Climate Impacts of Introducing Vaisala Radiosondes in the U.S. Observing Network

In 1995, the U.S. National Weather Service (NWS) changed to use of the Vaisala RS80 radiosonde at 21 stations. Prior to the change, most of these stations were using VIZ-B radiosondes although a few were still using the SDD radiosonde. Such changes may lead to shifts in the time series of monthly means that result from this change and make the determination of climate trends difficult and any trend estimates more uncertain. Our approach was to compare the time series at each Vaisala station with that of a neighboring station that did not change. The time series of differences between each station pair was used to identify any discernable shift in the mean.

Monthly mean temperature (T), dewpoint (Td), relative humidity (RH) and geopotential height (Z) were extracted from the dataset of monthly statistics, known as MONADS derived using radiosonde observations from the Comprehensive Aerological Research Data Set (CARDS, Eskridge et al., 1995). Monthly mean anomaly time series for the period 1991-1998 were used. At each station, for each of the four variables (T, Td, RH and Z) and for five pressure levels (850mb, 700mb, 500mb, 100mb and 50mb), difference time series were calculated of the anomaly difference between the station and its reference station at either 0000 UTC and 1200 UTC. A shift in the mean in the difference time series ( Vaisala - reference) was quantified by comparing the mean difference over the two year period prior to the change to that over the subsequent two years. We will show the vertical, geographic, seasonal and diurnal variations in the mean shift of the difference time series.