Subsurface Signal of Deforestation from a Climate Model Experiment: Implications for Borehole Temperature Reconstructions of the Common Era

Abstract:

A longstanding problem effecting paleoclimate reconstructions from borehole temperature data is related to the subsurface noise from land use changes, in particular deforestation. Here a series of model experiments were conducted to explore the effect of deforestation on subsurface temperatures. The experiments were conducted using the land surface component of the University of Victoria Earth System Climate Model (UVic ESCM) forced offline with CRU corrected NCEP reanalysis data. The model was spun-up for 500 years by repeating the 1900-1931 monthly climatology. Simulated deforestation was imposed globally at 100, 99, 90, 70, and 50% of the grid cell area. Deforestation was held for 10 years and thereafter ecosystems were allowed to recover. Baseline simulations were also conducted where no deforestation occurs. By examining the difference between the deforestation and baseline simulations the effect of deforestation on various model metrics is isolated. It is found that the change in subsurface temperature depends on the competing surface energy balance effects of an increase in surface albedo and decrease in surface roughness length induced from deforestation. In once heavily forested regions the net result is an increase in the surface and subsurface temperatures. The magnitude of the thermal noise due to deforestation can reach 1 K at 23 m depth in the decades following deforestation. The results of these experiments illustrate the need to take into account deforestation when conducting borehole derived climate reconstructions.