Climate sensitivity to tropical land surface changes with coupled versus prescribed SSTs

Tropical land cover change experiments with fixed sea-surface temperatures (SSTs) and with an interactive ocean are compared to assess the relevance of including the ocean system in sensitivity studies to land surface conditions. The results show that the local response to deforestation is similar with fixed and simulated SSTs. Over Amazonia, all experiments simulate a comparable decrease in precipitation and no change in moisture convergence, implying that there is only a change in local water recycling. Over Africa, the impact on precipitation is not identical for all experiments; however, the signal is smaller than over Amazonia and simulations of more than 50 years would be necessary to statistically discriminate the precipitation change. We observe small but significant changes in SSTs in the coupled simulation in the tropical oceans surrounding the deforested regions. Impacts on mid and high latitudes SSTs are also possible. As remote impacts to deforestation are weak, it has not been possible to establish possible oceanic feedbacks to the atmosphere. Overall, this study indicates that the oceanic feedback to land surface sensitivity studies is of second importance, and that the inclusion of the oceanic system will require ensembles of long climate simulations to properly take into account the low frequency variability of the ocean.