PROJECTIONS OF EXTREMES FOR BARCELONA, BRISTOL AND LISBON: ARE WE READY?

The analysis of extremes in meteorology and climatology presents two major problems that should be considered in every study.

The first problem is the definition of ‘extreme event’ itself. The definition necessarily assumes a low occurrence probability, and in most studies, this is combined with potential high-impacts on the studied area. In the framework of the RESCCUE project, when analyzing future climate scenarios it is important to consider different potential impacts because, in some cases, a low-recurrent event may cause an unappreciated impact. For instance, the rare snowfall events in Lisbon are not relevant because do not cause any problem. In any case, the most important element of the definition, the low occurrence, is related to the tails (extremes) of a given probability distribution (usually from the analysed climate variable). For a given distribution, the problem is to determine where each tail starts. Some authors use a threshold that splits a tail from the general distribution, but an arbitrary threshold could occur too many times or never. Other authors use classical quantiles as 0.90, 0.95 or 0.99 to determine the extremes from a distribution. Generally, these values correspond to relatively frequent events (happening every year) and then they usually do not cause potential impacts. The RECCUE project uses return periods (which correspond to very extreme quantiles close to 1) representing well the low occurrence required. This is combined with specific thresholds to indicate if an extreme event can cause problems in each city.

However, the low recurrence of the extreme events has a second problem: There are a limited number of observed extreme events and then statistical measures are less robust. To reduce the uncertainty, it is advisable to use a multi-method and multi-model ensemble. In the RESCCUE project FIC, Fundación para la Investigación del Clima, has been working during almost two years on generating future projections of climate extreme events for Barcelona, Lisbon and Bristol according to a set of 20 climate projections downscaled at local scale (in several station for each research site) and analysed by using several theoretical methods.

The RESCCUE results on climate change projections lead to a more extreme heat in the three cities, with a tendency towards more extreme rainfall behaviour. On one hand, extreme temperature could rise up to +5.0 +- 2.5°C and heat wave days will experience an increase from 5 to 40 days per year in Barcelona, 20 in Bristol and 17 in Lisbon, according to the median of the climate projections by the year 2100. On the other hand, daily and subdaily extreme rainfall (with at least 2y-return period) will increase in Barcelona and Bristol about +30%. For Lisbon, this change is expected only for 1-hour (or shorter) events. Moreover, extreme snowfall could also increase for 100y-return events in Barcelona, up to 40% during the next two decades. For the lesser extreme events (return periods from 2 to 10 years), snowfall would suffer a great decrease in Bristol and Barcelona due to the temperature rising.

Although significant changes in pluviometric drought are not expected, the water shortages (hydrological drought) will incerase due to a greater evapotraspiration.

Regarding the windstorms, extreme gusts could increase in Barcelona up to 10±3% in the next two decades for all return periods, while storm surge is expected to rise in Bristol for 2y-return period by 2100. However, non-significant changes are projected for the extreme wave heights in the RESCCUE cities.