论文部分内容阅读
【目的】以Landsat-8卫星数据为对照,探索国产GF-1卫星数据对林区积雪特征的估测能力,实现基于国产卫星对东北林区雪水水文过程的监测。【方法】以大兴安岭北部林区为研究区,结合同期的地面雪水当量野外观测数据,利用偏最小二乘回归与BP神经网络两种方法建立线性与非线性雪水当量反演模型。通过平均均方根误差(E_(RMSE))、平均相对均方根误差(r_(RMSE))和平均估测精度这3个评价指标对所建模型进行对比评价。同时,利用两种遥感数据建立的最优模型对研究区内雪水当量分布特征进行反演,并对反演结果进行对比分析。【结果】基于GF-1数据所建立的线性与非线性模型性能均略低于以Landsat-8数据构建的模型,其中GF-1数据最优反演模型精度为80.3%,较Landsat-8反演模型低1.6%;基于GF-1数据反演的雪水当量值与Landsat-8的基本相同;两类遥感数据反演得到的雪水当量在空间分布特征上基本一致,均反映了雪水当量与地形、植被及土地利用类型的高度相关性;由于山地林区植被和地形复杂,并且春季升温过程中地面积雪日消融速率快,GF-1数据以其高空间与时间分辨率上的优势能够更好地对研究区雪水水文过程进行监测。【结论】国产GF-1卫星能够替代Landsat-8卫星成为对大兴安岭北部林区雪水水文过程监测的遥感数据源。
【Objective】 Landsat-8 satellite data was used as a control to explore the capability of domestic GF-1 satellite data to estimate snow cover in forest area, and to monitor snow hydrological process in Northeast China based on domestic satellite. 【Method】 Based on field data of snow equivalent field in the same period, two models of linear and nonlinear snow water equivalent were established by using partial least squares regression and BP neural network. The model was evaluated by three evaluation indexes: average root mean square error (RMSE), average relative root mean square error (RMSE) and average estimation accuracy. At the same time, the optimal model established by two remote sensing data was used to invert the distribution of snow water equivalent in the study area, and the contrastive analysis was carried out. 【Result】 The performance of linear and nonlinear models based on GF-1 data was slightly lower than that of Landsat-8 data. The accuracy of GF-1 data inversion model was 80.3% The snow water equivalent value based on GF-1 data is basically the same as that of Landsat-8. The snow water equivalent retrieved from the two types of remote sensing data is basically consistent in the spatial distribution characteristics, both of which reflect the snow Because of the complexity of the vegetation and topography in the mountainous area and the rapid snow surface diurnal ablation rate during the spring warming season, the GF-1 data are highly correlated with the spatial resolution and temporal resolution Can better monitor the snow hydrological process in the study area. 【Conclusion】 Domestic GF-1 satellite can replace Landsat-8 satellite as a remote sensing data source for monitoring snow hydrological process in the northern part of Daxingan Mountains.