Extreme rainfall: can we anticipate the risk of flooding?

Jan Polch­er. Cli­mate change influ­ences the water cycle. As a col­league described it best: with cli­mate change, dry regions become dri­er and wet regions become wet­ter¹. This is explained by a well-known phys­i­cal process: the Clau­sius-Clapey­ron rela­tion­ship. An increase in the sur­face tem­per­a­ture of the atmos­phere leads to an increase in the amount of water vapour in the atmos­phere, of around 7% for each addi­tion­al degree.

As a result, pre­cip­i­ta­tion is more intense. Accord­ing to the lat­est report by the Inter­gov­ern­men­tal Pan­el on Cli­mate Change (IPCC), there has been an increase in the fre­quen­cy and inten­si­ty of extreme pre­cip­i­ta­tion events on a glob­al scale across all con­ti­nents since the 1950s, par­tic­u­lar­ly in Europe, North Amer­i­ca and Asia². But it should be not­ed that our under­stand­ing of this sub­ject has changed lit­tle in recent decades, and we can­not say that we have a good under­stand­ing of the impact of cli­mate change on extreme precipitation.

It has to do with the very nature of extreme pre­cip­i­ta­tion: it is both high­ly inter­mit­tent and geo­graph­i­cal­ly vari­able. It can rain heav­i­ly in one val­ley for a few hours, and not at all in the next. Our rain­fall obser­va­tion sys­tem is not at all suit­ed to record­ing this type of weath­er event, which requires the instal­la­tion of many instru­ments such as rain gauges. While some regions, such as Europe and North Amer­i­ca, are bet­ter equipped with instru­ments, this is not the case in many areas, par­tic­u­lar­ly trop­i­cal zones.

This lack of data is com­pound­ed by the per­for­mance of the com­put­er mod­els used to sim­u­late and bet­ter under­stand the cli­mate. Con­ven­tion­al mod­els sim­u­late the plan­et’s cli­mate by divid­ing it into grids of around a hun­dred square kilo­me­tres each, a res­o­lu­tion that is far too high to sim­u­late extreme rain­fall. We are work­ing to reduce this scale, but it is a colos­sal sci­en­tif­ic and tech­ni­cal challenge.

It is extreme­ly dif­fi­cult. The phys­i­cal process­es involved – such as the Clau­sius-Clapey­ron rela­tion­ship – will always be the same, and we can rely on them for our fore­casts. But there are many oth­er process­es that make a rain­fall event a flood: they can be bio­log­i­cal (such as veg­e­ta­tion), chem­i­cal (such as the num­ber of aerosols) or human (such as land use). 

All these para­me­ters are chang­ing at the same time as cli­mate change, so it is very dif­fi­cult to pre­dict hydro­log­i­cal trends, par­tic­u­lar­ly extreme rain­fall. The only wide­ly accept­ed out­come is the glob­al trend towards an increase in the fre­quen­cy and inten­si­ty of extreme rain­fall as the cli­mate warms. But this does not pro­vide any infor­ma­tion about local or sea­son­al impacts.

No, every­one is affect­ed. The impact of glob­al warm­ing on the water cycle is the most sig­nif­i­cant for human soci­eties. One of the main prob­lems is that mankind has learned to con­trol hydro­log­i­cal resources since Antiq­ui­ty, and this has been fun­da­men­tal to the devel­op­ment of mod­ern soci­eties. But hydraulic struc­tures – designed to store water and con­trol floods – are designed for a cli­mate of the past. Now, with today’s cli­mate (and that of the future) being so dif­fer­ent, our infra­struc­tures are no longer adapt­ed, and we are los­ing this control.

Flood­ing and extreme rain­fall are two dif­fer­ent con­cepts. Extreme rain­fall does not always cause flood­ing, and vice ver­sa. The risk of flood­ing depends on the cli­mate, but also on the direct impact of human activ­i­ties. This is an impor­tant fac­tor: in Europe, most of the hydro­log­i­cal vari­a­tions observed can be explained by the anthropi­sa­tion of the hydro­log­i­cal cycle³. This can be seen, for exam­ple, in soil seal­ing, which increas­es the risk of flood­ing, while cer­tain hydraulic struc­tures can reduce it. Many anthro­pogenic fac­tors influ­ence the con­ti­nen­tal water cycle – irri­ga­tion, urban­i­sa­tion, man­age­ment of riv­er nav­i­ga­bil­i­ty, hydro­elec­tric­i­ty, etc. It is very dif­fi­cult to sep­a­rate the effects of these fac­tors. It is very dif­fi­cult to sep­a­rate the impact of cli­mate change from that of human activ­i­ties on the hydro­log­i­cal cycle, which makes future flood pro­jec­tions even more complex.

It’s almost impos­si­ble to antic­i­pate it on a local lev­el. All we know, as the IPCC points out, is that dur­ing peri­ods of heavy pre­cip­i­ta­tion, the inten­si­ty is increased by the pres­ence of more water vapour in the atmos­phere⁴. But as floods are also affect­ed by human activ­i­ties, as well as oth­er cli­mat­ic phe­nom­e­na (melt­ing glac­i­ers, ris­ing sea lev­els, drought), it is impos­si­ble to know whether the risk of flood­ing will increase or even decrease in the future, for a giv­en location.

Interview by Anaïs Marechal