基于统计年鉴和MODIS的中国区域土地利用/覆盖变化特征研究

土地利用/覆盖变化(Land Use/Cover Change,LUCC)是影响区域气候变化不可忽略的因素,然而目前还没有十分可靠的土地利用/覆盖数据,因此其对气候变化影响的研究存在很大不确定性。基于此,利用中国统计年鉴(简称年鉴)和中分辨率成像光谱仪(MODerate-resolution Imaging Spectroradiometer,MODIS)遥感观测资料,以行政区为研究单元对3种典型土地利用/覆盖类型(森林、城市和农田)的变化进行了比较分析。结果表明:(1)2004~2011年,年鉴和MODIS森林覆盖率在中国各省(市、区)的分布和变化整体一致性较好,年鉴和MODIS数据都显示全国大部分省(市、区)份的森林覆盖率均有不同程度的增加,而仅MODIS数据中的北京、天津、吉林、黑龙江、上海、江苏森林覆盖率降低。MODIS数据能准确反映森林总体覆盖情况,由于类别精度不够,对单种森林类型及变化的描述存在较大误差。(2)年鉴表明中国东部城市建成区覆盖率及其增长远高于西部地区,黄淮海地区、东南部沿海地区城镇用地呈现加速扩张趋势,符合实际情况。但MODIS数据没有表征出中国区域近10年来的快速城市化进程,可能原因是城市面积较小,且地面地物分布复杂,受空间分辨率的限制,在离散的不集中的区域MODIS数据的土地覆盖类型分类精度较低,导致监测不到新城市的扩张。(3)MODIS农田面积和年鉴农作物播种面积、有效灌溉面积覆盖率的数值和空间分布均有较好的一致性,均是黄淮海地区覆盖率最大。但在2001~2011年,3种数据变化量的差异较大,尤其东部省(市、区)份,MODIS农田面积和年鉴有效灌溉面积增加,而年鉴农作物播种面积减少。(4)2004~2011年中国土地利用类型变化中年鉴城市建成区的变化速率最大(6.25%),其次为森林;耕地由于总量大,变化部分所占的比例较小,因而MOIDS农田及年鉴农作物播种面积和有效灌溉面积的变化速率均较小。

Characteristics of Land Use/Cover Change over China Based on China Statistical Yearbook and MODIS Data

Land Use/Cover Change (LUCC) is an influencing factor of regional climate change that cannot be ignored. However, there still exists no highly reliable Land Use/Cover data so far, and thus great uncertainties remain in the study of the effects of LUCC on climate change. For this reason, two datasets from China Statistical Yearbook and observed MODIS (MODerate-resolution Imaging Spectroradiometer) respectively are used to compare and analyze changes in three typical land use/cover types (forest, urban, cropland) over administrative regions. The results show that:(1) Spatial distribution and changes in the forest cover in provinces (municipalities, autonomous regions) of China overall agree well during 2004-2011 between China Statistical Yearbook and MODIS data. Both China Statistical Yearbook and MODIS data show that forest cover increases to different degrees in most provinces (municipalities, autonomous regions) of China but decreases in Beijing, Tianjin, Jilin, Heilongjiang, Shanghai, and Jiangsu based on MODIS data. MODIS data can accurately reflect the total forest cover, but there exist large errors in single forest types and their changes due to low classification accuracy. (2) The area of urban built districts from China Statistical Yearbook and its increment are much higher in eastern China than in the western region. Huanghuaihai region and southeastern coastal region of China show an accelerated expansion trend of urban area, which is consistent with the actual situation. However, MODIS data cannot reflect the process of rapid urbanization development in China over the past decade. The possible reasons are that urban areas are usually small with complex surface. Limited by spatial resolution, the classification accuracy of MODIS land cover is low in discrete regions, leading to no monitoring results of expansion of the new cities. (3) Both the values and spatial distribution are in good consistency among MODIS cropland, sown area of farm crops and irrigated area from China Statistical Yearbook, and the largest coverage of these three data are all in Huanghuaihai region. However, change quantities of the three data show great difference during 2001-2011. In particular, MODIS cropland and irrigated areas from China Statistical Yearbook increase in provinces (municipalities, municipalities) in eastern China, while sown area of farm crops from China Statistical Yearbook decrease. (4) The change rate of urban built districts in China Statistical Yearbook is the largest (6.25%) among all types of land use changes over China during 2004-2011, followed by forest. Because the quantity of arable land is large and the changes account for a relatively small proportion of the arable land, the change rate of MODIS cropland, sown area of farm crops and irrigated area are all small.