Remote sensing has proven its usefulness in many fields and applications. A critical point consists in the fact that the remotely sensed imagery needs to be converted into relevant data of geophysical interest, such as soil moisture, leaf area index, evapotranspiration, snow and ice cover, and wetland delineation. Within this framework, retrieval techniques ranging from physically-based modelling over data-driven approaches (e.g., neural networks, neuro-fuzzy modelling) to the more simple linear regressions, applied to optical, thermal, passive and active microwave, or lidar data play a key role and their development, refinement, and validation is important for assessing uncertainty on the retrieved values.
By integration of the remote sensing information into land surface process models, the simulations of the process model can be improved and forced to mirror a more appropriate and realistic land surface state. Uncertainties in the surface parameter inversion process as well as in the simulations obtained from the process model should be considered in this context. Modern data assimilation theory provides methods for optimally merging remote sensing observation with land surface process models by considering the process model and observation uncertainties. This allows for the estimation of the most probable surface state.
The session aims at the evaluation of the current state of art of retrieval techniques for a wide range of remote sensing products, uncertainty assessment on the retrieval, and data assimilation techniques of remotely sensed products for land surface modelling purposes. This includes examples of remote sensing data assimilation projects as well as data fusion approaches to combine the information content of various heterogeneous data sources.
This session therefore welcomes contributions on the following particular issues:
- retrieval algorithms for land surface parameters from remote sensing;
- assessment techniques for uncertainty estimation on the retrieval;
- disaggregation of remote sensing derived land surface information;
- fusion of different (complementary) data sources to improve surface parameter retrievals;
- remote sensing data assimilation techniques, including technical background and comparison of different assimilation strategies;
- comparative studies, evaluating the benefit of different assimilation techniques;
- examples of remote sensing data assimilation into land surface process models.