tratto da: ProClim

link alla notizia:
http://fm.proclim.ch/FMPro?-db=news....d=33292&-find=

Discrepancies between surface and lower troposphere temperature trends resolved

General news
Detection of errors in data correction and artefacts resolves major discrepancy between model results and observation



The discrepancy between model results and observations of tropospheric temperature trends especially in the tropics has been one of the most important open questions concerning the quality of climate models. While the understanding of physical processes in the atmosphere and climate models both suggested a warming of the tropical troposphere slightly stronger than at the surface, observations both from satellites and radiosondes showed a slight tropospheric cooling, while the surface has warmed considerably. In the last few years, other groups have calculated trends from satellite data (Microwave Sounding Unit MSU) with different methods, resulting in higher global warming trends (between 0.13 and 0.24ºC per decade) similar to the trend at the surface (0.17ºC per decade). However, the different data processing and correction methods remained disputed.
Now new research published in ScienceExpress on 11 August by Mears and Wentz, Santer et al. and Sherwood et al. demonstrates, that most probably errors in processing of satellite data and biases in radiosonde measurements are responsible for the discrepancies in the trends. New analyses of satellite data by Mears and Wentz show tropospheric warming trends slightly higher than at the surface. It has been detected, that the artifical trend due to the shift of the local equator crossing time in the traditional data set had been corrected erroneously in the wrong direction (opposite sign) in the UAH (Spencer/Christy) data set. After correction of this error, all satellite data sets now show tropospheric warming in the tropics. Updated Tropospheric temperature trends are now at 0.12 (UAH), 0.19 (Mears and Wentz), 0.20 (Fu et al., Nature 429:55) and 0.24ºC (Vinnikov and Grody, Science 302:269) per decade.
It has been known that radiosonde data is biased by daytime solar heating of instruments. This bias most probably has decreased over time due to development of better correction methods and thus produces an artificial cooling trend. Sherwood et al. show that there is a cooling trend in daytime compared to nighttime soundings which is much higher than can be explained by physical processes and than is observed at the surface. This probably artificial trend can explain the difference of radiosonde and surface trends. This conclusion is supported by the fact, that models and radiosonde observations are consistant at a monthly time scale and the difference only appears on the decadal timescale.