Modelling multi-scale atmosphere and land-surface interactions

As land-surface properties are heterogeneous over a broad range of length-scales, surface-induced fluxes are heterogeneous too. Representing land-surface heterogeneity and the corresponding fluxes is a challenging task in numerical prediction of weather and projection of climate.

In this work, we introduce the approach of 'para-real' ensemble modelling to investigate the dynamic effect of land-surface heterogeneity. We perform a large ensemble of high-resolution simulations using the Weather research and forecast model (WRF-ARW-LSM). The para-real simulation ensembles are externally forced by a reanalysis of a real case in spring 2013, but become exposed to different synthesized surface patterns (SP) generated as quasi-fractal Brownian surfaces (quasi-fBs) with exact control of the dominant wave length and fractal persistence.

The focus of this study is on the three inter-related land-surface and atmosphere coupling mechanisms--the thermodynamic coupling, aerodynamic coupling, and hydrological coupling. For each mechanism, a corresponding surface property is identified, namely surface albedo (α) for thermodynamic coupling, roughness length (z₀) for aerodynamic coupling, and soil type (s_t) for hydrological coupling. For each surface property, we generate a set of quasi-fBs with different dominant length scale and fractal persistence. In our para-real ensembles, the original fields of the surface properties are replaced by the quasi-fBs, for which we estimate the control parameters from the original data, i.e., the probability density distribution of the original data matches that of the quasi-fBs which eliminates the flux aggregation effect and allows us to focus on the dynamic effect.

We find, first, a strong impact of the length scale of the surface forcing on the intensity of coupling: while the dynamic effect of surface heterogeneity significantly impacts the state of the atmospheric boundary layer for all cases investigated,  the impact of the surface signal on the atmospheric state  grows with the length-scale of the surface heterogeneity. Second, we demonstrate that larger fractal persistence of the surface signal also strengthens the atmosphere--surface coupling. Third, the qualitative impact of the surface forcing is shown to depend on time, which eliminates the possibility of a simple linear forward propagation of the surface signal; there is strong sensitivity to the diurnal cycle, in particular with respect to the horizontal wind components: The maximum intensity of atmosphere--surface coupling (measured in terms of correlation) is found around noon for the atmospheric temperature, and some hours later (in the early afternoon) for water vapor. Fourth, among the different surface forcing investigated, we find that the heterogeneity of soil type is the most important to the atmospheric state--surface exchanges and its signal are detected in the atmospheric water-vapor up to 2km height; in particular, the soil-type pattern with the smallest length-scale causes a doubling of cloud-water above 500m height  whereas no impact on the bulk atmospheric state is found for patterns with other length-scales and fractal persistence or forcing of other surface variables. This illustrates the key part that hydrological coupling plays in connecting the atmosphere to the surface, and it underlines the relevance of improved hydrological process-level representation for improved parameterization of the coupled land--atmosphere system.