A new global gridded radiosonde temperature data base and recent temperature trends

Abstract

Abstract. We present a new analysis of global radiosonde temperature data. From 1979 onwards, the data from the Australasian region have been corrected for instrument-related discontinuities with the help of comparisons with collocated retrievals from satellite-based Microwave Sounding Units (MSU) and metadata: in future work, adjustments will be applied worldwide and extended to earlier years. The data are stored as monthly anomalies from a 1971-1990 reference period on a 5 ø latitude x 10 ø longitude grid at 8 levels from 50 hPa to 850 hPa. Seasonal and annual temperature anomalies have also been created on a 10 ø x 20 ø grid using an eigenvector reconstruction method to filter noise. Latitude-height profiles of zonal-mean temperature changes since the 1960s show significant cooling in the lower stratosphere, especially in middle and high latitudes of the Southern Hemisphere, but the cooling over Australasia is less than shown by unadjusted data. Warming dominates the troposphere but is not a maximum in the tropical upper troposphere. In the annual mean, tropospheric warming is greatest around 45øN and possibly in the data-sparse high latitudes of the Southern Hemisphere. Because of incomplete coverage, radiosonde data require substantial interpolation before they can be used to assess regionallyor globally-averaged changes, or to verify model simulations of climatic variations. We briefly describe our gfidding and interpolation techniques to this end. Finally we use the interpolated, gridded, adjusted data to estimate zonal-mean and hemispheric temperature changes in the troposphere and lower stratosphere in recent decades, comparing results with those using unadjusted data. We also compare the temperature changes with those observed at the surface.

Difference in radiosonde temperature trend for the period 1979–1998 of MSU data and the period 1959–1998 twice as long

“…Differences in radiosonde temperature trend with length of record have been considered qualitatively by Karl (1994) and Parker et al (1997). This paper does so quantitatively, and with estimates of the significance of the difference.…”

Assessing Bias and Uncertainty in the HadAT-Adjusted Radiosonde Climate Record

“…We excluded stations and levels that did not have at least five years of data containing at least three seasons for each of the three decades within the climatology period. We also excluded Indian stations because they have been found to be difficult to homogenize (Thorne et al 2005a) and problematic for subsequent analysis of long-term trends (Parker et al 1997;Lanzante et al 2003). The 0000 and 1200 UTC soundings were combined to produce a comprehensive merged dataset of 509 stations, 50 of which are within the (20°S-20°N) tropics.…”

Global and regional climate in 1999

“…In northern spring the pattern was similar but northern Asia was cool and there was a small area of unusual warmth centred over Labrador. In northern summer, parts of eastern Europe and the Middle East experienced Parker et al (1995).…”

“…l(a)). This is based on air temperature data collected from over 1000 land-based weather stations (Jones et al 1999), plus sea surface temperatures (SSTs) (Parker et al 1995) measured from the Voluntary Observing Fleet of about 7000 ships and 1000 buoys. The year was much less warm on a global average than 1998, which was the warmest in the instrumental record and 0.58degC warmer than average ( Fig.…”