摘要
黄淮海农区是中国重要的粮食生产基地,研究该地区不同等级生产力耕地的空间分布,对提高该地区生产力有重要的意义。该文在提取研究区冬小麦种植空间分布的基础上,对10a时间序列冬小麦MODIS-NDVI进行特征参数提取,并将冬小麦主要生长季多年NDVI特征值均值和年际变异系数,作为多年平均产量水平和稳产水平的指示指标,进行黄淮海农区冬小麦种植区耕地生产力评价,得到黄淮海农区冬小麦生产力空间分布图。结果显示:1)多时相MODIS-NDVI数据可以用于研究区冬小麦种植空间分布提取。经县级尺度验证,有较高的提取精度;2)县级尺度的冬小麦单产水平与其辖区内冬小麦生长关键期多个NDVI特征值有显著的相关关系,可以用来评价冬小麦生产力水平;3)研究区冬小麦种植区耕地以中低生产力水平为主,高生产力水平的耕地只占不到20%。高生产力的麦田大多分布在水热条件较好的黄淮平原亚区,中等生产力麦田大多分布在燕山太行山山麓平原亚区和鲁西黄灌区,而低生产力的麦田多分布在冀、鲁、豫低洼平原亚区。低生产力麦田分布集中连片的区域多为春旱易发、土壤粘淤或低洼积盐的地区。呈现出整体气候条件主导,局部土壤条件影响的高中低生产力空间分布特征。研究结果可以为黄淮海农区的耕地质量管理和中低产田改良提供依据。
Productivity is one basic property of farm land and the spatial pattern can be used as the baseline information in making and implementing appropriate agricultural policies. As an important winter wheat production area of China, food production in the Huang-Huai-Hai region has been receiving considerable attention for a long time. The Normalized Difference Vegetation Index （NDVI） derived from remote sensing techniques is a widely utilized vegetation index in assessing agricultural land distribution, productivity and crop growth conditions. The NDVI time series can be disaggregated into a set of quantitative metrics reflecting crop distribution and growth phenology. However, few studies take both mean crop yield and its inter-annual variability into account in evaluating the productivity of wheat land. In the present paper, we proposed a method to evaluate wheat productivity by using annual NDVI indices derived from the time series oflO years MODIS NDVI data over the Huang-Huai-Hai region. The calculated productivity can be used to monitor farmland quality. The proposed method composed of four steps： Firstly, we analyzed the reconstructed NDVI time series to generate a set of phenology indices; Secondly we extracted the distribution of winter wheat in the study area using the decision tree classifier; Thirdly we ranked each growing season＇s mean NDVI（NDVI,nean） of winter wheat cultivated area from reviving to maturity stage into five levels and the Coefficient of Variation（C.V） of ten years NDVImean into four levels; Lastly, we evaluated the winter wheat land productivity by applying a criterion established by both mean rank and C.V level. The extracted wheat planted areas were consistent with the wheat sown area obtained from the statistic database at the county and provincial scale. The statistically significant correlation between the NDVI,,ean and yields as calculated from the statistic database at the county levels suggested that the mean rank for ten years＇ NDVImean could indicate the yield level, whereas the C.V of the rank could measure the variation of yield within ten years. The relationship between the mean rank and the C.V for each 20 km^20 km grid showed that in most parts of the wheat land, the high ranked mean yield area had low temporal variations and the low rank＇s variations were high. It means the high variation of yield is related to a low productivity as well as a low yield level. Consequently, combining mean yield level and its inter-annual variability can generate more objective information for cropland productivity. The result showed that the percentage of high, middle and low level wheat land was 18.35%, 40.04% and 41.61% respectively, which suggested that there is still great potential to improve the winter wheat productivity in the study area. The spatial productivity pattem was driven by variations in hydrothermal condition, soil properties and water resources. The high productivity wheat lands were mainly located in the Huang-Huai Plain where hydrothermal condition is suitable for winter wheat. The low productivity lands were in Jin-Lu-Yu low Plain where drought is more likely to happen in spring, which is the key stage for winter wheat growth. Areas where wheat is low productive caused by infertile soil and water shortage, especially in Saline-alkali land, should be investigated in the future studies to help stabilize and increase grain yield in Huang-Huai-Hai region.
出处
《农业工程学报》
EI
CAS
CSCD
北大核心
2014年第2期153-161,F0003,共10页
Transactions of the Chinese Society of Agricultural Engineering
基金
中央高校基本科研业务费专项资金(2013QJ059)资助
“十二五”国家科技支撑计划项目(2012BAD05B02)
关键词
遥感
模型
农作物
黄淮海农区
冬小麦生产力
分级NDVI
产量变异
remote sensing
models
crops
Huang-Huai-Hai Region
winter wheat productivity
graded NDVI
ield variance
