Comparing Temperatures from Different Monthly-Mean Datasets
How closely matched are available datasets of upper troposphere and lower stratosphere temperatures? Climate models are increasingly being used to investigate changes due to volcanic aerosols, ozone depletion, subsonic and proposed supersonic flights near the tropopause. The tropopause refers to the boundary between the troposphere, the region of the atmosphere where clouds and rain occur, and the stratosphere, where the protective ozone layer lies. If the models' results are to be believable, their mean temperatures should correspond to actual values. For example, if a model's tropopause temperature is too warm, it will likely overestimate the amount of water vapor from commercial subsonic aircraft exhaust flowing into the stratosphere. This incorrect result would then create errors in modeled stratospheric heating and chemistry.
Various temperature datasets are being used by individual modeling groups to judge models' performances. As many regions are difficult to observe, not all of these datasets necessarily agree. A similar problem has been recognized for other near-tropopause datasets, such as ozone concentrations and high-altitude clouds. A comparison of datasets enables modelers to evaluate their model results against a broader range of collated observations and to better understand the uncertainties in the datasets themselves.
We have compared four widely used monthly-mean temperature data sets, focusing on the upper troposphere/lower stratosphere region. Each of these temperature datasets suffers from its own calibration, instrumentation, or data coverage issues. Our comparison suggests where such shortcomings can degrade climate model evaluations.
Mean temperatures from four datasets — the Microwave Sounding Unit (MSU) data, the GFDL/Oort Radiosonde data, the COSPAR International Reference Atmosphere (CIRA) mix of data, and the new 13-year National Center for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) Reanalysis Project's mix of many types of data — agree extremely well in the lower stratosphere. The top and middle panels of the figure show the MSU and NCEP/NCAR lower stratospheric brightness temperature maps. Brightness temperature is an average of atmospheric temperatures over a layer in the atmosphere, for example, a temperature representing the temperatures across the layer between 13 km and 24 km. Differences between NCEP/NCAR and MSU 18 km-centered temperatures are under 2°C year-round across the tropics and 5°C in southern winter polar latitudes. Artificial land-ocean outlines of roughly 2°C do appear in maps of the NCEP/NCAR lower stratospheric brightness temperature between 30°S-30°N. These artificial outlines may be due to sparseness of inputted radiosonde data at 13-24 km heights that are not completely corrected by satellite coverage in the NCEP/NCAR data-integration process.
Climate modeling groups are having to judge their models' temperatures in regions that are sparsely observed. Since a model's mean temperatures control the atmospheric moisture and circulation, which in turn influence many other aspects of a simulated climate, mean temperatures need to be right. Our work compares different datasets used to examine a climate model's "star performer" — temperature — at high altitudes that are difficult to observe.
Kalnay, E., M. Kanamitsu, R. Kistler, W. Collins, D. Deaven, L. Gandin, M. Iredell, S. Saha, G. White, J. Woollen, Y. Zhu, M. Chelliah, W. Ebisuzaki, W. Higgins, J. Janowiak, K.C. Mo, C.Ropelewski, J. Wang, A. Leetmaa, R. Reynolds, R. Jenne, D. Josep 1996. The NCEP/NCAR 40-yr reanalysis project. Bull. Amer Meteorol. Soc. 77, 437-471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
Shah, K., and D. Rind. 1995. Use of microwave brightness temperatures with a general circulation model. J. Geophys. Res. 100, 13841-13874.
Shah, K., and D., Rind. 1998. Comparing upper tropospheric and lower stratospheric temperatures: Microwave sounding unit, radiosonde, COSPAR International Reference Atmosphere, and National Center for Environmental Prediction/National Center for Atmospheric Research reanalysis monthly-mean climatologies. J. Geophys. Res. 103, 31569-31592.