A spatially explicit reconstruction of cropland cover in China around 1850 C.E. employing new land suitability based gridded allocation algorithm.

A credible spatially explicit historical cropland dataset is the basis for accurately evaluating the impacts of anthropogenic land cover change on global changes. The credibility of the historical cropland datasets relies on both the reconstructed regional cropland area and the allocation algorithm. In the case of adopting the more reliable reconstructed historical cropland area in a given region, the credibility of the gridded cropland cover data mainly depends on the allocation algorithm, which makes a major contribution to the uncertainties of the spatial distribution of the cropland cover. In this study, a new land suitability-based gridded allocation algorithm for historical cropland cover was used to allocate the provincial cropland area into 5' × 5′ grids in China around 1850 C.E. The algorithm was constructed using the dominant physiogeographic factors that vary from region to region, and these factors were assessed by the modern cropland cover. The main conclusions are described as follows: (1) the cultivation intensity under a 5' × 5′ cell can be effectively indicated by the physiogeographic factors in a 0.5° × 0.5° unit, although the degrees of correlation showed a diverse variety with positive, negative correlation or irrelevance relationships in different regions. (2) The land suitability calculated from dominant regional physiogeographic factors could appropriately represent the cultivation intensity. The proportion of positive correlations above 0.7 between the modern cropland fraction and land suitability in China is more than 93%. (3) Employing the newly developed allocation algorithm, the 5' × 5′ cropland fraction data in China around 1850 C.E. was generated from the reconstructed provincial cropland area. The traditional farming regions were maturely cultivated in 1850 C.E., while the distribution of cropland was relatively small and had a lower cultivation intensity in the frontier region (Xinjiang Province and Northeast China, etc.). (4) The new allocation algorithm is expected to be applicable in regions with only very coarse spatial resolution cropland reconstruction results, and the reasonability of gridded allocation results will be improved, especially in regions at the early cultivation stages with very low intensity in China. [ABSTRACT FROM AUTHOR]