Systematic Differences in Aircraft and Radiosonde Temperatures

W ind and temperature data from radiosondes and aircraft are main sources of in situ information to data assimilation systems for NWP (see Table 1 for acronym definitions). Currently, there are about 150,000 automated aircraft reports per day, used at the NCEP, with approximately 3 times as many temperature observations than there are from radiosondes. NCEP operational analyses use automated aircraft report data known as AMDAR, which include automated aircraft data from U.S. aircraft referred to as ACARS. Automated aircraft reports from outside the United States are referred here as NUS-AMDAR for clarity. For more information on automated aircraft data, see Moninger et al. (2003) and Painting (2003). A newer type of U.S. automated aircraft data, known as TAMDAR (Moninger et al. 2006), are not yet used in operational models at NCEP, and therefore are not analyzed. However, discussions of some past tests with these data and suggestions for possible field tests are given. Zapotocny et al. (2000) showed that with the data assimilation system of the NCEP Eta Model, winds and temperatures from ACARS data had a significant impact on analyses and short-range forecasts, but generally less than that of radiosondes for 12-h forecasts; however, no verification scores were presented. Graham et al. (2000) studied the impact on 60-h forecasts of mean sea level pressure with the Met Office model, with various data types withheld for 15 h prior to the analysis. These impact tests showed that both radiosondes and aircraft data were important, with winds generally more useful than temperatures, but the aircraft temperatures were important in some oceanic cases. Cardinali et al. (2003) analyzed the SYSTEMATIC DIFFERENCES IN AIRCRAFT AND RADIOSONDE TEMPERATURES Implications for NWP and Climate Studies