A multi-proxy reconstruction of anthropogenic land use in southwest Asia at 6 kya: Combining archaeological, ethnographic and environmental datasets

Land use and land cover (LULC) changes have important biophysical and biogeochemical effects on climate via a variety of mechanisms. Several climate modelling studies have demonstrated the impact of LULC scenarios on past climate reconstructions. Testing the impact of anthropogenic land use on mid-Holocene climate thus requires reconstructions of land use that accurately reflect this time frame. To address these concerns, the PAGES LandCover6k working group aims to create data-driven gridded global reconstructions of land use and land cover to provide the climate modelling community with inputs for sensitivity testing of the impact of LULC changes on global climate. As one of the earliest global centres of domestication, agricultural production, and population nucleation, Southwest Asia represents one of the areas of the world expected to display the greatest land use impact and human-induced land cover change at 6 kya, and is therefore critical for the mid-Holocene time frame. Here, we reconstruct land use for Southwest Asia for the 6 kya time frame at a regional scale. We draw on environmental data to reconstruct the range of possible land uses within each particular environment and on archaeological and historical data to reconstruct actualized land use. We then compare this reconstruction to common global LULC models, including the most recent HYDE and KK10 iterations. The reconstruction presented here differs from these previous reconstructions in its methodological approach, spatial extent and resolution. It also differs from both models in population density distribution and land use allocation. While the output of our reconstruction is generally more similar to HYDE 3.2 than KK10, particularly in terms of reconstructed pastoral land use, we model greater agricultural land use than HYDE across the entire region, and less land use overall compared with KK10. The paper provides a method for systematically incorporating archaeological data into models of past land use and demonstrates the value of such an approach for enhancing empirical validity.