CLIMATE CHANGE, EXTREME PRECIPITATION, AND FLOODING IN THE MIDWEST

Global temperatures have risen by more than 1⁰C when compared to pre-industrial averages due to anthropogenic emissions of greenhouse gasses. Climate change has also brought about changes in storm patterns, including an increase in extreme precipitation events in the midwestern U.S., in part due to an increase in mesoscale convective systems. Mesoscale convective systems are large thunderstorms that extend for at least 100 km in one direction and are common in the spring and summer in the Midwest. The increase in mesoscale convective system frequency and intensity in the Midwest from 1979 to 2014 was not uniform, with some areas seeing more change than others. The impacts of enhanced mesoscale convective systems on flooding in the Midwest is not well known.

To determine if there is a link between mesoscale convective systems and flood frequency in the Midwest, we first queried the USGS stream gauge database for streams in the Midwest that have daily mean discharge data from 1979 to 2014 and have a drainage area of < 150 mi². Google Earth historic imagery was used to determine if there were any dams on these streams, and if the watershed had undergone land use change over the study period. Streams with dams were discarded. A peaks-over-threshold analysis was performed on the remaining streams. For each stream, we selected a discharge representing approximately the 0.5-year flood and then quantified the number of events each year that exceeded that threshold. Next, we performed a Poisson regression on the number of exceedances each year using time as the predictor. A second Poisson regression was also run using only the number of threshold exceedances that occurred in April, May and June, peak mesoscale convective system season. The spatial distribution of these results was compared to the distribution of changes in mesoscale convective system rainfall found by Feng et al. (2016). 115 streams were analyzed, 19 of them exhibited a statistically significant trend in flood frequency (p < 0.05), with 16 showing increased flood frequency and 3 showing decreased flood frequency. During the spring, 27 streams exhibited a statistically significant trend in flood frequency (p < 0.05), 25 of which showed an increase and 2 of which showed a decrease in flood frequency. 11 streams showed an increase in both year-round and springtime flood frequency.