Exploring Elevation-Dependent Climate Warming since the Last Glacial Maximum over the Mountains of East Africa Using Semi-Idealized Regional Climate Model Experiments

Recent studies have used paleoclimate proxies to reconstruct temperatures over the mountainous regions of East Africa extending back more than 20,000 years and at sites with elevations as high as about 3 km above sea level. When reconstructed temperatures at these sites are compared between the recent preindustrial period (PI) (about 250-2,000 years ago) and the last glacial maximum (LGM; about 20,000-23,000 years ago), warming (PI minus LGM) is found to depend strongly on elevation. Near surface conditions at sea level are inferred to have warmed by about 2° C, whereas high-elevation sites are inferred to have warmed by more than 5°C, suggesting a reduction in the surface temperatures lapse rate. A reduction in tropical free-tropospheric lapse rates is expected with climate warming due to the role of the moist convection in setting tropical lapse rates. However, the magnitude of this reconstructed lapse rate change is in excess of the free-tropospheric changes predicted by global climate models (GCMs), which has been hypothesized to indicate a bias in model representation of the tropical troposphere’s response to climate forcing.

Here we explore the possibility that local processes over complex terrain might explain the apparent discrepancy between paleoclimate reconstructions and GCM simulations. We conduct high-resolution (4-km grid spacing) convection permitting regional climate model simulations using the WRF model over the mountains of East Africa. A semi-idealized approach is used to concentrate on the regional response to a prescribed large-scale climate forcing. First, simulations of the modern climate are conducted, forced by reanalysis boundary conditions. Then, simulations broadly representative of the LGM climate are conducted by perturbing the reanalysis boundary conditions with the average PI minus LGM change predicted by an ensemble of GCMs and with LGM radiative forcing. This “pseudo global warming” approach focuses on the mesoscale response to large-scale thermodynamic and radiative forcing. Using results from these experiments, we characterize the extent to which changes in surface lapse rates are indicative of changes in free-tropospheric lapse rates. The potential contributions to lapse rate changes of local processes such as albedo feedbacks, vegetation changes, soil moisture feedbacks, and orographically influenced cloud feedbacks are considered.