Vegetation Optical Depth (VOD) is a parameter in microwave radiative transfer models that describes the attenuation of microwaves in the vegetation layer. In theory, VOD is depending on vegetation water content (VWC), above-ground biomass (AGB) and live fuel moisture content (LFMC). Additionally, relationships between VOD and leaf area index (LAI), productivity, biomass and vegetation water status were found by past studies. VOD has also been used to assess changes in biomass, vegetation isohydricity and tree mortality.
However, VOD is not directly measured. Additionally, various VOD products exist that differ by the sensor used (i.e. AMSR-E/2, SMOS or SMAP), the retrieval algorithm (e.g. LPRM or MT-DCA) and also by the microwave frequency (Ku-, X-, C-, or L-band). Therefore, substantial challenges remain to infer vegetation properties and related ecological dynamics from the derived VOD signal.
Here we aim to assess several VOD products with respect to ecological interpretability and to quantify the co-varying sensitivities of high- and low-frequency VOD products to vegetation properties such as AGB, LAI and LFMC on a global spatial scale and over multiple years. We use VOD products from different sensors and wavelengths but all retrieved with the LPRM algorithm. Specifically, we use the VODCA datasets for shortwave VOD and the L-band VOD from SMOS and SMAP. The relationships and sensitivities of these products to the vegetation properties LAI, AGB and LFMC are examined using Generalized Additive Models differentiated by land cover types.
First results show remarkably small differences in the sensitivity of short-and long-wavelengths to AGB, LAI and LFMC. At monthly time scales, VOD and LFMC show a linear relationship for both wavelength domains in croplands. But the relationships of VOD with LFMC in forests and AGB are non-linear and non-monotonic. VOD and LAI are correlated at low LAI but surprisingly VOD saturates at higher LAI (in both wavelength domains).
Our results suggest that local or short-term changes in VOD cannot be easily linked to vegetation properties. In conclusion, the effects of vegetation on microwave signals needs to be further studied especially with respect to short-term processes in vegetation in order to fully harvest the potential of long-term global VOD datasets for ecosystem studies.