Paper Reviewed

Ji, M., Huang, J., Xie, Y. and Liu, J. 2015. Comparison of dryland climate change in observations and CMIP5 simulations. Advances in Atmospheric Sciences 32: 1565-1574.

In an enlightening study published in Advances in Atmospheric Sciences, Ji et al. (2015) report how they investigated both historical area changes and climate changes of drylands and subtype regions using both observational datasets and CMIP5 simulations for the period 1948-2005. And what they found by so doing was most interesting.

First of all, the four researchers found that the global drylands in the observations they analyzed expanded remarkably during the 58-year study period, but that (1) "this expansion was severely underestimated in the CMIP5 multi-model ensemble (MME)." They also noted that the dryland area expansion was accompanied by a decline or drying trend in aridity index (AI), as a result of decreased precipitation and increased evapotranspiration across all dryland subtype areas, especially in the semi-arid and dry sub-humid regions," but that (2,3) "the CMIP5 MME average performed poorly with regard to the decreasing trends of AI and precipitation." And in this regard they say that (4) "the overall bias of AI in the simulations, compared with observations, was largely due to limitations in the simulation of precipitation."

Other interesting findings revealed that (5) "on the global scale, the simulated precipitation was overestimated compared with observations across all subtype areas, especially in the semi-arid and dry sub-humid regions," that (6) "in specific dryland regions, AI was over-estimated or under-estimated consistently, with over- or under-estimation of precipitation, regardless of whether PET [potential evapotranspiration] was over- or under-estimated." And they further remark that (7) "recent studies also demonstrate that the large uncertainties in precipitation are mainly caused by natural variations in sea surface temperatures, which are often not captured by climate models," additionally citing Dai (2013) in this regard.

Consequently, there was little more that Ji et al. could do in concluding their discussion of their findings than admit that (8) "as the CMIP5 MME cannot capture the observed variability of dryland precipitation, further investigations are needed to validate the regional precipitation simulation capability of specific models."

Reference
Dai, A.G. 2013. Increasing drought under global warming in observations and models. Nature Climate Change 3: 52-58.