甘肃河西走廊土地荒漠化严重影响了当地居民的生产生活环境, 高光谱遥感技术是荒漠化土地退化程度、 土地类型识别、 遥感反演等的重要研究手段, 以河西地区荒漠化土地为研究对象, 分析其光谱特征与植被退化程度、 植物类型、 季节变化等的关系, 探讨河西地区荒漠化土地的光谱特征。 主要结果有: (1)当植被覆盖度小于20%, 同一类型不同退化阶段的植被光谱对沙地光谱的影响很小, 沙地光谱反射率与裸地接近, 尤其当植被盖度小于10%时, 沙地与裸地的光谱曲线几乎重合, 仅从植被景观很难反映出土地的沙化程度。 (2)不同的植被类型对沙地光谱的反射率有一定的影响, 以白刺为建群种的沙地光谱反射率较高, 其次为梭梭沙地, 多枝柽柳沙地相对较低, 植被不同演替阶段下指示性植物的沙地光谱可以反映土地的沙化过程。 (3)在植物生长季, 沙地光谱反射率受土壤、 植物含水量及植物物候期的影响, 8月—10月高于其他月份, 7月最低, 沙地光谱波形曲线的季节变化规律可以反映出沙地土壤含水量的变化。 研究结果对荒漠化土地遥感监测中土地沙化程度判定、 季节信息提取、 植被覆盖度估算等提供研究基础。

The desertification of Gansu Hexi Corridor is serious, which seriously affects the environment of living and production activities of local residents. Hyperspectral remote sensing technology is an important research method for desertification land degradation degree, land type identification and remote sensing inversion. This paper takes desertification land in Hexi area as the study object. The object was analyzed for the relationship between its spectral characteristics and vegetation degradation degree, plant type, season and soil type, and the spectral characteristics of desertified land in Hexi region were discussed. The main results are as follows: (1) When the vegetation coverage is less than 20%, the vegetation degradation has little effect on the sand spectrum in different degradation stages of the same vegetation type, and the spectral reflectance of the sand is close to the bare land, especially when the vegetation coverage is less than 10%, the spectral curves of sand and bare land almost coincide, and it is difficult to reflect the degree of desertification of the land only from the vegetation landscape. (2) Different vegetation types have a certain influence on the reflectivity of sandy land spectrum. The spectral reflectance of sandy land with Nitraria tangutorum Bobr as constructive species is higher, followed by Haloxylon ammodendron sandy land, and the Tamarix ramosissim sandy land are relatively low. The sandy spectrum of the indicator plants under different succession stages of vegetation can reflect the desertification process of the land. (3) During the plant growing season, the spectral reflectance of sandy land is affected by soil, plant water content and plant phenology, which is higher than other months from August to October and lowest in July. The seasonal variation of the spectral curve of sandy land can reflect the changes in soil water content in sandy land. The research results provide a research basis for the determination of land desertification degree, seasonal information extraction and vegetation coverage estimation in remote sensing monitoring of desertification land.