摘要:草原植被长势遥感监测具有时效性高、覆盖范围广的特点, 及时高效的草原长势遥感监测信息对于草地资源的保护与合理利用具有重要意义。为明确不同作物长势遥感监测方法在草原长势信息监测中的差异及其应用的局限性, 本研究利用MODIS数据的NDVI产品, 将直接监测法、植被生长过程曲线法、同期对比法、基于NDVI百分位数法分别在内蒙古自治区西乌珠穆沁旗草原进行了应用和适宜性研究, 并对其长势遥感监测结果进行了验证。结果表明, 直接监测法利用草原NDVI数值直观地反映了草原植被长势的空间异质性, NDVI数值与草原单位面积产草量干重显著相关(R2=0.5502), 但对不同类型草原内部的长势差异信息反映不够清晰; 植被生长过程曲线法集总地反映了监测区域草原整体长势随时间的变化, 以及相对于参考年份的差异, 但不同草原类型需要单独监测才能反映其各自的生长过程, 本研究中草甸草原和典型草原NDVI过程曲线的峰值分别为0.73和0.55, 与全区域集总式监测结果(峰值为0.60)差异较大; 同期对比法因参考年份的不同产生不同的评价结果; 基于NDVI的百分位法能够定量地评价草原的长势, 长势评分与草原单位面积产草量干重相关性的决定系数为0.5047。实际监测中应依据草原监测目标选择适宜的方法或组合。随着天-空-地一体草情长势监测平台的建设和发展, 将能提供实时的地面监测辅助信息, 有效提高草原植被长势遥感监测的效率与精度。
关键词:草原植被/
长势/
遥感/
NDVI/
MODIS数据
Abstract:The research and application of grassland vegetation growth monitoring methods have important scientific significance and application value for the sustainable use of grassland resources and the improvement of the ecological environment. Remote sensing monitoring has characteristics such as high timeliness and wide coverage, and several remote sensing monitoring methods have been increasingly used in crop growth monitoring. As the monitoring objects are all plants, these monitoring methods were tried to introduce to monitor grassland vegetation growth in this study. We applied four remote sensing monitoring methods for the crop growth: direct monitoring, vegetation growth process curve, same period comparison, and NDVI percentiles to the grassland vegetation growth monitoring of West Ujimqin County in Inner Mongolia; to clarify the suitability and limitation of these remote sensing monitoring methods for crop growth when monitoring grassland vegetation growth using MODIS Vegetation Index Products (NDVI). The monitoring results provided by the direct monitoring method and the NDVI percentile method were compared with the ground sampling data. The direct monitoring method could intuitively reflect the spatial heterogeneity of grassland vegetation growth by the grassland NDVI, and the NDVI value was significantly correlated with the dry weight of grassland yield per unit area (R2=0.5502). However, this method could not provide details on the growth of different types of grasslands owing to the limitation of the NDVI grade. The vegetation growth process curve method could only collectively reflect the changes in the overall growth of the grassland in the monitored area over time, showing that the growth was better or worse than that of the reference year. In this study, the NDVI peak values of the vegetation growth process curves for the meadow grassland and typical grassland were 0.73 and 0.55, respectively, significantly different from the whole regional lumped monitoring results (NDVI peak value was 0.60). This means that different grassland types should be monitored separately to reflect their respective growth processes. For the same period comparison method, if the selected reference year was different, the method would provide different monitoring results for grassland growth; the results from grassland growth monitoring were semi-quantitative comparative. Using the statistical analysis of NDVI data of different grassland types for 5 years, the NDVI percentile method could quantitatively evaluate the growth of corresponding grassland types, as shown in this study. The determination coefficient of the correlation between the growth score provided by the NDVI percentile method and the dry weight of grassland yield per unit area was 0.5047. To achieve a reasonable classification of these semi-quantitative monitoring results of grassland growth, the assistance of ground grassland monitoring information is required. There is an urgent need to establish a sky-air-ground integrated grassland growth monitoring platform to improve the efficiency and accuracy of grassland vegetation growth monitoring.
Key words:Grassland vegetation/
Growth/
Remote sensing/
NDVI/
MODIS data
PDF全文下载地址:
http://www.ecoagri.ac.cn/article/exportPdf?id=b00ce998-8b93-41ed-b5b3-1f0e98b3f508
