利用CDOM吸收系数估算博斯腾湖水体表层DOC浓度

基于2021年夏秋两季西部干旱区博斯腾湖表层50个采样点位实测数据，通过对DOC特征的统计学分析、内外源因素影响的时空差异以及电导率相关性验证，将博斯腾湖划分为河口和非河口区域.然后以CDOM特征波长吸收系数a₂₅₀和a₃₆₅为自变量，以DOC浓度（c_DOC）为因变量，分别构建了基于CDOM的河口和非河口水域DOC浓度估算模型.结果表明：河流与湖泊进行水体交换的同时，会将大量陆源DOC输送到博斯腾湖中，使得河口区域DOC浓度明显高于非河口区域（t-tests，P<0.01），河口区域的DOC浓度（c_DOC）约是非河口区域的2.2~2.3倍，且在河口区域c_DOC与电导率呈现显著相关关系（夏季：R²=0.81，P<0.01；秋季：R²=0.84，P<0.01）.本文构建拟合模型（c_DOC=α+β·α₂₅₀+γ·α₃₆₅），并通过交叉验证的方法来检验模型精度.将夏季和秋季同区域数据统一建模，河口和非河口区域CDOM与DOC均存在较好的相关关系，且模型精度较高（河口区域：R²=0.60，RMSE=8.56%；非河口区域：R²=0.66，RMSE=8.77%）.本文所建立的模型可以在不增加环境因子变量的前提下提高精度，有利于实现卫星遥感反演.同时，本研究揭示了河流输入对博斯腾湖DOC分布和估算的时空影响，提出可利用CDOM估算DOC浓度，但需根据水文特征和c_DOC等因素区分河口和非河口区域.本研究对实现新疆水资源合理开发、有效保护以及综合治理提供科学依据，对我国西部干旱区湖泊DOC遥感动态监测具有重要意义.

Estimation of DOC concentration in the surface water of Lake Bosten using CDOM absorption coefficient

This paper divide Lake Bosten into estuarine and non-estuarine regions based on 50sampling sites in the surface layer of Lake Bosten in the western arid zone in summer and autumn 2021, through statistical analysis of DOC characteristics, temporal and spatial differences in the influence of internal and external factors, and electrical conductivity correlation verification. Then, with the CDOM characteristic wavelength absorption coefficients a₂₅₀ and a₃₆₅ as independent variables and DOC concentration (c_DOC) as the dependent variable, the CDOM-based estimation model of DOC concentration in estuarine and non-estuarine waters was constructed respectively. The results showed that the exchange of water between the river and the lake would transport a large amount of terrestrial DOC into Lake Bosten, making the DOC concentration in the estuary area significantly higher than that in the non-estuarine area (t-tests, P<0.01), and the DOC concentration in the estuary area was about 2.2~2.3times that in the non-estuarine region, and c_DOC was significantly correlated with conductivity in the estuarine region (summer:R²=0.81, P<0.01; autumn:R²=0.84,P<0.01). This paper construct a fitting model(c_DOC=α+β·α₂₅₀+γ·α₃₆₅), and test the accuracy of the model by cross-validation. Unified modeling of the same area data in summer and autumn, there is a good correlation between CDOM and DOC in both estuary and non-estuarine areas, and the model accuracy is high (estuary area:R²=0.60, RMSE=8.56%; non-estuarine area:R²=0.66, RMSE=8.77%). The model established in this paper can improve the accuracy without increasing the environmental factor variables, which is beneficial to the realization of satellite remote sensing inversion. At the same time, the spatiotemporal effects of river input on DOC distribution and estimation in Lake Bosten were revealed, and it was proposed that CDOM could be used to estimate DOC concentration, but estuarine and non-estuarine areas should be distinguished according to factors such as hydrological characteristics and c_DOC. This study provides a scientific basis for the rational development, effective protection and comprehensive management of water resources in Xinjiang, and is of great significance to the dynamic monitoring of lakes in the western arid zone of China.