基于遥感方法的黑河中游荒漠绿洲区生长季蒸散量估算

发布时间：2018-02-28 21:07

本文关键词： 生长季蒸散遥感模型对比模型验证　出处：《西北农林科技大学》2016年博士论文　论文类型：学位论文

【摘要】：黑河中游流域是典型的干旱荒漠绿洲区,对流域水资源有显著影响的土地覆盖类型有农田、防护林、绿洲-荒漠过渡带人工草地和荒漠植被等,了解不同土地覆盖类型蒸散量的时空变化特征对于流域水资源的合理配置与利用具有十分重要的意义。本论文利用Landsat系列卫星影像资料,结合地面气象观测数据、物候观测数据和土地覆盖类型,将METRIC模型应用于黑河中游,分析了样带尺度蒸散量的时空变化特征;在不同时间尺度、不同土地覆盖类型,采用涡度观测蒸散数据对三种基于干湿限空间特征法的遥感蒸散模型的估算结果进行了验证和对比;利用样带内63个监测点的土壤水分数据和Hydrus-1D模拟蒸散数据,在像元尺度分析METRIC模型估算季节尺度蒸散量的准确性,并利用空间插值方法将Hydrus-1D模拟的蒸散数据扩展到样带尺度,与METRIC的估算结果进行对比。主要研究成果如下:(1)提出在干旱区应用METRIC模型估算不同土地覆盖类型的蒸散时,需要将研究区根据下垫面特征划分为子区域,分别选取干点运行模型。绿洲区与荒漠区下垫面土壤性质、植被分布等差异显著,不同土地覆盖类型蒸散为零的像元地表温度可能存在较大差异。研究结果表明:干点温度每降低2 K,区域显热通量平均值可增加10-30 W/m2;不同干点选择影响不同土地覆盖类型蒸散量估算结果。利用涡度观测数据验证黑河中游METRIC模型估算结果,发现:在荒漠区和绿洲区分别选取干点,METRIC模型可以满足黑河中游干旱区不同时间尺度蒸散量的估算需求。(2)分析了黑河中游样带尺度不同土地覆盖类型生长季蒸散量时空变化特征。不同土地覆盖类型日蒸散量差异明显,水体(河渠、水库坑塘和沼泽地)蒸散量最大,其他土地覆盖类型蒸散量从高到低依次为:农田、滩地、林地、过渡带(灌木林地、疏林地、草地)、未利用土地(沙地、戈壁和盐碱地);各土地覆盖类型月蒸散量均在7月达到最大值。农田蒸散量主要随作物生长阶段进行变化,月际差异最大;滩地受降水影响较大,在较大降水之后,蒸散量接近潜在蒸散量。林地及过渡带月际变化较小。未利用土地蒸散受降水影响明显,7月-8月蒸散量最大。受到作物种类及生长年限、土壤性质、地下水水位等影响,相同地类蒸散量的空间变化也很显著。(3)在不同时间尺度,利用站点涡度观测数据验证了METRIC、Ts-VI、SSEB在干旱荒漠绿洲区的蒸散估算精度;并对比了三种方法估算不同土地覆盖类型的蒸散量。结果表明:METRIC模型在瞬时、日、月的不同时间尺度蒸散验证中,估算精度都较高,而SSEB模型和Ts-VI三角法估算蒸散量的精度,与像元可利用能量、最大可能蒸散量等的估算精度密切相关。在农田区,三种模型的估算结果都较为一致;在荒漠区及荒漠-绿洲过渡带子区域的低值区,SSEB模型和Ts-VI三角法表现出明显的高估趋势。在估算非卫星过境日蒸散时,不同土地覆盖条件推荐采用不同的时间尺度扩展方法,建议在绿洲区采用参考蒸散比不变法,在荒漠区采用蒸发比不变法。(4)利用样带内63个监测点的水分数据和Hydrus-1D模拟蒸散量,分析了METRIC估算不同土地覆盖类型蒸散量在季节尺度的分布规律,发现:在像元尺度上,应用METRIC和Hydrus-1D模型得到的农田蒸散估算结果在日、月、季节尺度均具有较好的一致性;但在估算荒漠蒸散时,日蒸散值存在“失真”现象。在估算月尺度或季节尺度蒸散时,这种影响将会降低,METRIC模型估算的月蒸散总量和季节蒸散总量与Hydrus-1D模拟结果仍具有较好的一致性。此外,在地表资料(气象数据、土壤水分数据、植被结构数据,灌溉数据等)充足的条件下,Hydrus-1D虽然可以提供准确的像元尺度ET的变化特征,但无法通过简单的地统计方法将像元ET扩展到样带或区域尺度。
[Abstract]:The middle reaches of Heihe basin is a typical oasis in arid desert area, has a significant impact on water resources in the basin land cover types farmland protection forest belt, artificial grassland and desert vegetation in oasis desert transition, to understand the spatial and temporal variation characteristics of evapotranspiration of different land cover types of great significance for the rational allocation and utilization of water resources in the basin has. This paper uses Landsat series of satellite image data, combined with meteorological data, phenological observation data and land cover types, the application of METRIC model in the middle reaches of Heihe, analyzes the transect scale spatial and temporal variation characteristics of the bulk vapor; scale at different time, different land cover types, the vorticity observation data of three evapotranspiration for estimating evapotranspiration model and limited space feature method based on the results of the verification and comparison; utilization of water soil within the transect of 63 monitoring points and data points Hydrus-1D simulation of evapotranspiration data in pixel scale analysis of METRIC model to estimate the accuracy of seasonal evapotranspiration, and the evapotranspiration data using spatial interpolation method Hydrus-1D simulation is extended to the transect scale, compared with the METRIC estimation results. The main research results are as follows: (1) proposed to estimate evapotranspiration of different land cover types in the application of METRIC drought model, need to study area according to the features of the underlying surface is divided into sub regions, were selected to do the operation model. The surface area and soil properties of desert oasis, vegetation distribution and other significant differences in different land cover types, evapotranspiration may be zero pixel surface temperature has a significant difference. The results show that: the temperature of each point by 2 K, the average area of sensible heat flux can be increased by 10-30 W/m2; effect of different land cover types of evapotranspiration estimation results of different choices. By using vorticity observation point Data validation in the middle reaches of Heihe METRIC model estimation results, that were selected do in desert and oasis area, METRIC model can meet the demand of different time scales to estimate evapotranspiration in arid area in the middle reaches of Heihe. (2) analysis of the middle reaches of Heihe River transect scale in different land cover types growing season evapotranspiration temporal variation characteristics of different land cover. The type of daily evapotranspiration significantly, water (river, reservoir and pond and swamp evapotranspiration) maximum evapotranspiration, other land cover types from high to low: farmland, woodland, beach, transition zone (shrub forest, woodland, grassland, unused land (sand), Gobi and salina) the land cover types; monthly evapotranspiration amount reached the maximum in July. Farmland evapotranspiration with the major crop growth stage changes, the largest monthly difference; beach affected by precipitation in large, large precipitation, evapotranspiration close to potential In the transition zone and forest evapotranspiration. Monthly variation is small. The unused land evapotranspiration was affected by precipitation and evapotranspiration. The maximum -8 months of July by crops and different growth years, soil properties, effects of groundwater level, the same kind of spatial variation of evapotranspiration is also significant. (3) in different time scales. Validation of the METRIC, using the site vorticity data of Ts-VI, SSEB in the arid desert oasis evapotranspiration estimation accuracy; and compares the estimated evapotranspiration of different land cover types of three methods. The results show that the METRIC model in the day, the month is not instantaneous, with time scale evapotranspiration verification, the estimation accuracy is high, and the SSEB model and the Ts-VI triangulation method for estimating evapotranspiration accuracy and pixel energy availability estimation accuracy, the maximum potential evapotranspiration is closely related. In farmland area, estimation results of the three models are more consistent; in the desert area and desert green The low value zone of transition belt Delta region, SSEB model and Ts-VI triangulation method showed obvious overestimation of the trend. In the estimation of non satellite transit daily, the condition of different land cover recommended by different time scale expansion method, recommended reference evapotranspiration in oasis in the desert area ratio method, the evaporation ratio method. (4) simulated evapotranspiration using transect 63 monitoring points in the water data and Hydrus-1D, analyzes the METRIC estimation of different land cover types in evapotranspiration distribution, seasonal scales found on pixel scale, evapotranspiration estimation results obtained using METRIC and Hydrus-1D model in the day, month, season scale has good consistency; but in the estimation of evapotranspiration in desert, there are "distortion" in the daily value. In the estimation of monthly or seasonal scale evapotranspiration, this effect will be reduced, the monthly evapotranspiration model to estimate METRIC The total and seasonal evapotranspiration with the simulation results of Hydrus-1D still have good consistency. In addition, the data (surface meteorological data, soil moisture data, vegetation structure data, irrigation data etc.) under the condition of enough variation while Hydrus-1D can provide accurate pixel scale of ET, but not by simple statistical methods to the ET pixel is extended to transect or regional scale.

【学位授予单位】：西北农林科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：S181

【相似文献】