Different aggregations of Corine land cover for WRF simulations of a pre-Alpine valley

The land cover defines what is on the earth’s surface. For atmospheric research, land cover is a leading influencer of surface exchange rates, roughness length and the surface heat flux. Nowadays, with finer resolution, the importance of land cover being correct increases. Especially over heterogeneous terrain, as gradients created by short-distance variability can influence local meteorology. For processing purposes, land covers maps group all land covers in classes based on certain conditions. Every land cover map has different classes and conditions to serve their purpose best.

The land cover map USGS is the default map for weather simulations with WRF. Since Pineda et al., 2004 developed a method to convert the Corine Land Cover (CLC) to a WRF readable format, is CLC available for simulations over Europe. CLC is a land cover map that focuses primarily on the EU and a few collaborating countries.

Our objective is to investigate the influence of land cover on valley winds. Our study focuses on the central part of the pre-Alpine Durance valley. Simulations exist out of a domain with three nested domains and verified with the KASCADE 2013 data. In previous simulations with USGS, land cover is in the domain near the Cadarache site almost homogeneous. Results and representation improved with the introduction of the CLC as land cover map.

This CLC study will act as an update and extension of a previous study of the area. In this study, we use an updated version of CLC, and we look at different methods of aggregation of the CLC data to quantify their effects on simulation results. CLC has 44 categories and 3 levels of detail. We compare direct aggregation to an aggregation, which takes into account the available detail levels.

First comparisons of aggregation methods show a mismatch ranging from 1.4 % in at 300 m resolution to values around 11 % for resolutions of 3 km and coarser.