Reconstruction of cropland cover changes in the Shandong Province over the past 300 years

Reconstruction of historical land cover changes has progressed significantly over the past 20 years with the establishment of regional and global land cover data sets1,2,3,4,5. For Chinese historical land cover reconstruction, scholars often use the part of China that is included in the Global Land Use Database (termed the SAGE data set), which was established by the Center for Sustainability and the Global Environment at the University of Wisconsin–Madison2, the Historical Database of the Global Environment (termed HYDE data sets), which was produced by the Netherlands Environmental Assessment Agency6, and the PJ datasets, which were established by the Planck Institute of Meteorology in Germany4. However, the accuracy of these global datasets when applied at regional scales is uncertain. Zhang et al.7 compared the reconstruction of Chinese cropland areas in traditional agricultural areas during the middle of the 17th century with the portion of China that is included in the aforementioned global data sets. The cultivation ratios in the global HYDE and PJ data sets appear to be lower than those in the Chinese regional data sets, which are based on historical cropland records. Comparing regional reconstructions of cropland cover in northeastern China over the past 300 years (termed CNEC data sets) and historical data from the Scandinavian Peninsula with the two global land cover data sets, the following conclusions can be drawn: the trends of the data sets and the phase characteristics exhibit differences, and there are clear deviations in the spatial distribution of cropland in the global land cover data8. Because of inaccuracies and regional deviations in the global data sets, the results, when applied to Chinese regional climate modeling, often deviate from reality9,10. Reconstruction of Chinese regional land cover to advance global change research is necessary, and will fulfill any impending needs in research regarding the impact of land use/cover changes and climate modeling. The accuracy, resolution and accessibility of historical land cover data sets are key issues. At present, a reconstruction of cropland cover in traditional Chinese agricultural areas, with uniform spatial and temporal resolutions, is necessary for national-scale research. These data are sourced from historical documents and are validated using data-calibration methods. The reconstruction results are also used to produce gridded data sets10,11,12. At the same time, many regional research studies in China show that long-term regional land cover changes are nonlinear and highly dynamic, and these changes are also the result of intensive interactions with the local social history11,13,14,15,16,17,18,19,20,21,22,23. The popular global-scale reconstruction method is not always suitable for reconstruction of Chinese regional land cover24. In China, there are still gaps in these regional-scale reconstructions. Therefore, it is important to strengthen the exploration of regional-scale reconstruction methods, establish a series of methods that are suitable for historical Chinese regional cropland cover reconstruction, and obtain regional cropland data sets. Reconstruction studies of China’s historical cropland cover have unique characteristics and other differences that distinguish them from Western countries. Reconstruction in China uses historical documents as the primary proxy data; these documents are tax unit records, not actual survey data, and contain more nonstandard descriptions than uniform statistical data. We collected data from 244 county gazetteers in Shandong that were compiled during the Qing Dynasty and the Republic of China period as well as statistical data since 1916 that was gathered from the central government and modern remote-sensing land cover data. The historical document data were calibrated via a correlation analysis of data from different sources using trend substitution and spatial interpolation. We examined five time periods—the late 17th century, the 18th–19th centuries, the beginning of the 20th century, the 1980 s, and the beginning of 21st century—to reconstruct the cultivated land area at the county level in Shandong. Next, a spatial distribution map of the reclamation rate was produced. A modern-day map of the suitable cropland area was then compared with historical data to estimate the saturation level of land reclamation in different regions and during different periods. The contributions of this study are the following: (i) it explores applicable reconstruction methods to study the regional historical cropland cover change in China and to establish regional data sets; (ii) it provides data sets of historical cropland cover changes in Shandong that are based on the county administrative unit; and (iii) the findings are valuable for local land use planning and land policy formulation.