Understanding extreme precipitation events in climate simulations of the twentieth and twenty-first centuries in Madison, Wisconsin

Understanding extreme precipitation events in the current climate system, as well as the future, is an important aspect of assessing the impact of climate change for both adaptation and mitigation purposes. The current study investigates the character of extreme precipitation events over Madison, Wisconsin during the late twentieth and late twenty-first centuries using reanalysis data, along with 19 coupled ocean-atmosphere general circulation models that participated in the Intergovernmental Panel on Climate Change Fourth Assessment Report (IPCC AR4).

Results show an increase of approximately 9.5% in the multi-model average of annual precipitation received in Madison by the end of the twenty-first century with the largest increases projected to occur during Northern Hemisphere spring (MAM) and winter (DJF). In addition, results show the observed seasonal cycle of precipitation in Madison is not accurately captured by the models. The multi-model average shows a strong seasonal peak in May, whereas observations indicate a mid-summer peak. Also misrepresented in model simulations of the current century is the seasonal cycle of extreme precipitation, namely the wettest five percent and wettest one percent of all days in Madison. Model results show a great deal of disagreement in the temporal distribution of the extreme precipitation events, with a preference towards a bimodal distribution peaking during spring and autumn. The annual cycle of precipitation in late twenty-first century projections over Madison shows a seasonal peak occurring in the spring, which is consistent with twentieth century simulations but different from observations. Although issues in the timing of extreme events exist, results show an increase in the magnitude of extreme precipitation in the wettest ten, five, and one percent of days by the end of the twenty-first century. The largest increase in magnitude occurs in the heaviest precipitation category (wettest one percent), which is approximately 15% by the late twenty-first century.

A quantitative analysis of vertically integrated moisture flux convergence (MFC) is used to investigate the physical mechanism responsible for producing extreme precipitation events in model simulations for the twentieth and twenty-first centuries. Spatial composites of MFC during the wettest one percent of days from the late twentieth century simulations agree well with results from the North American Regional Reanalysis (NARR) dataset. A budget of the moisture flux convergence term during the wettest one percent of days indicates the convergence component dominates over the advective component, representing 64% and 72% of the total MFC for the late twentieth and twenty-first centuries, respectively. Changes in the magnitude of average MFC between the twenty and twenty-first centuries are found in all models. The correlation between the change in MFC and the change in extreme precipitation is greater than the correlation between the change in surface specific humidity and the change in extreme precipitation. The multi-model average of MFC shows an increase of 50% by the end of the twenty-first century.