From the spring of 1995 to the present the number of radiosonde reports from the former USSR dropped by about 41 % from 196 to 80 per day. There are several reasons for the reduction; the most important being budget reductions. An obvious question is, "Has the loss of these data had an impact on the skill of NWP predictions over the Northern Hemisphere or subregions therein?"

Several possibilities exist for examining this question. One would be to conduct Observing System Experiments (OSEs) where data are denied from the assimilation to determine what would be the impact. A simpler, less costly technique proposed would be to examine existing forecast statistics to see if an impact could be discerned. In March 2000, the World Meteorological Organization's Commission for Basic Systems organized an Expert Team to use routine verification statistics in an attempt to determine if the loss of the Russian RAOBS had a noticeable impact on the NWP forecasts produced by the main operational centers. Early results of the WMO study suggest that the signal, if any, will be difficult to detect.

NCEP, therefore, under the auspices of the North American Observing System (NAOS) program, has decided to conduct a simple OSE using the NCEP Reanalysis/Reforecast system for at least the 1994-1995 cold season (December, January, and February). NWP runs will be with the full complement of Russian RAOBs available during the selected period, and only those available at sites existent during the 1999-2000 cold season.

This paper reviews the studies conducted to examine the impact of the loss of the Russian RAOBs, and preliminary results from the NCEP Observing System Experiment. The implications of the results to the makeup of the Global Obsering System (GOS) will also be discussed.