FOR2131-P7

Data Assimilation for Improved Characterization of Fluxes Across Compartmental Interfaces (Phase 2)

01/10/2017 to 30/09/2020

The Research Unit addresses the use of in-situ and remote sensing observations to confine parameters and state variables of integrated terrestrial model systems on the scale of mesoscale hydrological catchments in order to improve state estimation and predictions. In the scientific community these efforts are subsumed under inverse modelling and/or data assimilation. While data assimilation has achieved considerable success in the atmospheric sciences and has long become the backbone of numerical weather prediction, inverse modeling for parameter estimation is more common and developed in hydrological sciences. The Research consists of seven research (P1-P7) and one central project (C1), supported by the coordination project Z, and follows the hypothesis that improving the parameters and state variables in integrated models by data assimilation will lead to better and more consistent state analyses, model parameters and predictions across all compartments in comparison to data assimilation of individual compartments.

Project P7 investigates (a) whether and how groundwater observations and the estimation of aquifer properties improve the overall state estimation and prediction by data assimilation in coupled atmosphere-surface-subsurface systems, and (b) whether and how updating states and properties of neighbouring compartments improves the state estimation of the groundwater compartment. We expect added value by a consistent treatment of fluxes across aquifer boundaries in all neighbouring compartments within a coupled modeling framework. But incorporating neighbouring compartments increases computational efforts, and adds conceptual and parametric uncertainty besides mismatching spatial and temporal scales between the compartments. The project will develop the best strategy for updating groundwater states and parameters both in a weakly and strongly coupled data assimilation framework based on TerrSysMP-PDAF.