Climate change increases the risk of landslides

Gilles Grand­jean. Land­slides are mul­ti-fac­to­r­i­al process­es: they occur when the slope is steep enough and the geol­o­gy favourable. Oth­er fac­tors can also play a part: rain­fall, for exam­ple, sat­u­rates slopes and encour­ages their desta­bil­i­sa­tion. So, as rain­fall increas­es, so does the num­ber of land­slides, accord­ing to a cor­re­la­tion that can vary spa­tial­ly¹. As cli­mate change affects pre­cip­i­ta­tion (by increas­ing the fre­quen­cy of extreme events), it also has an impact on the risk of land­slides. How­ev­er, it remains very dif­fi­cult to estab­lish the direct role of cli­mate change in the occur­rence of landslides.

First­ly, because we don’t have com­pre­hen­sive mon­i­tor­ing of land­slides. On the very large land­slides that move reg­u­lar­ly, such as Super-Sauze or La Clapière in the French Alps, a large num­ber of sen­sors are installed: lasers, teleme­ters, radar, weath­er sen­sors and so on. How­ev­er, many oth­er moun­tain slopes are not equipped. More­over, land­slides are extreme­ly com­plex process­es. We use dig­i­tal mod­els to under­stand them bet­ter. Dur­ing a land­slide, the physics of mate­ri­als can be dif­fi­cult to analyse: the same event can be sub­ject to the phys­i­cal laws of brit­tle, vis­cous or even flu­id mate­ri­als, which adds com­plex­i­ty to our models.

Here again, it is dif­fi­cult to give a pre­cise answer because of the lack of data. There are no glob­al data­bas­es that exhaus­tive­ly record all land­slides. In France, the Land­slide Data­base (BDMvt) suf­fers from this prob­lem, par­tic­u­lar­ly in unin­hab­it­ed areas where these events are rarely record­ed. Teams are work­ing on the use of satel­lite imagery to sys­tem­at­i­cal­ly detect land­slides, but fur­ther devel­op­ments are still needed.

On the oth­er hand, in small areas where sen­sors have been installed, we are see­ing the effects of cli­mate change. We are see­ing an increase in the fre­quen­cy of grav­i­ty insta­bil­i­ties, which are very close­ly linked to pre­cip­i­ta­tion. We are also see­ing the base of glac­i­ers melt­ing as a result of ris­ing tem­per­a­tures, releas­ing a lot of sed­i­ment – a col­lec­tion of mate­ri­als such as grav­el and sand – which accu­mu­late in the sur­round­ing tor­rents. When it rains heav­i­ly, the sed­i­ment-laden water rush­es down the slopes, caus­ing much more dam­ag­ing floods down­stream and increas­ing the risk of cre­at­ing logjams.

The sci­en­tif­ic com­mu­ni­ty is cur­rent­ly look­ing into this sub­ject. Moun­tain areas are grad­u­al­ly adapt­ing to changes in cli­mate: slopes are being mod­i­fied by ero­sion, plant species are chang­ing, and so on. But it is chang­ing so rapid­ly that we fear that the slopes will not have time to adapt. This could lead to cat­a­stroph­ic events, includ­ing land­slides. On the oth­er hand, the direct impact of human activ­i­ties may lim­it the nat­ur­al adap­ta­tion of moun­tain regions.

The risk of land­slides is also affect­ed by human activ­i­ty. Slopes can be weak­ened by the con­struc­tion of infra­struc­ture (for exam­ple, if they are cut into to build a road) or by land clear­ance. Anoth­er effect that can be ampli­fied by human activ­i­ty is bank ero­sion. When ero­sion is severe, river­banks can become unsta­ble and small land­slides can occur. Through a regres­sive effect, they can gen­er­ate a larg­er land­slide across the entire slope.

Final­ly, risk is defined as the com­bi­na­tion of a nat­ur­al phe­nom­e­non and the vul­ner­a­bil­i­ty of inhab­it­ed areas: even if a land­slide occurs, with­out infra­struc­ture or peo­ple, the risk is zero. By build­ing more infra­struc­ture in moun­tain­ous areas, vul­ner­a­bil­i­ty is mul­ti­plied, and the risk is increased.

Gen­er­al­ly speak­ing, the risk of land­slides is not going to decrease, due to the increase in the num­ber of infra­struc­tures in moun­tain areas. But it is impos­si­ble to draw any gen­er­al con­clu­sions. In a study pub­lished in 2018, we assessed future risks in Pyre­nean and Alpine val­leys on the basis of IPCC sce­nar­ios and socio-eco­nom­ic sce­nar­ios devel­oped with local stake­hold­ers².The results are very var­ied: the risk increas­es in some areas, but not every­where. It all depends on the ori­en­ta­tion of the val­ley, its lat­i­tude, its vul­ner­a­bil­i­ty, etc. Assess­ing the impact of cli­mate change requires stud­ies spe­cif­ic to each valley.

 The sci­en­tif­ic com­mu­ni­ty is heav­i­ly involved in these issues, par­tic­u­lar­ly with­in the pri­or­i­ty research pro­gramme and equip­ment (PEPR) Risques coor­di­nat­ed by BRGM with CNRS and Greno­ble Alpes Uni­ver­si­ty. We are also work­ing on moun­tain risk projects, such as IRIMONT and the ANR with VIGIMONT, devel­op­ing ear­ly warn­ing sys­tems in particular.

There are solu­tions to reduce the risk, but there are also prob­lems. For exam­ple, some nature-based solu­tions involve leav­ing nat­ur­al areas around rivers to absorb floods, or replant­i­ng slopes, which can lim­it the devel­op­ment of eco­nom­ic activ­i­ties. The oth­er prob­lem con­cerns the vul­ner­a­bil­i­ty of infra­struc­ture. When works are car­ried out to reduce the like­li­hood of land­slides, this can encour­age more to be built. But these struc­tures – fences, retain­ing walls, etc.- require main­te­nance. In the absence of main­te­nance, and if the infra­struc­ture fails, the risk of land­slides is even greater than before because of new con­struc­tion. What’s more, these struc­tures are some­times adapt­ed to a past cli­mate and the stan­dards no longer cor­re­spond to cur­rent or future cli­mat­ic hazards.

Interview by Anaïs Marechal