Project Details
Projekt
Integrating Remote Sensing Information in the Cosmic Ray Neutron Sensing Signal for Soil Moisture Modelling and Sampling
Subject Area
Hydrogeology, Hydrology, Limnology, Urban Water Management, Water Chemistry, Integrated Water Resources Management
Term
since 2018
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 357874777
The results of the first phase of the research unit Cosmic Sense proved that there is a strong link between CRNS-derived Soil Moisture Content (SMC) with active and passive remote sensing signals. This was shown for different UAS data as well as Sentinel-1 and -2 satellite information, especially for grassland. However, there is still a high degree of uncertainty regarding the spatial and temporal variability of the different measurement systems (CRNS and Remote Sensing), which need to be entangled in further detail to find the best scales of transferring the CRNS footprints to a larger spatial coverage with remote sensing imagery. Although the empirical investigations show plausible results, they are still relying on regional specifics and require an additional physically-based modelling to understand the relations, especially regarding the spatial distribution (horizontally and vertically) of different types of hydrogen pools. This can lead to a clear separation of the soil moisture signal from other hydrogen pools in the footprint. Both empirical and physical models rely on measured input variables that cover the whole range of the observed target variable under different connected environmental conditions. Finding such representative sampling points is crucial for a future catchment scale SMC monitoring. High resolution remote sensing combined with new AI methods data can take up this challenge. Our scientific objective is to understand distribution and temporal changes of SMC in large-scale landscapes with their diverse mosaics. Therefore, we have to improve the understanding of neutron transport at areas of heterogeneous biomass distribution. In doing so, we want to couple Radiative Transfer Models (RTMs) and neutron transport models such as URANOS. RTMs deliver information on plant water content and biomass proxies and can be used, when inverted, to be included in the URANOS calculation of neutron transport. With this modeled information on vegetation water content we aim to identify characteristic locations for CRNS probes within the Joint Field Campaign (JFC). This will be implemented in novel deep learning (DL) methods, which consider temporal and spatial characteristics of the SMC in relation to a predictor data set consisting of remote sensing and other spatial data. In the final consequence, the DL methods can be used to derive SMC variability maps. CRNS measurements from planned joint field campaigns and roving campaigns will be used to validate and compare the derived representative sampling as well. A further step towards high-resolution SMC monitoring at the catchment scale will be the use of this extended CRNS dataset from the JFC to create large-scale, high-spatial-resolution SMC map for the Elbe catchment using deep learning approaches. These results will be the basis for a comparison of SMC patterns at the different spatial resolutions from meter to the kilometer scale using UAS and global remote sensing SMC products.
DFG Programme
Research Units
