| dc.description.abstract | Radiosonde data from weather balloons are critical because they are essential inputs for numerical weather prediction models and are used for climate research. However, radiosonde programs are costly to maintain, in particular in the remote regions of the Arctic. The climate of this data-sparse region is poorly understood and forecast data assimilation procedures are designed for more general, global applications. Thus, observations may be rejected from the data assimilation because they are too far from the forecast model expectations. Here, we evaluate how radiosondes launched twice daily (0 and 12 UTC) from Summit Station, Greenland, (72.58°N, 38.48°W, 3210 masl) influence the European Centre for Medium Range Weather Forecasting (ECMWF) operational forecasts from June 2013 through May of 2015. Several stakeholders use this information for forecasting and research, such as investigating the changing climate, including that of the Greenland Ice Sheet (GIS). Therefore, a statistical analysis is conducted to determine the impact of these radiosonde observations on the forecast model, and the meteorological regimes that the model fails to reproduce will be identified. First, the frequency of the deployment of radiosondes is calculated, and approximately 90% or more of the deployments were successful. Next, the climatology of the GIS temperature inversions is calculated, and it is found that ECMWF underestimates temperature inversions. The assimilation rates of meteorological variables that influence the variability of the inversion strength – temperature, specific humidity, are lowest in the layer where the inversion strength is located. Therefore, there is a likely relationship between model simulations with and without these meteorological variables assimilated, and the underestimation of this particular meteorological regime. Meanwhile, winds – another influence on the variability of temperature inversion – had high assimilation rates, likely caused by being the only observational source for the model. Based on a statistical assessment, the magnitude of the mean model bias for simulations without temperature and specific humidity data assimilated is significantly reduced in comparison model with radiosonde observations assimilated. Overall, the radiosonde data provided by the Summit Station radiosonde program improves meteorological forecasts in ECMWF. This likely improves the underestimation of surface features on the GIS, such as shallow temperature inversions, which leads to improved understanding of the climatic and atmospheric dynamics of the GIS. | en |
