On the uncertainties introduced by land cover data in high-resolution regional simulations

Abstract

We investigate the impact of implementing an up-to-date and detailed land cover dataset in high-resolution regional climate simulations. We used the Weather Research Forecast (WRF) model version 3.6.1 on a high horizontal resolution of 5 km × 5 km, with 29 vertical levels, covering mainland Europe. We performed simulations within the year 2050, using future Representative Concentration Pathway 8.5 mid-century projections, for 2 winter (January, February) and the 2 summer months (June, July) to investigate the seasonal dependency of the impact of the land cover datasets on and their interaction with the different meteorological conditions prevailing in summer and winter. We compare simulations using the CORINE Land Cover dataset (100 × 100 m) and the standard United States Geological Survey (USGS) (~ 1 × 1 km) land use data for the same periods. Our analysis shows that simulated meteorological variables (temperature at 2 m, wind speed, sensible and latent heat fluxes and PBL heights) differ significantly between the WRF simulations, linked to the land cover parameterization. We quantify and discuss the modelling uncertainties arising from surface-type classifications and motivate the use of high resolution, and continuously updated land use inventories in climate modelling, especially for future projections. Our findings are particularly important for the summer season and over large urban centers, and we strongly recommend the use of high-quality resolution land use data in modelling experiments studying heat waves in synergy with the urban heat island phenomenon and land–surface interactions.