随着水质遥感研究的不断深入，水质遥感的应用也越来越多、越来越广。但是到目前为止，对于水质遥感产品、方法和模型仍没有一个系统的评价指标，在很多水色产品中，只有初级的不确定性估计，而没有针对每个数据产品的不确定性分析，使得很多水色遥感数据产品只是定性而非定量的描述。由于水体遥感反射率信号相对较弱，人们就希望用水体偏振遥感信号作为信息补偿，提高水质参数遥感估算能力。与非偏振信息相比，偏振信息能得到更多目标的特征信息，近几年也被广泛应用于遥感应用中。但是，对水体偏振波谱的测量和水体偏振遥感数据的获取都有相当的难度，无论是常规水质遥感产品还是水体偏振研究，都没有对结果进行的细致的不确定性和敏感性分析，使得对很多产品和结果的描述并不完善，一定程度上影响了水色遥感数据产品的应用和水质遥感的进一步发展。本论文在总结前人工作的基础上，对水质遥感中影响水质参数估算结果的各种不确定性因素的产生和影响方式、程度进行归纳总结，并通过水体辐射传输模型模拟和室内水体偏振测量详细分析了其中较敏感参数对模拟结果和测量值的不确定性影响。重点分析了铜绿微囊藻溶液和泥沙悬浊液浓度变化、水面风速变化、天空云量变化和观测方位变化对离水辐亮度和遥感反射率数据获取的不确定性的影响；通过水体偏振测量实验分析了偏振离水辐射信号对观测方位和水体成分浓度的敏感性。结果发现，非偏振离水辐射信号对观测方位的敏感性随方位的变化而变化，偏振离水辐射和观测方位具有较好的正相关关系；偏振和非偏振离水辐射信号对泥沙悬浊液浓度具有很强的敏感性；天空云量对遥感反射率和离水辐亮度的影响明显；离水辐亮度和遥感反射率对风速变化不太敏感。

Applications water quality remote sensing becomes more and more broad. But water quality information retrieved from remotely sensed data are generally distributed or reported without estimation of their uncertainties. Users have to use water quality remote sensing products as a qualitative descriptor of patterns rather than a quantitative variable.Comparing to remote sensing with normal light, polarized light may contain more characteristic information of the objects, which greatly forced the polarized water quality remote sensing’s development. However, neither normal water quality sensing nor polarized sensing data has corresponding uncertainty estimates, which has obstructed the further development of water quality remote sensing research and further use of water color products.In this paper, the main uncertainty sources in water quality remote sensing and the way in which they affect the measured and retrieved results are presented. The influence of the key uncertainty factors were evaluated by using physical modeling and statistical analysis methods. Studies focused on how the algae concentration, sediment concentration, wind speed, cloud and viewing azimuth variations influence the water-leaving radiance and remote sensing reflectance. Results showed that non-polarized signals are sensitive to viewing azimuth; there is a positive correlation between polarized water-leaving radiance and azimuth. Both polarized and non-polarized water-leaving radiance and remote sensing reflectance are sensitive to the variations of sediment concentration. The cloud influences the total irradiance into water body, thus has a stronger influence on remote sensing reflectance than on water-leaving radiance. Wind speed is not a main uncertainty source in water quality remote sensing.