利用卫星遥感技术对水中重金属浓度进行监测具有重要研究意义。 但由于自然水体重金属含量低, 其遥感反演可行性仍被学界怀疑, 以致该领域发展比较缓慢。 鉴于此, 以中国环境卫星(HJ-1A)高光谱成像仪(HSI)反演水中硫酸铁浓度为例, 采用遥感敏感性分析模型(DDE模型), 模拟得到硫酸铁在理论清深水体和三种常见自然水体: 富营养化水体、 浑浊水体与重金属尾矿污染水体, 共计四种水体类型背景下的可遥感反演下限浓度光谱曲线, 明确地给出不同情形遥感反演下限浓度最小值及其所在波长, 进而分析不同水体类型的遥感敏感性变异规律。 研究结果发现, 在不同水体类型背景下, 硫酸铁的遥感反演下限浓度最小值及所在波长位置均是变化的。 在理论清深水体背景下, 该值的最小值为4.63×10^-4 mg·L^-1, 出现在468.530 nm处(对应HSI传感器第5波段, 简写为Band5, 下同), 而扩充至最小值10%增量后, 覆盖的波段范围为460.040~479.600 nm(Band 1—10)。 但在波长范围721.605~951.540 nm(Band 81—115), 下限浓度值已不存在。 同样地, 富营养化水体、 浑浊水体、 重金属尾矿污染水体三种自然水体背景下, 下限浓度最小值分别为: 6.30×10^-2、 2.78×10^-2和1.64×10^-1 mg·L^-1; 对应所在波长分别为: 577.865 nm(Band 46)、 587.900 nm(Band 49)、 669.285 nm(Band 70); 对应的扩充最小值10%增量后的覆盖波段范围分别为555.725~587.900 nm(Band 39—49)、 568.160~612.740 nm(Band 43—56)、 627.895~687.410 nm(Band 60—74)。 以上特征波段, 均是未来硫酸铁遥感反演模型敏感波段集合的重要参考依据。 这四种水体背景下的下限浓度最小值大小关系是: 理论清深水体<<浑浊水体<富营养化水体<重金属尾矿污染水体, 说明相比理论清深水体, 三种自然水体背景下遥感敏感性相对有所下降。 从该最小值出现的波长位置(或范围)来看, 三种自然水体背景下的下限浓度最小值所在波长相比理论清深水体均向卫星图像受大气影响作用越来越小的长波方向移动。 这个波长位置的移动方向由硫酸铁自身的吸收系数光谱和背景水体反射率光谱二者共同决定。 以上研究成果为将来实现水中硫酸铁浓度遥感反演打下了理论基础, 进一步推动水体重金属遥感领域发展。

Using satellite remote sensing technology to monitor heavy metals in water is of great research significance. However, due to the low content of heavy metals in natural water, the feasibility of remote sensing inversion is still doubted by the academic community, resulting in relatively slow development in this field. Because of this, taking the retrieval of iron sulfate concentration in water by the Chinese HJ-1A Satellite Hyperspectral Imager (HSI) as an example, the remote sensing sensitivity analysis model (DDE model) was used to numerically simulate the lower limit concentration spectrums in remote sensing inversion of iron sulfate under the background of four kinds of water, including: theoretical clear deep water and three kinds of common natural water: eutrophic water, turbid water and heavy metal polluted water. The minimum value of the remote sensing inversion lowers the limit concentration, and its wavelength is given. The variation pattern of remote sensing sensitivity under the background of different kinds of water is then analyzed. The research results found that under the background of different kinds of water, the simulated minimum lower limit concentration of iron sulfate and its wavelength position changed. Under the background of theoretically clear deep water, the minimum lower limit concentration is 4.63×10^-4 mg·L^-1, appearing at 468.530 nm (the fifth band of the HSI sensor, abbreviated as Band 5, the same below). After expanding to the minimum 10% increment, the covered band range is 460.040~479.600 nm (Band 1—10). However, the lower limit concentration value no longer exists in the wavelength range 721.605~951.540 nm (Band 81—115). Similarly, under the background of three kinds of natural water: eutrophic water, turbid water, and heavy metal polluted water, the minimum lower limit concentrations are: 6.30×10^-2, 2.78×10^-2, and 1.64×10^-1 mg·L^-1; the corresponding wavelengths are: 577.865 nm (Band 46), 587.900 nm (Band 49), 669.285 nm (Band 70); the corresponding coverage band ranges after expanding the minimum value by 10% are: 555.725~587.900 nm (Band 39—49), 568.160~612.740 nm (Band 43—56), 627.895~687.410 nm (Band 60—74). The above characteristic bands are all important references for the future sensitive band collection of iron sulfate remote sensing inversion models. The relationship between the minimum lower limit concentration under the background of four kinds of water is: theoretical clear deep water<<turbid water<eutrophic water<heavy metal polluted water, which shows that compared with the theoretical clear deep water, the sensitivity of remote sensing under the background of three kinds of natural water has relatively decreased. Judging from the wavelength position (or range) where the minimum value appears, the wavelengths of the minimum retrievable concentrations under the background of three kinds of natural water all move toward the longer wavelength direction where satellite images are less affected by the atmospheric influence than theoretically clear deep water. The moving direction of this wavelength position is determined by both the absorption coefficient spectrum of iron sulfate itself and the reflectance spectrum of the background water. The above research results have laid a theoretical foundation for the remote sensing inversion of iron sulfate concentration in water in the future, further promoting the development of the field of remote sensing of heavy metals in water.