Consistent hydro-meteorological large eddy simulation of water budgets in mesoscale catchments: Overkill or necessity?

Abstract:

The modeling of water budgets in mesoscale catchments is confronted by several challenges: (1) Quantitative precipitation estimation (QPE) is erroneous due to the lack of high-resolution observing systems, (2) soil-vegetation-atmosphere exchange processes and feedbacks are not included in the forcing of offline land-surface models, (3) land-surface models as an essential component of weather and climate simulations include only poor representations of water budget components, e.g., runoff and streamflow, (4) the influence of vegetation on the surface energy balance closure including transpiration is oversimplified, (5) crucial processes such as mesoscale circulations induced by land-surface heterogeneity and orography are often not sufficiently resolved, (6) a stringent methodology for the parameterization/regularization of the sub-grid variability of land surface processes is missing. This calls for the development of hydro-meteorological models, which simulate the physical-chemical-biological processes of the soil-vegetation-atmosphere system in a consistent manner. We present the development of a corresponding model system based on WRF-NOAH-MP, which can be considered as contribution to the WRF-HYDRO program. The model is operated over a large domain in Europe with convection-permitting resolution (3km). Initial conditions are optimized by a spin-up run continued by a rapid update cycle with a resolution of 10 min - 1 h in combination with 3-dimensional variational data assimilation (DA) of a huge set of observations. These include surface in-situ networks, radio soundings, global positioning system, aircraft measurements, and satellite products. A recent highlight is the assimilation of reflectivity and Doppler wind measurements of the European radar network, which led to a substantial improvement of QPE and quantitative precipitation forecasting (QPF). Thus, this model system can be applied for reanalyses of the regional weather and climate down to the meso-gamma scale. Current limitations and ongoing improvements of the DA system are presented and discussed such as the development and incorporation of a new model forward operator for polarization radar. By setting up a series of nests, the model simulations can be downscaled to 100 m, hence, reaching the scale of large-eddy simulations (LES). This is a significant advance because in contrast to many previous downscaling efforts, upscaling of fluxes to the size of mesoscale catchments becomes possible while resolving the most important physical processes determining 2-dimensional flux fields. Furthermore, due to the series of simulations with different resolutions, new strategies for the parameterizations of key processes such as land-surface exchange and atmospheric turbulence can be developed. It is also interesting to study the influence of meso-gamma scale circulations on the surface fluxes. First results of LES are presented with respect to simulations over the Attert catchment in Luxembourg. An example is depiected in Fig.1. The 2-m atmospheric temperature field with its color scale, a vertical cross-section of specific humidity, and grey isosurfaces of boundary layer clouds are shown. The linewise structures in the temperature field are due to horizontal convection rolls, a special mesoscale organization of the flow, which clearly influence the surface temperature and fluxes. The vertical plane indicates turbulent eddies responsible for the vertical transport of water vapor resulting in soil-vegetation-atmosphere feedback in the convective atmospheric boundary layer. Even though these results demonstrate the scientific applications of LES forced to realistic weather conditions, there are still several topics, which need to be addressed towards the goal of consistent hydro-meteorological modeling: A correct representation of soil properties and hydraulic coefficients as well as calibrated run off and streamflow schemes are required. The vegetation and the heterogeity of the land cover must be well represented including leaf area dynamics and transpiration. The status of our research in this direction is introduced and an outlook is given with respect to future developments and applications.