An approach to reconstructing spatial distribution of historical cropland with grid-boxes by utilizing MODIS land cover dataset:A case study of Yunnan Province in the Qing Dynasty

Highly precise Land Use and Cover Change(LUCC) dataset plays a key role in improving simulations of effects of LUCC on climate and ecosystem.Historical LUCC usually has no precise spatial location information.This shortage limited usage of historical LUCC dataset in the global environmental changes simulations.So,it is needed to develop an effective way to reconstruct historical cropland spatial distribution with grid-boxes.In this study,we develop a new way in which the spatial distribution of historical cropland was reconstructed effectively.This approach was built on a reasonable hypothesis that historical cropland was located in the domain of present cropland area.This hypothesis was derived from a feature that cropland area increased all the time generally in the past 300 years.This approach includes two steps:(1) estimating the easiness for reclamation one pixel by one pixel within the cropland domain determined by MODIS land cover product;(2) by descending order of easiness for reclamation,filling in the pixels with cropland from historical inventories;it would not stop until the total area of cropland pixels was equal to inventory cropland area.As a case study,we reconstructed the spatial distribution with a 90-m resolution of cropland in the Yunnan Province in 1671 and 1827 by using this approach.The results show this approach could reconstruct the historical cropland spatial distribution with high resolution.