As the old adage goes, when it rains, it pours. To speak less metaphorically, however, sometimes it pours, then stops for a while, then sprinkles, then pours again.
Rainfall is an important component of many environmental models, which are crucial for studying flooding, drought, water quality, and more. Yet model representations of rainfall are typically static, whereas, in reality, rain is often intermittent. A model that fails to account for rainfall that is not continuous will have inaccuracies, which can affect the quality of the conclusions that scientists are able to draw from it.
A new study by Lombardo et al. helps improve our understanding of this problem and illustrates ways to refine representations of rainfall in environmental models.
The team has developed a method for breaking up, or disaggregating, rainfall measurements into many individual blocks of time, each one less than a month long. First, the researchers calculated the probability of rainfall occurring, as well as how that probability would randomly evolve over time.
They then compared their calculations to data collected at a rain gauge in Viterbo, Italy, between 1995 and 2005. The gauge was set up to take rainfall measurements every 30 minutes. This comparison allowed the researchers to test the model’s ability to successfully simulate the behavior of real-life rainfall at more frequent time intervals.
It is one thing to measure the amount of rain a given location receives in a year or a month, but being able to represent when rainfall is occurring on finer timescales is even more useful. This study has produced a more realistic, yet simple, model with which scientists can better simulate rainfall. (Water Resources Research, https://doi.org/10.1002/2017WR020529, 2017)
—Sarah Witman, Freelance Writer