【作者】 王万同；

【导师】 孙九林；

【作者基本信息】 河南大学 ， 地图学与地理信息系统， 2012， 博士

【摘要】 区域地表蒸散的变化特征反映了陆面过程中能量和水分收支状况的演变趋势,同时也进一步影响区域气候和水资源总量分布,进而对区域经济发展产生影响。随着遥感技术的发展与逐渐成熟,研究大尺度范围的地表蒸散变得更为便捷。因此,开展基于遥感手段的区域地表能量与水分动态监测研究具有重要的意义和研究价值。本文以遥感技术为主要手段,分别利用ETM+、MODIS两种不同尺度的遥感数据,并引入DEM数字高程数据,对广泛应用的SEBS蒸散模型加以改进,拓展其在复杂地形条件下的适用范围,同时探讨两种不同尺度的遥感数据对蒸散估算精度的影响,并融合ETM+和MODIS数据构建了多尺度遥感模型,改善了MODIS数据的尺度效应(混合像元)问题。本文以伊洛河流域为研究区域进行了实证研究,并对研究区2002年、2008年的地表蒸散时空分布格局进行了分析和探讨,具有比较重要的理论意义和实践价值。研究的主要结论如下(1)利用ETM+和MODIS遥感数据可以较好地估算区域地表蒸散量。其中,利用ETM+数据估算的结果比利用MODIS数据估算的结果精度要高,与地面实测数据的对比分析表明,平均相对误差分别在10%和20%左右,高分辨率的影像数据对蒸散估算的精度有一定程度的提高。(2)基于DEM改进的净辐射通量估算山区模式可以用于地形复杂区域的净辐射通量估算。山区模式考虑到区域海拔高度、坡度、坡向的不同以及周围地形的遮蔽,从三个方面进行太阳直接辐射的地形修正：首先,依据坡度、坡向、纬度等因子修正坡面太阳入射角的余弦值；其次,利用坡面太阳入射角的余弦值与太阳天顶角的余弦值的比值来修正坡面接收的辐射量；最后,引入一个二值函数——地形遮蔽因子来修正邻近地形的阴影遮蔽带来的影响。在研究区的实证结果表明,未考虑DEM情况下的净辐射通量估算结果不能准确反应研究区的地表特征,存在较大偏差,因此在地形复杂条件下的区域净辐射通量估算必须考虑地形的影响。(3)改进的SEBS模型估算的蒸散精度较高,模型在研究区的适应性和稳定性较好。以广泛应用的SEBS模型为基础,引入DEM数据,对坡面上太阳辐射、大气透射率、地表温度、地表粗糙度等地表参数及动力学参数进行了地形(海拔高度、坡度、坡向)修正使得模型更能适应复杂地形条件下的区域地表蒸散估算,拓展了蒸散模型的适用范围。经过在研究区的应用结果表明,改进的SEBS模型估算的日蒸散量和实测值的误差较小,估算精度较高,模型在研究区的适应性和稳定性较好,利用ETM+数据对研究区地表蒸散的估算结果精度要高于MODIS数据,改进的SEBS模型能够反映出研究区的地表覆盖类型和水分分布的时空格局。(4)分析了MODIS数据在蒸散估算中由混合象元造成的尺度误差,并基于空间增强方法构建了多尺度遥感模型,提高了模型估算的精度,使得进行高频率的区域地表能量与水分的动态监测具有可操作性,也拓展了MODIS和TM/ETM+数据的应用范围,提高了数据的利用率。多尺度遥感模型结合TM/ETM+的高空间分辨率数据对MODIS精度较高的地表温度数据进行增强,发挥二者的优势,以得到的30m高分辨率LST为输入参数,提高了模型估算的精度。尤其在F垫面破碎度高、土地利用类型复杂、混合像元严重的区域,多尺度遥感模型可以使估算的蒸散量精度得到一定程度的改善甚至显著提高。同时,MODIS的多时相优势也能得到充分发挥。(5)最后,选择了2002年和2008年四个季节中晴空无云或云量较少的典型日MODIS数据(8d),对伊洛河流域地表蒸散的时空格局进行了分析。研究结果表明,研究区地表蒸散总体特征为地区分布差异大、年内分配不均。研究区地表蒸散量受植被覆盖度、地表温度、地形条件的影响较大,其时空分布特征也与这些因素的变化相一致。流域内地表蒸散的空间分布有三个特点：具有明显的地域差异；地表蒸散量受地形影响显著；地表蒸散受供水条件的影响较大。另外,日蒸散量在空间分布上也与地表覆盖类型的分布具有很好的一致性。日蒸散量与NDVI、地表温度呈线性相关,与海拔高度呈对数相关,与坡度、坡向无明显相关性。未考虑地形影响的日蒸散量平均偏大0.17mm/d,在不同坡向上的变化不大,并随坡度的增加呈增加趋势,这与研究区的实际情形不符,因此,在地形复杂的山区,必须要考虑地形因素的影响。从地表通量及蒸散的年内变化看,季节性差异特征明显,夏季最大,春秋次之,冬季最小,年内分布呈“单峰”变化趋势。2008年的地表净辐射、土壤热通量、日蒸散量均高于2002年,2008年年内日蒸散量差异比2002年大,表明2002年至2008年间,研究区内人类活动的影响加剧,地表蒸散过程逐渐增强,蒸散量逐年增加,年内季节性差异显著。更多还原

【Abstract】 The variation of regional land surface Evapotranspiration (ET) reflects the trend of energy and water balance in land surface process. Also it further affects distribution of the regional climate and water, and thus has an impact on regional economic development. With the development of remote sensing technology, it becomes more easily to explore the land surface ET of large scale. Therefore, it is a promising theme to carry out a research related to monitoring land surface energy and water balance on regional scale based on the remote sensing technique.Based on the remote sensing technique, using ETM+and MODIS, two different scales of remote sensing data, and also bringing DEM data, the thesis improves the generally-applied SEBS model and expands its application in the situation where the topography is complicated. At the same time, it explores how these two different data with different scales affect the accuracy of ET. It also builds a multi-scales remote sensing model by using data of ETM+and MODIS to improve the scale effect (mixed pixel) of the MODIS data. The thesis regards the YILUO River basin as a case, and also analyses and explores the temporal and spatial distribution of land surface ET in the year2002and2008, which has more important theoretical and practical value. Specific research work and conclusions of this paper are as follows.1) ETM+and MODIS can be used to estimate the regional land surface ET better. Especially, the estimate result which uses the ETM+data is more accurate than that uses the MODIS data. Comparison with ground truth data shows that the average relative error is of10%and20%and high-resolution image data can improve ET estimation accuracy to some extent.2) DEM-based net radiation flux estimation mountain area mode can be used in the net radiation flux estimation of complex terrain area.Taking into account of the regional sea level elevation, ratio of slope, aspect and the shelter of the topography around in a mountain area model, it corrects the topography affected by the sun radiation from three aspects:firstly. according to the factors such as ratio of slope, aspect, latitude to correct the cosine of slope angle of the sun; Secondly, using the cosine ratio of the slope angle of the sun and solar zenith angle to correct the radiation quantity the slope received; Finally, it introduces a binary function-terrain masking factor to correct the impact of near terrain shadowing. The empirical results in the experimental area show that the net radiation flux estimation result without the consideration of DEM can not reflect the feature of the land surface, so we must take the effect of the topography into consideration.(3) The improved SEBS model of ET is of higher accuracy, and has a better adaptability and stability in the studied area.Based on the widely-applied SEBS model and introducing DEM data, it carries out the terrain (elevation, slope, aspect) amendments on the surface parameters and Kinetic parameters such as the solar radiation, atmospheric transmittance, land surface temperatures and land surface roughness, which makes the SEBS model can even adapt to complicated geography condition, and expands the application of the regional surface ET model. The application result of the studied area indicates that the daily ET amount estimated by the improved SEBS model has a smaller error with the measured value, and a higher accuracy. The suitability and stability of the model is better. Making use of ETM+data, the surface ET estimation in the studied area is more accurate than a MODIS data. The SEBS model can better reflect the land cover type and pattern of spatial and temporal distribution of water in the studied area.(4) This paper analyzes the size error caused by the mixed pixel in ET estimation with MODIS data. It sets up the multi-scale remote sensing model based on the space enhancement method, and raises the accuracy that the model estimates, making it operable for high-frequency dynamic monitoring of regional land surface energy and water. And also it expands the application of MODIS data and TM/ETM+data, and raises the data utilization ratio.(5) Finally, it selects the MODIS data for typical days of clear sky having no cloud or less cloud during four seasons in2002and2008, and analyzes the temporal and special pattern of land surface ET in YiLuo River. Research result shows that the major characters of the land surface ET in the studied area are the large differences in regional distribution and the uneven distribution of the year. The land surface ET is greatly influenced by vegetation cover, surface temperature and terrain conditions. Temporal and spatial distributions are also consistent with the changes in these factors.Inside the river valley, the spatial distribution of land surface ET has three characteristics:it has apparent regional differences; the land surface ET is greatly affected by topography; the land surface ET is also under the impact of water supply condition. In addition, the spatial distribution of daily ET is also in accordance with the distribution of different land cover types. The daily ET has a linear correlation with NDVI and surface temperature, has a logarithmic correlation with altitude, and has no significant correlation with slope and aspect. Without considering effects of terrain, the daily ET is a large0.17mm/d on the average, and there is little change in different aspects, increasing with the slope’s increasing trend.This is not in accordance with the actual situation in the studied area. Therefore, at mountain area with complicated topography, the terrain factors must be taken into consideration.From the land surface flux and ET changes in the year, the character of seasonal variation is obvious, largest in summer, followed by the spring and autumn, winter in the end. Distribution during the year has a "single peak" trend. The surface net radiation, soil heat flux and daily ET quantity in2008are all higher than those in2002, and differences of daily ET quantity within the year in2008are bigger than those in2002, which shows from2002to2008, the activities of the human beings have an increasing impact in the studied area. There exists a gradual increase in both land surface ET process and ET quantity by the year, and the seasonal variation inside the year is obvious.更多还原