Evaluation of the Sensitivity of the Weather Research and Forecasting Model to Parameterization Schemes for Regional Climates of Europe over the Period 1990–95

Abstract

The Weather Research and Forecasting model (WRF) is used to downscale interim ECMWF Re-Analysis (ERA-Interim) data for the climate over Europe for the period 1990-95 with grid spacing of 0.448 for 12 combinations of physical parameterizations. Two longwave radiation schemes, two land surface models (LSMs), two microphysics schemes, and two planetary boundary layer (PBL) schemes have been investigated while the remaining physics schemes were unchanged. WRF simulations are compared with Ensemble-Based Predictions of Climate Changes and their Impacts (ENSEMBLES) observations gridded dataset (E-OBS) for surface air temperatures (T2), precipitation, and mean sea level pressure (MSLP) in eight subregions within the model domain to assess the performance of the different parameterizations on widely varying regional climates. This work shows that T2 is modeled well by WRF with high correlation coefficients (0.8 , R , 0.95) and biases less than 48C. T2 shows greatest sensitivity to land surface models, some sensitivity to longwave radiation schemes, and less sensitivity to microphysics and PBL schemes. Precipitation is not well modeled by WRF with low correlation coefficients (0.1 , R , 0.3) and high root-mean-square differences (RMSDs; 8-9 mm day 21 ). Precipitation shows sensitivity to LSMs in summer. No significant bias has been observed in the MSLP modeled by WRF. Correlation coefficients are typically in the range 0.7 , R , 0.8 while RMSDs are in the range 6-10 hPa. MSLP output is sensitive to longwave radiation scheme in summer but is relatively insensitive to either microphysics or the choice of LSM. The optimum combination of parameterizations for all three state variables examined is strongly dependent on subregion and demonstrates the need to carefully select parameterization combinations when attempting to use WRF as a regional climate model.

Ammonia Concentrations Over Europe – Application of the WRF-Chem Model Supported with Dynamic Emission

“…Similar findings are reported by Kryza et al [43] for the area of Poland, where the air temperature bias is low in winter (mean bias -0.6 K), but where summer temperatures are significantly overestimated (up to +1.0 K). A bias in WRFcalculated air temperatures were also reported by Mooney et al [44] and Miglietta et al [37]. The model performs well at simulating wind speed for Europe, with mean bias not exceeding 0.5 m s-1 [37,[45][46][47], which is the second variable affecting ammonia emissions in this study.…”

Regional climate modeling on European scales: a joint standard evaluation of the EURO-CORDEX RCM ensemble

“…This indicates a contribution of the model-specific method to reduce simulated surface pressure to mean sea level, and the pronounced biases in the mentioned regions should not be overinterpreted. Still, the underestimation of mean sea-level pressure by several hectopascal over large parts of continental Europe particularly in wintertime seems to be a robust feature of the WRF experiments and is also described by Mooney et al (2013) in a sensitivity study of WRF in Europe.…”

“…The fact that the temperature bias range of the three WRF-11 experiments often corresponds to the bias range of the entire ensemble illustrates the uncertainty introduced by the choice of parameterizations and parameter settings (e.g., Bellprat et al, 2012a;Mooney et al, 2013). This, however, is not apparent for precipitation biases where the different WRF setups approximately agree on sign and magnitude of their bias.…”

Diurnal cycle of precipitation over the British Isles in a 0.44° WRF multiphysics regional climate ensemble over the period 1990–1995

“…A complete description of the WRF multiphysics ensemble used in this study can be found in Mooney et al (2013), and only a brief summary is included here for completeness. The ensemble consists of thirteen WRF simulations driven by ERA-Interim (Dee et al 2011) over the period 1989-1995.…”

“…The computational cost of such a high resolution over a large domain for an extended period when several parameterizations are being investigated is difficult to justify. As discussed in Mooney et al (2013), the model is allowed to spin up for 1 year and the period covering 1990-1995 is used in the analysis. Table 1 summarises the parameterizations used in each of the 13 simulations.…”