An observational study on the evaluation of snow-covered area by remotely-sensed data with different spatial resolutions and on the snowmelt-runoff analysis

为实现本研究的目的，具体分析内容和取得的结果如下：1.利用Landsat/ETM+计算了新近提出的积雪指数，并将其应用于新泻县津南町的针叶林。在野外实地调查的基础上，估算了利用积雪指数区分积雪和针叶林的叶面积指数阈值。结果表明，当叶面积指数小于4时，雪指数可以区分针叶林和雪。（Shimamura等人，2006).2.通过在东京日原村和津南町的实地调查，研究了直接测定叶面积指数（LAI）的方法及其组合对柳杉人工林LAI估算的影响。该过程包括样树的选择、叶片的取样和叶面积的测量。每个过程都有几种方法， ...更多信息 各处理估算的叶面积指数差异不大。然而，这些差异的累积导致了高达40%的估算叶面积指数，即使在经过修剪或间伐的日本雪松种植林中也是如此（Fujiwara等人，2005).3.在Makihata山进行了积雪调查，并在泉水中同步观测了Landsat卫星通过时的积雪情况，从而能够估计Muikamachi河排放站以上的Uono河的雪水资源。Landsat/ETM+的两个场景的分析所揭示的融雪的估计量大于观测径流。这意味着精确分析融雪径流过程的重要性（Shimamura等人，2005).4.评价了卫星数据对监测植被的太阳/传感器几何形状的依赖性的影响。对森林火灾后演替阶段的植被进行了调查。归一化植被指数（NDVI）的差异值在0.1-0.2之间，这是由与太阳/传感器几何形状相关的BRDF引起的。对BRDF，近红外波段的反射率归一化的总和，其他波段是最强大的植被存在的网站。全球环境监测指数在植被稀少的地点最为稳健，而在这些地点，NDVI受到BRDF的强烈影响（Hasegawa等人，2006).5. AMeDAS年平均温度的空间代表性，这是重要的融雪径流分析，使用变异函数进行了研究。在变异函数中，空间代表性的度量值被称为“极差”。研究发现，日本所有地区的年平均温度至少具有100 km的空间代表性（Hadano et al.，2004年）。少

To accomplish the objectives of the present study, the concrete contents of the analyses and the results obtained are summarized as follows.1.The snow index recently proposed was calculated by Landsat/ETM+ and was applied to the coniferous forests in Tsunan-town in Niigata prefecture. The threshold of LAI (Leaf Area Index) to distinguish between snow and coniferous forests using the snow index was estimated based on in situ field survey. As a result, it was clarified that the snow index could discern snow from coniferous forests if LAI was smaller than 4.(Shimamura et al., 2006).2.Effects of differences of the processes of direct measurements of LAI and their combinations on the estimation of LAI for planted stands of Japanese cedar (Cryptomeria japonica D.Don) were estimated by in situ field survey in Hinohara-village in Tokyo and Tsunan-town. The processes consist of selection of sample tree(s), sampling of leaves, and planimetry of leaf area. Each process has several methods, and th … More e differences of estimated LAI in each process were not so large. However, the accumulation of these differences led to as large as 40% of the estimated LAI, even in the planted stands of Japanese cedar which were taken care of such as pruning or thinning (Fujiwara et al., 2005).3.Snow surveys in Mt.Makihata and synchronized observations of the existence of snow with the passage of Landsat satellite were carried out in springs which made it possible to estimate the snow water resources in the Uono river above Muikamachi river discharge station. The estimated amount of snowmelt as revealed by the analyses of two scenes of Landsat/ETM+ was larger than the observed runoff. This implied the importance of precise analyses of snowmelt-runoff processes (Shimamura et al., 2005).4.The effect of the dependence of the satellite data on sun/sensor geometry (BRDF) for monitoring vegetation was evaluated. The vegetation in the succession stages after forest fire was investigated. The difference of Normalized Difference Vegetation Index (NDVI) of the value around 0.1-0.2 was caused by BRDF in relation to the sun/sensor geometry. Against BRDF, the reflectance of the near-infrared band normalized by the sum of other bands was most robust in the site where vegetation existed. Global Environmental Monitoring Index was most robust in the site where vegetation was scarce, while NDVI was strongly affected by BRDF in these sites (Hasegawa et al., 2006).5.The spatial representativity of the annual mean temperature of AMeDAS, which was important for snowmelt-runoff analyses, was investigated using the variogram. In the variogram, the value called as "range" was regarded as the measure of the spatial representativity. It was found that the annual mean temperature had the spatial representativity of at least 100km in all regions in Japan (Hadano et al., 2004). Less