基于多维光谱空间的干旱指数

Multi-dimensional spectral space based drought indices

遥感干旱指数是用遥感数据定量描述区域受旱程度以及作物产量灾损的机理性数学模型.干旱指数在农业、林业和生态环境监测方面有较高的研究价值与非常广阔的应用前景,是当前的研究前沿和热点之一.介于遥感干旱指数起步至今还未形成研究系统,开展了这方面工作的梳理和归纳.首先,对多维光谱特征空间概念作延伸,将多维光谱特征空间的多维变量由光谱值拓展为与一切干旱相关的遥感产品,将光谱特征统一归纳到多维坐标框架当中,使得干旱指数的数学建模意义更明确,构造更清晰;其次,对30种干旱指数进行归纳,从"三要素"的角度将干旱指数分为四类,并评价其各自优势和局限性,将干旱遥感监测模型家族抽象为一个金字塔形式,从底层到顶层分析模型家族的一般构成要素和复合形式,解析干旱指数通用的数学构建规律及其物理内涵;最后,分析大气影响、土壤、传感器差异等环境影响因素,指出干旱指数未来发展重点在数据源、监测对象、表达形式三个方面.通过对干旱指数的研究工作有助于进一步完善发展干旱指数体系,推进干旱遥感监测模型在农情遥感监测、水资源规划和环境科学管理等领域的应用.

Multiband drought indices （Dis） have been developed for quantitatively characterizing the degree of region drought-hit and crop yield loss. Nowadays, drought indices are currently one of the hot topics with potential applications in agriculture, forestry and ecological environment monitoring. However, it is still lack of a systematic research for the multiband drought indices. This paper made a review of latest remote sensing derived drought indices. The concept of multi-dimensional spectral space was extended, where multi-dimensional variables are expanded from the remote sensing spectral bands to all the satellites derived products concerned with drought. Over 30 drought indices developed during the past years were categorized by constitution elements into four groups. Advantages and limitations of each group were an- alyzed. Further more, the family of drought indices was viewed as a ＂pyramid＂, and the whole family＇ s composition were deconstructed from the bottom to the top. Environmental impacts like atmosphere, sensor calibration and soil effects on the drought monitoring are analyzed and expressed as improved factors into the models.