Modeling the temperature of alpine streams

Anyone who has been to the mountains knows that alpine streams are refreshingly cold. But because of a warming climate, their temperatures are on the rise, potentially threatening local biodiversity. Researchers at EPFL have developed a mathematical model that combines meteorological and geomorphologic data to simulate stream flow and temperature in snow-dominated catchments. The results of their model, applied to the alpine Dischma Valley near Davos, Switzerland, have been accepted for publication in the journal Water Resources Research.

River networks connect distant ecological niches, tying them together into complex webs. According to Francesco Comola, the first author of the study, biological species present in each part of the network have preferential temperature ranges within which they thrive. But as the climate continues to warm, alpine streams are following suit. Over the past 20 years, maximum annual water temperatures have increased by as much as 1.5-2 degrees Celsius in some Swiss rivers. This leaves few options for resident organisms that are sensitive to temperature changes: either they adapt to their new environment, they move upstream to higher altitudes, or they are driven out altogether.

Several phenomena act in parallel to give rise to the stream temperature that can be measured where the stream flows out of a catchment. At the land surface, snowmelt water and rainfall infiltrate into the soil at different temperatures. Once underground, the water exchanges heat with the soil it flows through, before eventually entering the stream. Finally, once in the stream, the water is exposed to the atmosphere, where its temperature progressively approaches that of the air.

Today, most models estimate stream temperature based on atmospheric temperature alone. But as Comola explains, this is only valid under very specific conditions. “Air temperature is a good indicator of stream temperature in large catchments where water spends most of its travel time in the stream and temperature is controlled by atmospheric exchanges. The alpine catchments we are interested in are too small for this approximation to be valid,” he says. Accurately determining the stream temperature in such small catchments requires taking into account all of the phenomena mentioned above.

The biggest challenge, says Comola, is modeling what goes on underground. Because it is extremely challenging to model all the local thermodynamic processes that occur as infiltrated rainwater travels through the soil to the stream, Comola and his colleagues used a stochastic approach to simulate the ensemble average response of the system. This involves estimating the time it takes for each volume of water to travel underground until it reaches the stream. During this journey, its temperature approaches the temperature of the soil, matching it only if the journey is long enough.

“With our model, we confirmed a finding that was recently reported by Canadian researchers. During the summer, they observed the stream temperature to be higher than that of the water that flows into the stream from the soil,” says Comola. And the opposite was found to be true in the winter. This finding further underscores the importance of taking into account underground processes to be able to predict stream temperatures.

The WSL Institute for Snow and Avalanche Research has extensively studied the Dischma Valley near Davos, leading to an abundance of meteorological data needed to drive the model. Provided sufficient meteorological data, Comola’s work, which is particularly relevant to ecologists and for the management of water resources, especially in the context of a warming climate, could be applied to alpine catchments anywhere in the Alps.