Validation and climatology of ERA5 using in-situ observations in the Arctic

The Arctic is currently warming at a rate three to four times greater than the global average. This region also has important implications for sea level rise and the global radiation balance. Therefore, the Arctic requires special attention in the study of meteorological conditions, which is often accomplished through the use of reanalysis models. Reanalysis models are weather forecasting systems that use a combination of observational data and numerical models to produce a best guess of atmospheric variables at high resolution and for a large period. When data from these reanalysis models are used, it is crucial to know the uncertainty and potential biases that are associated with each parameter. In this study, we used data from 17 meteorological stations located in the Arctic to validate the performance of the ERA5 reanalysis model. The surface meteorological variables include air temperature at 2m, relative humidity, surface pressure, wind speed/direction, and shortwave downward radiation. We then performed a climatology of the Arctic region using ERA5 data. The validation results show that the temperature (R2 mean = 0.94 with σR2 = 0.02) and surface pressure (R2 mean= 0.98 with σR2 = 0.02) predictions are very accurate. However, the predictions of relative humidity are worse with large variations (R2 mean = 0.42 with σR2 = 0.20). Finally, the performance of the wind speed prediction (R2 mean = 0.49 with σR2 = 0.12) is quite variable depending on the topographic situation of the station and the validation of shortwave downward radiation flux gives satisfactory results (R2 mean = 0.87 with σR2 = 0.03). Compared to other studies, the temporal resolution used (hourly in this study versus daily or monthly in other studies) shows a strong influence on the quality of the wind and shortwave radiation predictions and no influence on the temperature evaluation. The climatology of the surface variables shows the spatial and seasonal variations of the stations distributed in all regions of the Arctic. This study, unprecedented in the higher number of years used, the number of variables considered and its high temporal resolution, provides the scientific community with recommendations on the confidence to be placed in each of the surface variables produced by ERA5. The climatology provides a good overview of common weather conditions in the Arctic that can be used to contextualize future measurement campaigns.