The U.S. radiosonde network is in the process of changing to the Radiosonde Replacement System (RRS). In preparation for this change, a strategy was developed to ensure climate continuity over the RRS transition. Two aspects are addressed by the strategy. The first focuses solely on the differences between the new and old observing systems using dual sonde observations at selected stations. The dual sonde flights will take place at a limited number of stations carefully selected for both climate considerations, such as spatial coverage, and station histories indicating that the stations have good continuity of their historical data. These two factors identified 17-19 NWS radiosonde climate stations: four stations in Alaska, nine in the CONUS, five in the tropical Pacific and one NWS station in the Caribbean.

Based on the variability of radiosonde relative biases in World Meteorological Organization radiosonde intercomparison studies data, approximately 200 dual sonde flights will be needed at each station in order to adequately assess the bias in temperature caused by the RRS transition. Two hundred flights spread out over all four seasons are also enough dual sonde flights for the 95% confidence limit on the impact of the RRS transition induced discontinuity on a CONUS averaged tropospheric time series to be less than 0.05°C. The 0.05°C threshold was selected because it is one third of the controversial difference between two global satellite-derived tropospheric temperature time series adjustments for the discontinuity associated with the NOAA-9 polar orbiting satellite.

The dual sonde continuity flight portion of the strategy can assess the bias in the RRS system relative to the earlier radiosondes. The second aspect of the strategy is to employ reference radiosondes to help determine the bias of the RRS to true atmospheric temperature and humidity. With significant investment in development, the cost of reference radiosonde would be expected to come down dramatically over the course of a few years. However, even with lower costs, reference radiosonde flights would still be too expensive to fly at more than a few stations. To make the best use of this new source of high quality data, the reference radiosonde flight schedule would be determined to optimize not only comparisons with RRS sondes but also satellite calibration and basic research into atmospheric processes such as upper tropospheric water vapor feedback.