Added value of hourly climate-model ensembles for process-based local-scale water-balance projections
This study demonstrates that using native hourly resolution from convection-permitting climate ensembles, rather than aggregated daily data, significantly alters long-term local water-balance partitioning and drought-stress diagnostics by capturing process-relevant intra-day dynamics, particularly increasing interception evaporation and reducing runoff at forested sites.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine trying to predict how a thirsty forest will drink from a rainstorm. For decades, scientists have looked at the weather in broad, daily strokes, like checking a calendar to see if it rained "today." But nature doesn't wait for a calendar page to flip; it happens in seconds, minutes, and hours. A sudden downpour might soak the leaves of a tree before the water even hits the ground, or a hot, dry afternoon might evaporate moisture before the soil can drink it. This is the world of hydrology—the science of how water moves through the Earth. To understand it, we need to know not just how much water falls, but when and how fast it falls. As the climate changes, these tiny details matter more than ever. If we get the timing wrong, our predictions about droughts, floods, and how much water is left for our cities and farms could be completely off. It's like trying to bake a cake by only knowing the total weight of the ingredients, without knowing if you added the sugar before or after the eggs.
This paper is a deep dive into a new kind of weather forecast: one that speaks in hours instead of days. The researchers used a super-detailed climate model called NUKLEUS, which simulates the atmosphere with a resolution so fine it can "see" individual clouds and storms. They asked a simple but tricky question: Does feeding this hour-by-hour weather data into a water-balance computer program change the story of how water moves through the soil and plants, or is it just a fancy way of saying the same thing we already knew from daily data?
The team, led by Ivan Vorobevskii and colleagues from TUD Dresden University of Technology, ran a massive experiment using a digital model of the water cycle called BROOK90. They tested this model across six different German landscapes, ranging from the snowy Bavarian Alps to the flat grasslands of the North Sea coast. They compared two scenarios: one where the model received weather data every single hour (the "native" high-speed data), and another where that same data was squashed down into a single daily average (the "daily" data). They also tested how "bias correction"—a mathematical fix to make the model's rain match real-world records—changed the results.
Here is what they found, and it's a bit of a plot twist. First, the daily bias correction was a heavy lifter for the amount of water. It changed the total numbers significantly, especially in mountainous areas where the raw model tended to over-predict rain. But when they looked at the timing—comparing the hour-by-hour data against the daily data—the story changed. The total amount of water didn't change much, but where that water went did.
For forests, the difference was dramatic. When the model saw the rain falling in hourly bursts, the trees acted like sponges that got wet and then evaporated the water back into the air before it could reach the ground. The study found that at forested sites, using hourly data increased the amount of water caught by tree leaves (interception) by up to 60% compared to daily data. In Garmisch-Partenkirchen, this meant an extra 193 mm of water evaporated from the canopy, which in turn reduced the water flowing into streams (runoff) by a massive 246 mm. The trees were essentially "drinking" more of the rain before it could run off. However, for grasslands and croplands, the effect was much weaker or even different; the grass didn't have the same "umbrella" effect, so the water mostly just soaked into the soil or ran off as usual.
The researchers also looked at the future, simulating what happens when the world warms by +2 K and +3 K. They discovered that not all parts of the water cycle are equally predictable. The amount of water evaporating from plants (evapotranspiration) showed a clear, consistent signal across all their model runs, especially in winter. But the amount of water flowing into rivers (runoff) was a different beast. The models disagreed wildly on whether rivers would get wetter or drier; in some cases, the models predicted a flood, and in others, a drought, for the exact same location. This suggests that for rivers, the "noise" of the different models is just as important as the "signal" of climate change.
Finally, they checked the "stress" on the plants. They defined a "hot-dry day" as a day when the air is super thirsty (high vapor pressure deficit) and the soil is dry. Using hourly data, the models predicted more of these stressful days, but only because the hourly calculation of "thirsty air" was more accurate. The air felt thirstier when you measured it every hour than when you averaged it over a whole day. However, this didn't automatically mean more drought stress for the plants. The plants only suffered more if the thirsty air happened to coincide with dry soil. The hourly data showed that these two stressors lined up more often in certain years and places, particularly in the forests of central Germany, but not everywhere.
In short, the paper suggests that while daily weather data might be good enough for a rough guess, it misses the subtle dance of water in forests. The hour-by-hour data reveals that trees catch and evaporate a surprising amount of rain before it hits the ground, and that this detail changes how much water is left for rivers and how stressed plants get during heatwaves. The authors conclude that for accurate local water management, especially in forests, we need to stop squashing our weather data into daily averages and start listening to the clock. But they also warn that predicting river flows remains tricky, as the different climate models still tell very different stories about the future.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.