王蕊, 李红军, 雷玉平. 基于多年MODIS NDVI 分级的河北平原农田生产力评价[J]. 中国生态农业学报(中英文), 2011, 19(5): 1175-1181. DOI: 10.3724/SP.J.1011.2011.01175
引用本文: 王蕊, 李红军, 雷玉平. 基于多年MODIS NDVI 分级的河北平原农田生产力评价[J]. 中国生态农业学报(中英文), 2011, 19(5): 1175-1181. DOI: 10.3724/SP.J.1011.2011.01175
WANG Rui, LI Hong-Jun, LEI Yu-Ping?. Evaluation of cropland productivity in the Hebei Plain via graded multi-year MODIS-NDVI data[J]. Chinese Journal of Eco-Agriculture, 2011, 19(5): 1175-1181. DOI: 10.3724/SP.J.1011.2011.01175
Citation: WANG Rui, LI Hong-Jun, LEI Yu-Ping?. Evaluation of cropland productivity in the Hebei Plain via graded multi-year MODIS-NDVI data[J]. Chinese Journal of Eco-Agriculture, 2011, 19(5): 1175-1181. DOI: 10.3724/SP.J.1011.2011.01175

基于多年MODIS NDVI 分级的河北平原农田生产力评价

Evaluation of cropland productivity in the Hebei Plain via graded multi-year MODIS-NDVI data

  • 摘要: 根据中等分辨率航天成像光谱仪(MODIS)遥感数据计算的归一化植被指数(NDVI)被广泛用于作物长势监测和产量预报, 但由于NDVI 数值在不同年份的同一时期变化较大, 直接用于评价农田生产力会有较大误差。本文以河北平原所在的北纬37°~39°之间连续种植的冬小麦农田为研究区域, 通过对多年冬小麦MODISNDVI 数据进行比较和分级, 尝试用每季NDVI 在区域内的高低级别评估区域农田生产力。Landsat 卫星数据用于对不同时相MODIS 图像进行精确配准, 从而实现像素尺度上长时间序列数据的统计分析。首先, 对区域内2000~2008 年间每年作物返青期到成熟期的NDVI 平均值及各生育阶段NDVI 平均值分别进行高低分级, 以了解河北平原农田生产力的空间变异, 结果显示其中高水平农田分布在太行山山前平原, 指数等级水平并没有完全按南北走向趋势分布, 表明该研究方法受纬度差异的影响较小。不同年份分析结果显示, 2008 年东部地区也出现了较高等级的田块。其次, 利用NDVI 分级结果计算出9 年间NDVI 等级的变异系数, 对采用不同生育期NDVI 可能带来的误差进行了分析, 结果显示不同小麦生育期NDVI 等级的变异系数不同, 返青期和成熟期变异系数较大, 且具有一定的地理差异。最后, 利用GIS 空间分析方法以9 年NDVI 分级结果为基础制作了以县为单元的麦田生产力等级图, 结果显示河北平原农田生产力高低分区, 同时也表明中低水平区块有较大提升空间, 为河北县级土地管理和耕地质量管理提供理论依据。

     

    Abstract: Moderate resolution imaging spectroradiometer sensor (MODIS) provides high-temporal, medium-spatial resolution science-quality global land cover data. Calculated normalized differential vegetation index (NDVI) from MODIS data is widely used for crop growth monitoring and yield forecasting. However, variations of NDVI values in the same season of each year lead to uncertainties and biases in forecasting cropland productivity. This study presented a method for evaluating cropland productivity and indirect monitoring of soil quality, using annual NDVI calculated from multi-year MODIS data in the Hebei Plain. Calibrated Landsat data with high spatial resolution were used to process geometric registration of the MODIS data. This was used to determine the exact region with continuous winter wheat cultivation for many years in the plain. Mean NDVI of winter wheat cultivated area from reviving to maturity stage was calculated for 2000~2008 and ranked into different levels in order to analyze for spatial variations in farmland productivity. The study showed that while high productivity croplands were mainly located in the piedmont of Mount Taihang, medium or low productivity croplands were in the central and eastern regions of the plain. Longitudinal change in NDVI was not regular, indicating that the method used was not affected by latitude difference. The distribution pattern was driven by variations in local land use types and cropland quality management. The study also analyzed the causes of biases in the coefficient of variation for the 9-year NDVI data. The analysis showed that the coefficient of variation was relatively high during the reviving and maturity stages of winter wheat, with also spatial differences. Combined geo-spatial and NDVI analysis was used to map cropland productivity at county level. The map showed spatial differences in productivity with huge potential croplands where low/medium productivity was located at. The study therefore laid the theory basis for land control and cropland quality management in Hebei Province, which could be vital for improving cropland quality at county level.

     

/

返回文章
返回