Intercomparison of methods for disentangling local and nonlocal biophysical impacts of land cover changes

The climatic impacts of land cover changes (LCCs) due to altered biophysical properties include local effects in LCC areas and nonlocal effects in both LCC and non-LCC areas resulting from atmospheric feedback. The biophysical impacts of LCC simulated by climate models typically represent only the total effects, that is, the sum of local and nonlocal effects. The respective local and nonlocal effects have been disentangled from climate model simulations using several methods. However, a systematic intercomparison of these methods under a comparable modeling framework is still lacking, hindering the complete understanding of LCC's biophysical effect on the methodological nature. This study employs a series of unified global deforestation experiments to assess the performance of existing methods in quantifying the local and nonlocal effects of LCC on land surface temperature. Results show that local effects derived by different methods exhibit overall consistent latitudinal and seasonal patterns compared to those in observational data sets, including the transition zone between warming and cooling across the Northern Hemisphere. The separated nonlocal effects, which dominate the total effects, can be reproduced with broad similarity across methods. Nevertheless, regional discrepancies exist, reflecting the different assumptions of each method. We further discuss and summarize the strengths and limitations of each approach and offer practical suggestions on their usage. This study provides methodological guidance for the quantitative assessment of biophysical climate effects of LCC across multi-scales for climate change research.