A new study from Pennsylvania indicates that extreme short-duration rainfall events in spring are growing larger and happening more often

Heavy rains and flash flooding have long been a concern in the US (Schumacher, 2017; Marvel et al., 2023). As our climate continues to warm (Leung et al., 2023), some of the largest rainfall events – known as extreme rainfall events – seem to be intensifying.

Such events are now dumping more rain than they did in the past. As the saying goes: when it rains, it pours. A growing body of research reports that extreme rainfall is indeed intensifying around the world (Fowler et al., 2022). Most evidence points to the intensification of daily rainfall extremes (Westra et al, 2013). Yet, subhourly rainfall extremes – those that occur in less than an hour – appear to be increasing faster than hourly and daily rainfall extremes (Westra et al., 2014; Fowler et al., 2022). 

Why are we seeing more intense rainfall? 

One explanation lies in a nearly 200-year-old equation called the Clausius-Clapeyron (CC) relationship. According to the CC relationship, the amount of water the atmosphere can hold increases by 7% for each 1.8 °F (1 °C) rise in temperature (Witze, 2018; Means, 2021). Generally, extreme daily rainfall increases with temperature at roughly the CC rate of 7% per 1.8 °F (Westra et al, 2013). Scientists refer to this as CC scaling. In some cases, subhourly extremes increase at rates greater than the CC relationship would suggest (Guerreiro et al., 2018; Schroeer and Kirchengast, 2018). Scientists call this super-CC scaling. 

To date, many studies worldwide have assessed changes in subhourly rainfall extremes. But only a few of these studies have come from the US. Thus, there is a unique opportunity for researchers to describe how subhourly rainfall extremes are changing in the US. If subhourly rainfall extremes are in fact increasing, then where are they increasing, by how much, and at what time(s) of year? 

Springtime subhourly rainfall extremes appear to be intensifying in Mahantango Creek

From 1968 to 2022, Mahantango Creek experienced an increase in the magnitude and frequency of subhourly (15-minute) rainfall extremes in the spring. We also found a notable uptick in the relationship between subhourly rainfall extremes and temperature in the watershed. 

This uptick suggested a possible role of thunderstorms at higher temperatures. Unfortunately, thunderstorm conditions were not directly tracked in Mahantango Creek over the entire 55-year study period. So, we explored rainfall contributions from thunderstorms using variables that indirectly represented their occurrence. In the following sections, we delve into these findings in a bit more detail. 

Long-term increases in magnitude and frequency

Changes in the magnitude (size) and frequency of rainfall extremes often go hand in hand (Slater et al., 2021). Our findings about springtime subhourly rainfall extremes supported this idea. 

For example, the largest subhourly (15-minute) rainfall events in the spring increased by 0.17 inches over the 55-year study period (Figure 2, left panel). To put a finer point on it, an extreme 15-minute rainfall event in 1968 might have produced around 0.25 inches of rain. In 2022, that same 15-minute rainfall extreme might have generated closer to 0.42 inches of rain. 

Subhourly (15-minute) rainfall extremes were also occurring more often (Figure 2, right panel). In our study, we defined a 15-minute rainfall extreme as any 15-minute rainfall amount that exceeded 0.3 inches of rain. Such extremes represented the top 1% of 15-minute rainfall amounts over the 30-year baseline period from 1971 to 2000. Using this method, we estimated that there were 1.65 more 15-minute rainfall extremes in 2022 than in 1968 in Mahantango Creek.

Super-CC scaling at temperatures above 52 °F

Increases in temperature and atmospheric water availability exert strong controls on rainfall extremes (Westra et al., 2014). Scientists often use dew point temperature to capture these effects because the dew point represents the combined influence of temperature and humidity. 

In the spring, we found that 15-minute rainfall extremes increased at roughly the CC rate (~6% per 1.8 °F) up to dew point temperatures of around 52 °F (Figure 3). Above this temperature, 15-minute rainfall extremes increased at more than twice the CC rate (~18% per 1.8 °F). Such shifts from sub-CC to super-CC scaling are thought to indicate changes in the types of storms causing extreme rainfalls (Haerter and Berg, 2009; Berg et al., 2013). For instance, large storm systems with stratiform rain of light to moderate intensity often occur at lower temperatures. As temperatures rise, smaller scale convective storms like thunderstorms increase in importance.

A possible growing role of thunderstorms

The spring season in the Northeast is a period of transition from cold, winterlike weather to the warmer, more humid weather of summer. As climate change pushes springtime temperatures higher, there are some indications that the thunderstorm season will expand due to increased periods of warming-induced instability and moisture availability. In Mahantango Creek, many of the largest 15-minute rainfall amounts in the spring were due to lines of intense thunderstorms passing through the watershed (see Figure 4). Indeed, thunderstorms are well known producers of extreme rainfall in the US  (Schumacher, 2017; Schumacher and Rasmussen, 2020). Could thunderstorms be affecting the intensification of subhourly rainfall extremes in Mahantango Creek?  

To address this question, we first tested for trends in a simple indicator that measured the amount of daily rainfall that could be due to convective storms like thunderstorms. In essence, we divided the largest 15-minute rainfall amount each spring by the daily rainfall total to which it contributed (following Barbero et al, 2019). Notably, the Mahantango Creek watershed saw this rainfall fraction rise from 0.36 in 1968 to 0.55 in 2022 (Figure 5, left panel). This increasing trend indicated that rainfall extremes in the spring were concentrating in smaller parts of the day.

We then turned our attention to convective available potential energy (CAPE), which is measured by weather balloons and simulated by global weather models. CAPE reflects the amount of energy that is available to fuel thunderstorms, and it is typically reported in units of Joules per kg. We used CAPE to assess favorable thunderstorm environments, which we defined as the number of hours each spring with CAPE greater than 150 Joules per kg (Taszarek et al., 2021). Such environments rose by 0.7 hours per year in the spring in Mahantango Creek (Figure 5, right panel). 

Taken together, these results implied that thunderstorms may be playing an increasingly important role in the intensification of springtime 15-minute rainfall extremes in Mahantango Creek.

Relevance of findings and broader implications

Our study sheds light on the changing nature of short-duration rainfall extremes in the Mahantango Creek watershed. Long-term rainfall data from Mahantango Creek clearly suggest that subhourly rainfall extremes are intensifying in the spring. However, our findings come with some caveats: 

• We are unable to determine whether the trends in Mahantango Creek represent a broader regional pattern, or if they are more locally relevant. Thus, there is an opportunity to study other sites with long-term subhourly rainfall observations. Such sites include USDA’s networks of benchmark agricultural watersheds (Goodrich et al., 2021) and experimental forests and ranges (Mukherjee et al., 2023), as well as the US Climate Reference Network (Diamond et al., 2013) and NOAA’s COOP network (Lawrimore et al., 2020). 

• We also cannot say if the trends in Mahantango Creek will continue. Indeed, our results reflect historical changes in subhourly rainfall extremes that are based on rainfall observations made during the 55-year study period. Such results do not offer predictions about the future. Instead, scientists are using regional climate models that simulate heavy rain from small scale storms like thunderstorms (Prein et al., 2015). These models can predict how short-duration rainfall extremes might evolve this century (Prein et al., 2017).

As new research is undertaken with observations and models, we expect to learn much more about the future intensification of short-duration rainfall extremes in the US.

These caveats aside, findings from our study do raise several practical concerns for farming interests in the vicinity of the Upper Chesapeake Bay watershed where Mahantango Creek is situated. Spring is critical season for farmers that involves field work, planting, and fertilization. Extreme rainfall events have the potential to upend many of these farming activities (Walsh et al., 2020). For instance, extreme rains early in the spring season may increase the risk of erosion from bare fields that have yet to be planted. Later in the spring, extreme rains may damage newly planted crops (Rosenzweig et al., 2002). Extreme rains also pose acute risks to water quality (Costa et al., 2022). This is especially true if they occur shortly after manures and fertilizers are applied to fields. 

As a result, we need to continue investigating and updating trends in short-duration rainfall extremes using the latest available data. Such information will be vital to farmers (and foresters) who are trying to cope with changes in rainfall extremes in a rapidly warming climate.