Experimental Cases of an Arctic System Reanalysis

The importance of establishing an Arctic System Reanalysis has been recognized by the National Oceanography and Atmospheric Administration (NOAA); and an initiative is undergoing at the University of Alaska Fairbanks, the Ohio State University, the Colorado State University, and the University of North Dakota. This presentation is to report the results from our experimental case study of an Arctic reanalysis system. Our past studies have been focused on the impact of model resolution, customized background error for three-dimensional variational (3DVAR) data assimilation, and satellite data utilization. Here in this study, an experimental reanalysis is conducted on two case studies; and independent verification and intercomparison with existing reanalyses (e.g., ERA-40 and NCEP/NCAR reanalyses) is performed. Based on preliminary studies and experiments, an Arctic regional reanalysis modeling system, including a regional numerical weather prediction model, a combination of data assimilation approaches, and a verification system, has been setup for a pan-Arctic domain. The domain covers all the regions north of 55 N. The model resolution is setup at 30 km horizontal grid spacing and 41 terrain-following sigma levels. In this system, conventional surface and upper-air observations and satellite retrieved data are assimilated via methods including Newtonian nudging to both an analysis and observations and 3DVAR approach. To examine the Arctic reanalysis system, experimental cases of two three-day periods, one for winter (14-17 January 1998) and one for summer (20-23 July 1998), have been chosen for our case study (potentially, a full year run has been planned). The testing case studies have been chosen from 1998 because independent verification of the reanalysis results can be performed by utilizing the SHEBA (Surface Heat Budget of the Arctic Ocean) program observational data. The reanalysis is compared with the ERA-40 and the NCEP/NCAR reanalysis, based on the independent verification and intercomparisons. The results will be shown at the conference.