surface soil moisture can only be reliably measured under a light/medium canopy. Vegetation is usually considered a barrier to soil moisture retrieval using microwave remote sensing. As the vegetation water content increases, the opacity of the canopy reduces the sensitivity of the observations to the surface soil moisture. Recent research has demonstrated that there is a detectable difference between the ascending and descending tracks of the ERS scatterometer data and that the most significant differences correspond to locations and periods of water stress. The goal of this research is to use this difference in backscatter to infer quantitative information on plant water status and hence root zone soil moisture. Backscatter depends on both the dielectric and geometric properties of the canopy. However, we assume that the observed difference in backscatter is due to diurnal variation in dielectric properties. A diurnal cycle in the dielectric constant of tree components and crops has been observed in the field by other researchers. Our working hypothesis is that the manner in which water is stored in the plant when it is stressed controls the dielectric constant. The theoretical models of the dielectric behaviour of vegetation will be briefly reviewed. Results will be shown from a numerical study in which the difference in backscatter due to diurnal variation in vegetation water was determined, and its magnitude compared to the error budgets for current and future radar-based soil moisture products. Finally, preliminary results from laboratory experiments to measure the dielectric response to variations in plant water content will be presented.