Predicting the seasonal streamflow supply of water in a mountainous basin is critical to anticipating the operation of hydroelectric dams and avoiding hydrology-related hazard. This quantity partly depends on the snowpack accumulated during winter. The study addresses this prediction problem using information from streamflow data and both direct and indirect snow measurements. In this study, the prediction is improved by integrating the data information into a basin-scale hydrological model.

What drives young water fraction, F^*_yw (i.e., the fraction of water in streamflow younger than 2–3 months), variations with elevation? Why is F^*_yw counterintuitively low in high-elevation catchments, in spite of steeper topography? In this paper, we present a perceptual model explaining how the longer low-flow duration at high elevations, driven by the persistence of winter snowpacks, increases the proportion of stored (old) water contributing to the stream, thus reducing F^*_yw.

Under the joint action of climate–human activities the use of runoff data whose mathematical properties have changed has become the key to watershed management. To determine whether the data have been changed, the number and the location of changes, we proposed a change-point detection framework. The problem of determining the parameters of wavelet transform has been solved by comparing the accuracy of identifying change points. This study helps traditional models adapt to environmental changes.

Their flexible growth provides the plants with a strong ability to adapt and develop resilience to droughts and climate change. But this adaptability is badly included in crop and climate models. To model plant development in changing environments, we need to include the survival strategies of plants. Based on experimental data, we set up a simple model for soil-moisture-driven root growth. The model performance suggests that soil moisture is a key parameter determining root growth.

Powerful hybrid models (called δ or delta models) embrace the fundamental learning capability of AI and can also explain the physical processes. Here we test their performance when applied to regions not in the training data. δ models rivaled the accuracy of state-of-the-art AI models under the data-dense scenario and even surpassed them for the data-sparse one. They generalize well due to the physical structure included. δ models could be ideal candidates for global hydrologic assessment.

A simplified hydroclimatic modelling workflow is proposed to quantify the impact of climate change on water discharge without resorting to meteorological observations. Results confirm that the proposed workflow produces equivalent projections of the seasonal mean flows in comparison to a conventional hydroclimatic modelling approach. The proposed approach supports the participation of end-users in interpreting the impact of climate change on water resources.