Models with more detailed snow physics better captured runoff timing
Researchers compared 13 global models across 1,455 basins where snowfall strongly shapes runoff. Models that represented snow processes in greater detail tended to better capture when meltwater flowed through the basin. The relationship was weaker for total runoff and peak flow and stronger for timing among models that used the same weather inputs.
Science··Evening
Timing improved most clearly with more detail
The researchers tested 13 global hydrology and land-surface models across 1,455 basins where snow is a major part of the water cycle. Models that represented snow processes in more detail generally did better at placing runoff within the melt season. The relationship was much weaker for the season’s total runoff and for the highest flow. It was stronger for timing among models that ran on the same weather data.[1]
Models differed most in how they represent melting
The researchers scored models on selected physical processes. The largest differences lay in how they simulated melting, with further variation in whether snow caught on trees, how much went directly into the air as water vapor and how the models split precipitation into rain and snow. For some processes, including snow caught in the canopy, the researchers observed larger differences when models left them out altogether than when models represented them in simpler ways.[1]
Weather and water storage can shape the comparison
The authors also consider precipitation, the water held in the basin and its passage through rivers when discussing differences in runoff volume. The authors say differences in weather inputs, basin characteristics and model calibration could affect the results. They call for controlled tests to identify which snow processes are most closely linked with runoff timing.[1]