Exploring extreme floods in two Italian watersheds through unseen ensemble scenarios

Panàro and Reno are two rivers in N-Italy having adjacent mid-sized watersheds. Given their spatial proximity they are expected to respond similarly to meteorological forcing leading to extreme floods. Three different river discharge (RD) datasets are analyzed: 1) the historical daily-mean observed timeseries by ARPAE in the stations of Bomporto (Panàro) and Casalecchio (Reno); 2) the EFAS5 historical simulations of daily-mean, computed using the LISFLOOD hydrological model forced with the EMO1 dataset (1992-2023); and 3) the EFAS seasonal-reforecast daily-mean, computed by LISFLOOD forced by the 25-member SEAS5 ensemble (2000-2023).

The UNSEEN approach (UNprecedented Simulated Extremes using ENsembles) uses many simulations to identify “plausible” extremes. Building upon this approach, an ensemble of 100 “surrogate” timeseries is built by concatenating 4-month-blocks from the original 25-EFAS members, skipping the first 3-months of each timeseries to ensure statistical independence. This is done starting from four different initial dates (beginning of April, May, June, and July). RD in two stations are compared by inspecting quantiles, interannual trends, and annual cycles. It is found that the 100 surrogate timeseries in Bomporto can produce RD-extremes higher than those seen in the historical EFAS dataset. The surrogate timeseries show a mostly bell-shaped distribution, centered around zero, for the slope of the interannual trend, while both the historical EFAS timeseries and the ARPAE observations show a positive trend, in recent years. With respect to the annual cycle, autumnal extremes appear to be strongly underestimated in the surrogate timeseries (particularly at Casalecchio), due to an underestimation of the local precipitation from September to December.

Possible relations with RD are explored by examining the meteorological synoptic conditions associated with the most extreme floods, in both the historical-ERA5 and in the surrogate-SEAS5 worlds. The meteorological conditions leading to floods in Bomporto and Casalecchio (only 33 km apart) seem different: the most extreme events at Bomporto are typically associated with deep trough over the Mediterranean or a cyclone over Central-Italy, while those at Casalecchio seem to be linked to a less meridional flow and deeper Atlantic depressions. The frequency of synoptic configurations associated with RD-extremes is studied in both the historical and surrogate worlds, to assess whether the surrogate simulations reveal unseen patterns, or a higher frequency of extreme-RD configurations. Lastly, Z500 and MSLP are used together with local precipitation, to develop a CNN-based statistical model able to characterize the RD-extremes and compare performances on historical and surrogate worlds.