Estimating long-term changes in China's village landscapes

Abstract

Over the past 50 years, China's ancient agricultural village landscapes have been transformed by unprecedented social, technological, and ecological changes. Although these dense anthropogenic mosaics of croplands, settlements, and other used lands cover more than 2 million square kilometers across China, the nature of these changes and their environmental impacts remain poorly understood because their spatial scale is generally too small to measure accurately using conventional land-change methods. Here, we investigate the regional consequences of fine-scale landscape changes across China's village regions from 1945 to 2002 using high-resolution, field-validated ecological mapping of a regionally stratified sample of village landscapes at five sites across China, with uncertainties estimated using model-based resampling and Monte Carlo methods. From 1945 to 2002, built surface areas increased by about 7% (90% credible interval = 2-17%) across China's village regions, an increase equivalent to about three times the total urban area of China in 2000. Although this striking result is explained by a near doubling of already large village populations and by lower housing density per capita in rural areas, two unexpected changes were also observed: a 9% net increase (-4% to +21%) in regional cover by closed canopy trees and an 11% net decline (-30% to +3%) in annual crops. These major regional changes were driven primarily by intensive fine-scale land-transformation processes including tree planting and regrowth around new buildings, cropland abandonment, and by the adoption of perennial crops and improved forestry practices. Moreover, the fragmentation, heterogeneity, and complexity of village landscapes increased over time. By coupling regional sampling and upscaling with observations in the field, this study revealed that fine-scale land-change processes in anthropogenic landscapes have the potential for globally significant environmental consequences that are not anticipated, measured, or explained by conventional coarser resolution approaches to global and regional change measurement or modeling. © 2008 Springer Science+Business Media, LLC.