Long-term trends and probability distributions of river water quality variables and their relationships with climate elasticity characteristics.

The statistical characteristics of surface water quality play fundamental roles in climate change-related studies. Climatic drivers will affect surface water quality, and such potential effects will vary between different regions and climate types. Here, we studied the long-term trends and probability distributions of water quality variables (including water temperature, pH, turbidity, DO, C-,N-,P-variables, etc.) and their relationship with climate elasticity, a non-parametric estimator of the sensitivity of the response of water quality to climate drivers, based on three typical watersheds: the Yukon, the Mekong, and the Murray. Significant decreasing trends were observed in the Yukon and Murray watersheds for the majority of water quality variables, except turbidity and filtered nitrate plus nitrite, whereas increasing trends were exhibited by most water quality variables in the Mekong watershed. Compared with the Yukon and Murray watersheds, the probability distributions of most water quality variables and their corresponding percentage change-based elasticity estimator samples are characterized by a heavy-tailed distribution in the Mekong watershed. The precipitation elasticity results are statistically meaningful in the Mekong and Murray watersheds, whereas temperature elasticity is significant in the Yukon watershed. The revealed characteristics of long-term trends and probability distributions pattern of basic water quality variables are helpful for water quality modeling. The findings suggested that the increasing trends, heavy-tailed probability distribution patterns, and the response of water quality to precipitation and temperature, especially in densely populated developing areas, can be modulated by restoration efforts, which will reduce the potential impacts of climatic and non-climatic factors.