Observation and numerical climate model prediction do not always match when it comes to atmospheric temperature trends in Earth’s tropics. Some skeptics have focused on the mismatch, and the question of a systematic problem in the models has been raised. Recently, Flannaghan et al. set out to find the source of the problem.
The researchers studied the impact of two commonly used, but subtly different, data sets of sea surface temperature on multiple climate models—the HadISST1 and the Hurrell data sets. Both are based on measurements from ships and buoys as part of the Comprehensive Ocean-Atmosphere Data Set project, but after 1981 the data sets diverge following the onset of including measurements from satellite instruments using different algorithms. The team plugged both data sets into the Geophysical Fluid Dynamics Laboratory’s High Resolution Atmospheric Model (HiRAM) and the National Center for Atmospheric Research’s Community Earth System Model (CESM) and found that sea surface temperature trends are the biggest driver of uncertainty—more important even than which model was used.
The group discovered that several years around 1982/1983 were particularly problematic in the data, but the main problem is that the two sea surface temperature data sets show the largest trend differences right in the regions where they matter most for the atmosphere, namely, in the regions that form tropical convective storms. Trend differences there are about a factor of 3 larger than for the tropical average, leading to a 50% difference in the models’ atmospheric temperature trends for the period 1984–2008. The team concludes that unraveling the discrepancies will be crucial to understanding recent tropical climate trends. (Journal of Geophysical Research: Atmospheres, doi:10.1002/2014JD022365, 2014)
—Eric Betz, Freelance Writer
Citation: Betz, E. (2015), Conquering uncertainties in tropical climate forecasts, Eos, 96, doi:10.1o29/2015EO028263. Published on 17 April 2015.