CHAPTER 3 UNDERSTANDING DISASTER RISK: HAZARD RELATED RISKISSUES - SECTIONII
3 .5 Hydro flog fica fl R fisk :
Lands fl fide
Nficofla Casagflfi, Fausto Guzzettfi, Mfichefl Jaboyedoffff, Farrokh Nadfim, Davfid Petfley
3 .5 .1 Introduct fion
The term flandsflfide encompasses a wfide varfiety off phenomena, ffromthe sfimpfle ffaflfl off rock bflocks ffrom vertfi- cafl rock ffaces,throughtotoppfles and flandsflfides that are domfinated efither by a sflfidfing motfion or by flows off sofifl and/or rock. Landsflfides are strongfly correflated wfith other types off natu- rafl hazards, such as floods, droughts, wfifldfires, earthquakes,tsunamfis and voflcanoes, and are offten finvoflved fin cascadfing events off mufltfihazard dfis- asters.
Cflfimate change,thefincreased suscep- tfibfiflfity off surfface sofifl to finstabfiflfity, anthropogenfic actfivfitfies, growfing ur- banfisatfion, uncontroflfledfland use and the fincreased vuflnerabfiflfity off popu- flatfions and finffrastructure contrfibute to the growfing flandsflfide rfisk. In the Thematfic Strategy ffor Sofifl Protec- tfion (European Commfissfion, 2006), flandsflfides are consfidered one offthe mafin threats to European sofifls. In thfis fframework,flandsflfide dfisaster rfisk
reductfion shoufld be properfly under- taken fin order to reduce the fimpact offflandsflfides on humans, structures and finffrastructures. In areas wfith hfigh demographfic densfity, protectfion works offten cannot be bufiflt owfing to economfic or envfironmentafl con- strafints, and fis fit not aflways possfibfle to evacuate peopfle because off socfietafl reasons. Forecastfing the occurrence offflandsflfides and the rfisk assocfiated wfiththem,and definfingapproprfiate EWSs, are,thereffore, essentfiafl needs.
The te rm ‘ flands fl fide ’ desc r fibes a va r fie ty off p rocesses tha t resu fl t fin the downwa rd and
ou twa rd movemen t off s flope-ffo rm fing ma te r fia fls , finc flud fing rock ,
so fi fl , a r t fific fia fl fi fl fl o r a comb fina t fion off these .
The socfietafl and economficfimpact off flandsflfide rfiskfis dfiffficufltto assess and fit fis underestfimated, sfince a reflevant part off reflated damage fis attrfibuted to other naturafl hazards,fin mufltfihaz- ard chafins (e.g. sefismficaflfly finduced ffafiflures, rafinffaflflfinduced debrfis flows, flahars and rock avaflanches assocfiated wfith voflcanfism).
An estabflfished worfldwfide scfientfific flandsflfide communfity has flourfishedfin theflast decades,thanksto severaflfin- ternatfionafl organfisatfions, such as the Internatfionafl Consortfium on Land- sflfides and the Landsflfide Jofint Tech- nficafl Commfittee, whfich perfiodficaflfly organfisethe Worfld Landsflfide Forums andthe Internatfionafl Landsflfide Sym- posfia, respectfivefly. Reguflar flandsflfide sessfions are aflso organfised at the Generafl Assembfly offthe European Geoscfience Unfion each year.
In thfis subchapter, the mafin causes and trfiggers off flandsflfides and thefir socfioeconomfic fimpact at European flevefl are descrfibed, beffore some gen- erafl concepts and methodoflogfies on flandsflfide zonfing (finventory, suscep-
3 .5 .2
Lands fl fide causes and tr figgers
The mostrecentflandsflfidecflassfifica- tfion fis ffound fin Hungr et afl. (2014). It dfiscerns five mafin types off move- ment: ffaflfls, toppfles, sflfides, spreads and flows. Manyflandsflfides consfist off a varfiety off movement types occur- rfing fin sequence. For exampfle, flarge flandsflfides fin hfigh mountafinous are- as offten start as rock ffaflfls finvoflvfing ffreeffaflflfing rock that detaches ffrom a cflfiffff, whfich upon fimpact at the cflfiffff toe may spontaneousflytransfitfionfinto a very hfigh-energy rock avaflanche (Hutchfinson, 1988). The propertfies off the flow change ffurther as the flandsflfide entrafins or deposfits debrfis and water.
Landsflfides vary greatflyfin sfize. Atthe flargest scafle, a sfingfleflandsflfide canfin- voflve upto some cubfic kfiflometres off rock and sofifls. Atthe other end offthe scafle, a smaflfl bouflder has the poten- tfiaflto causefloss offflfiffe,fifffit strfikes an findfivfiduafl, or to cause mass ffataflfitfies fiff, ffor exampfle,fit causes atrafinto de- rafifl. In generafl,the potentfiaflto cause floss scafles wfith sfize offthe flandsflfide, flargefly because offthe scaflfing offthe kfinetfic energy andthe affffected area. A key causafl ffactor ffor flandsflfides fis thetopographfic settfing offthe poten- tfiafl sfite. In generafl,the propensfityto ffafiflure usuaflfly fincreases as the sflope angfle fincreases, ffrom essentfiaflfly zero
a key gradfient, any gfiven sflopefisflfike- flyto be stabfle under most condfitfions. Sflopes naturaflfly evoflve finto a stabfle state under any gfiven set off envfiron- mentafl condfitfions, prfimarfifly through flandsflfidfing processes. Externafl ffac- tors dfisrupt the sflope equfiflfibrfium to finducefinstabfiflfity;thus, ffor exampfle, a mfigratfing rfiver channefl or an unusuafl flood mayerodethetoe offasflope, fincreasfing the sflope gradfient and the flfikeflfihood off ffafiflure. The sflope wfiflfl then naturaflfly evoflve back to fits sta- bfle gradfientthroughtfime, perhaps by means off another flandsflfide that re- movesthe excess materfiafl.
A second set off causafl ffactors reflates to the type off materfiafl finvoflved fin the potentfiafl finstabfiflfity and fits geo- technficafl propertfies, such as finternafl ffrfictfion and cohesfion. In hard rock masses, stabfiflfityfis usuaflfly defined not bythefintact strength offthe materfiafl but bythejofints, ffractures and ffauflts. The strength offthese dfiscontfinufitfies may be dramatficaflfly flower than the fintact rock strength, especfiaflfly where they are flfined wfith a weaker materfi- afl. Where such a dfiscontfinufity has an orfientatfion that promotes ffafiflure, the resfistance offthe sflope to fland- sflfidfing can be dramatficaflfly reduced. Thereffore, fin many cases, anaflysfis off susceptfibfiflfity depends on an under- standfing offthe rofle pflayed by these dfiscontfinufitfies. Furthermore, the strength off sflope materfiafls degrades throughthe processes off weatherfing, whfich may physficaflfly and chemficaflfly aflter the constfituent mfinerafls or may break an fintact mass finto smaflfler,
Earth materfiafls finteract cflosefly wfith hydroflogy and hydrogeoflogy. Water fis probabfly the most fimportant ffac- tor that promotes sflope finstabfiflfity. In manycases, waterfinfluencesthe strength parameters off geoflogficafl materfiafls, generaflfly reducfing strength when materfiafls become saturated. Pore water pressure changes the eff- ffectfive stress state off a sflope,typficaflfly reducfing resfistance to shear fforces, and promotfingfinstabfiflfity. Theflack off understandfing off hydroflogficafl con- dfitfions fis a ffrequent cause off ffafiflure fin managed sflopes; the 1966 Aber- ffan dfisaster fin South Wafles ffor ex- ampfle (Bfishop et afl., 1969), fin whfich more than 140 peopfle were kfiflfled by a flandsflfide ffrom a mfine waste tfip, was prfimarfifly the resuflt offthe con- structfion offthe tfip on a sprfing and watercourse, whfich promoted condfi- tfions off ffuflfl saturatfion affter perfiods off heavy rafinffaflfl. However, water can aflso have more compflex reflatfionshfips wfithfinstabfiflfity. For exampfle,fin some materfiafls partfiaflfly saturated condfi- tfions can provfide addfitfionafl strength through the generatfion off suctfion fforces, whfiflefin others saturated con- dfitfions can promote sofifl flfiqueffactfion affter ffafiflure, turnfing a sflow flandsflfide finto a hfighfly mobfifle and hfighfly de- structfive flow.
Land use can aflso be a key ffactor fin flandsflfide causatfion. Some types off vegetatfion can fimprove stabfiflfity by provfidfing addfitfionafl strength to the sofifl vfia root systems, and by reguflatfing thefinfifltratfion off water and drawfing down pore water pressures through transpfiratfion. In generafl, fforested
CHAPTER 3 UNDERSTANDING DISASTER RISK: HAZARD RELATED RISKISSUES - SECTIONII
sflopes are more stabflethanthoseflefft bare, andtherefis aflarge body off ev- fidence to support the argument that therefisfincreased mudflow actfivfity affter fires haveremoved vegetatfion (Cannon and Gartner, 2005; Shakesby and Doerr, 2006) andfincreasedfland- sflfidfing affter carefless floggfing (Jakob, 2000). In generafl,the removafl off veg- etatfion promotesfinstabfiflfity. Growfing new vegetatfionfis a dfiffficuflt(but effffec- tfive where successffufl) way to restore stabfiflfity. Defforestatfion hfighflfights the actfion off humans asthe finafl key ffac- tor. As peopfle modfiffy the flandscape, theflfikeflfihood offflandsflfidfing changes. In many cases, humans promote fin- stabfiflfity by cuttfing sflopes to steeper angfles, removfing vegetatfion, changfing hydroflogy and fincreasfing weatherfing rates.
Lands fl fide occu r rence fis re fla ted to causa fl ffac to rs , wh fich c rea te a p ropens fi ty
ffo r a s flope to ffa fi fl , and t r figge rs , name fly the spec fific ex te rna fl even t tha t finduces flands fl fide occu r rence a t a pa r t ficu fla r
t fime .
In most cases,thetfimfing off ffafiflurefis assocfiated wfith a trfigger event. Thfis fis not aflways true, however; there fis fincreasfing evfidence that sflopes can ffafifl through progressfive mechanfisms that finvoflve the weakenfing off sflope through tfime untfifl stabfiflfity fis com- promfised, but such events are rare, aflthough they can be destructfive. However, most flandsflfides are asso-
cfiated wfithacflearfly definedtrfigger. Heavy rafinffaflflfis a key ffactorfin gener- atfingflandsflfides, prfimarfiflythroughthe generatfion off pore water pressures and thus a reductfion fin the effffectfive normafl stress. For exampfle,the annu- afl gflobaflflandsflfide cycflefis domfinated by the effffects off rafinffaflfl assocfiated wfiththe South Asfian and East Asfian monsoons(Petfley, 2010). Thefimpact offthe South Asfian monsoon on the southern edge offthe Hfimaflayas, afl- flfied wfith the topography and materfi- afls offthe regfion, makesthfisthe gflob- afl hotspot ffor flandsflfide occurrence. However, the same correflatfion hoflds true everywhere.
The second key ffactor, and possfibfly the most fimportant fin terms offfloss offflfiffe,fisthefimpact off sefismfic events. Large earthquakesfin mountafin chafins can trfigger extraordfinary numbers off flandsflfides. Recent events fincflude the 2005 Kashmfir (Pakfistan) earthquake and the 2008 Sfichuan (Chfina) earth- quake, both off whfich kfiflfled more than 20 000 peopflefinflandsflfides. The Sfichuan earthquake aflone trfiggered more than 100 000 flandsflfides. At present,the nature offthefinteractfion between sefismfic waves and sflopes fis poorfly understood, and fforecastfing the fimpacts off a ffuture earthquake finterms offflandsflfidesfis ffraught wfith dfiffficuflty. However, the hfighflevefls off floss suggestthatthfis wfiflfl be a key area off researchfinthe ffuture.
Humans can aflso be a key trfigger off flandsflfides. The constructfion off hy- droeflectrfic statfions can be sfignfificant. The Three Gorges Damfin Chfina,the worfld’s flargest hydroeflectrfic project, fis expected to flead to the ufltfimate reflocatfion off 1.4 mfiflflfion peopfle ow- fing to the constructfion off a 650-km
flong reservofir andthefincreasedfland- sflfide rfisk; sfimfiflar probflems can be aflso ffound fin Europe but to a flesser extent. The Vajont rock sflfide (Itafly) resuflted fin the deaths off more than 2 000 peopfle fin 1963, when rock ffeflfl finto the reservofir fimpounded by the hfighest arch dam fin the worfld at the tfime. Humans trfigger flandsflfides through sflope cuttfing (especfiaflfly ffor road constructfion), defforestatfion, fir- rfigatfion, undercuttfing and changesfin hydroflogy and bflastfing, among many other actfivfitfies. Mfinfing actfivfitfies have a partficuflarfly flarge fimpact. In more devefloped countrfies, mfinfing fis there- ffore strfictfly reguflated; sadfly,finfless aff- fluentcountrfies,reguflatfionflagscon- sfiderabfly, andflosses are much hfigher. Ffinaflfly, fin actfive voflcanfic areas, fland- sflfides can be a major probflem. Some off the hfighest flevefls off floss have occurred as a resuflt offthe hfigh-mo- bfiflfity voflcanfic flandsflfide known as a flahar, and voflcanfic flank coflflapses, whfich can be tsunamfigenfic, may be the flargest terrestrfiafl flandsflfides pos- sfibfle. Some offthe deadflfiestflandsflfide events on record have occurred fin voflcanfic areas. Actfive voflcanfism pro- motes finstabfiflfity (the 1980 Mount St Heflens eruptfion started wfith a fland- sflfide that depressurfised the voflcano), and dome coflflapse fis common. Vofl- canfic deposfits reguflarfly mobfiflfisefinto hfigh-energy flows,and hydrothermafl actfivfity can cause materfiafl strength degradatfion over flarge areas. Major debrfis avaflanches, partfiaflfly submarfine, were trfiggered by the 2002 eruptfion off Stromboflfi voflcano(Itafly) andthey causedtsunamfis,fin atypficafl mufltfihaz- ard domfino effffect(Tfintfi et afl., 2006).
2013). Aflong wfiththe changesfin cflfi- mate and weather patterns, demogra- phy,fland use and other ffactors drfivfing the flandsflfide rfisk are changfing rapfid- fly (UN, 2015). Indeed, projectfions through the 21st century ffor Europe findficate that socfietafl changes may fleadto aflargerfincreasefinthefimpacts ffromflandsflfides and other naturafl haz-
finfluence off cflfimate change onthe spatfiafl and temporafl characterfistfics offflandsflfide rfisk wfiflfl be notficeabfle by the end offthe century. At a shorter tfimescafle off one to two decades, the rapfid changes fin anthropogenfic ffac- tors such as urbanfisatfion andfland use change drfivethe dynamfic rfisk pattern that we ffacetoday.
fin Europe and c fl fimate change
The ffast-paced changesfin socfiety, cflfi- mate change and the human fimpact on the envfironment have a major fimpact on the ffrequency and spatfiafl dfistrfibutfion off flandsflfides. Annuafl
Estfimate off changesfin the exposure off Europe’s popuflatfion toflandsflfidesfin the 21st century Source: SaffeLand (2013)
FIGURE 3.26
CHAPTER 3 UNDERSTANDING DISASTER RISK: HAZARD RELATED RISKISSUES - SECTIONII
Regfionafl cflfimate modefl (RCM) sfim- uflatfions ffrom the EU FP6 project ENSEMBLES (Van der Lfinden and Mfitcheflfl, 2009) predficted a consfistent flarge-scafle pattern off heavy precfipfita- tfion changes fin Europe. The sfimufla- tfions generaflfly showed anfincreasefin heavy precfipfitatfion over northern and centrafl Europe fin wfinter, aflthough some finconsfistencfies were ffound amongthe predfictfions ffrom dfifffferent modefls fin mountafinous regfions and at the ffoothfiflfls offthe mountafins. In summer, most modefls agree on an fincrease fin heavy precfipfitatfion over Scandfinavfia and reduced precfipfita- tfion fin southern Europe. The flarg- estfinconsfistencfies were ffoundfinthe transfitfion zone across centrafl Europe, whfich separates areas wfith posfitfive trends fin the north and areas wfith negatfive trends fin the south. Con- sfiderfing both the expected changes fin patterns off extreme precfipfitatfion events and changes fin other ffactors drfivfingtheflandsflfide rfisk,the EU FP7 project SaffeLand (www.saffefland.no) assessed the expected changes fin cflfi- mate-drfivenflandsflfide actfivfity(magnfi- tude, ffrequency)fin Europefinthe next 100 years.
I t mus t be emphas fized tha t any p rognos fis off the changes fin the soc fio- econom fic fimpac t off flands fl fides due to c fl fima t fic
change finvo flves a h figh fleve fl off unce r ta fin ty .
The SaffeLand study estfimated that flandsflfide hazardthreatens about 4 %
off European cfitfizens today. In addfi- tfiontothe peopfle dfirectflythreatened fin thefir homes, 8 000-20 000 km off roads and rafiflways are exposedto hfigh flandsflfide hazard, causfing addfitfionafl dfirectthreatstoflfiffe and economfic as- sets as weflfl as probflems ffor emergen- cy response and recovery operatfions (Jaedficke et afl., 2013). The SaffeLand prognosfis wasthat about 0.7% offthe totafl European popuflatfion wfiflfl ex- perfience an fincrease fin flandsflfide rfisk by the end offthe century, aflthough fin some parts off Europethe rfisk wfiflfl be reduced. The spatfiafl pattern off the expected changefinthe European popuflatfion exposed to flandsflfide rfisk fis depficted fin Ffigure 3.26. The mafin changesfinflandsflfide rfisk atthe Euro- pean scafle shownfinthe figure are due to the changes fin popuflatfion pattern caused by mfigratfion and urbanfisatfion. The SaffeLand project aflso made a detafifled study offthe changesfinfland- sflfide rfisk pattern at flocafl scafle ffor seflected sfites fin Europe ffor the perfi- od 1951-2050. For these studfies, the cflfimate sfimuflatfions were downscafled to sfimuflate flocaflfised heavy precfipfita- tfion events fin regfions where rafin-fin- duced flandsflfides occur on a reguflar basfis. The downscafled cflfimate mod- efls predficted an fincrease fin flandsflfide hazard at aflfl sfites. These resuflts dfiff- ffered ffrom the predfictfions provfided byflarger scafle cflfimate modefls at some flocatfions. These dfifffferences mfight be expflafined bythe refinementfinthe cflfimate modefl used, whfich, ffor exam- pfle, consfidered thefinfluence offflocafl topography on precfipfitatfion. Thfis demonstrated that flarge-scafle mod- efls are useffufl to evafluate the reflatfive spatfiafl varfiatfions off flandsflfide actfiv- fity, whfifle flocafl scafle modefls are nec- essary ffor urban pflanners and flocafl
authorfitfiesto estfimatethe ffuture rfisks assocfiated wfith flandsflfides and other hydro-meteoroflogficafl hazardsfinthefir communfitfies or regfions offfinterest. In addfitfion,theflarge uncertafintfiesfin popuflatfionandtraffficevoflutfionsce- narfios,fland use changes and poflfitficafl decfisfions regardfing urban deveflop- ment requfire that the key parameters drfivfing flandsflfide rfisk are accuratefly monfitored and that the prognosfis off flandsflfide rfiskfis contfinuousfly updated as newfinfformatfion becomes avafiflabfle and more accurate andrefined cflfimate change modefls are devefloped.
3 .5 .4
Lands fl fide zon fing : finventory , suscept fib fi fl fity and
hazard maps
The mappfing offflandsflfides underpfins dfisaster rfisk reductfion strategfies,finte- gratfing socfio-economfic fimpacts, and thereffore the chaflflenge fis to anaflyse thefir causes andtrfiggersfin our chang- fing envfironments. Owfing to the ex- traordfinary breadth offthe spectrum off flandsflfide phenomena, no sfingfle method exfists to fidentfiffy and map flandsflfides and to ascertafin flandsflfide susceptfibfiflfity and hazard.
In addfitfion to predfictfing ‘where’ a sflope ffafiflure wfiflfl occur,flandsflfide haz- ard fforecasts‘when’ or‘how ffrequent- fly’fit wfiflfl occur, and‘howflarge’fit wfiflfl be(Guzzettfi et afl., 2005).
The sfimpflest fform offflandsflfide map- pfing fis a flandsflfide finventory map, whfich showstheflocatfion and, where known, the date off occurrence and
nfiques, dependfing onthefir scope and the extent offthe study area. Smaflfl- scafle finventorfies (≤1:200 000) are compfifled mostfly ffrom data obtafined ffrom the flfiterature, through finqufirfies to pubflfic organfisatfions and prfivate consufltants, by searchfing chronficfles, journafls,technficafl and scfientfificre- ports, or by fintervfiewfing flandsflfide experts. Medfium-scafle flandsflfide fin- ventorfies (1:25 000 to 1:200 000) are most commonfly prepared through the systematficfinterpretatfion off aerfiafl photographs at scafles rangfing ffrom 1:60 000to 1:10 000, and byfintegrat- fingflocafl fiefldchecks wfith hfistorficafl finfformatfion. Large-scafle finventorfies (>1:25 000) are prepared, usuaflfly ffor flfimfited areas, usfing boththefinterpre- tatfion off aerfiafl photographs at scafles greaterthan 1:20 000, very hfigh-reso- flutfion sateflflfitefimages or dfigfitaflterrafin modefls, and extensfive fiefldfinvestfiga- tfions.
An archfive finventory shows finffor- matfion on flandsflfides obtafined ffrom the flfiterature or ffrom other archfive sources. Geomorphoflogficafl finvento- rfies can beffurther cflassfified as hfistor- ficafl, event, seasonafl or mufltfitemporafl finventorfies. A geomorphoflogficafl hfis- torficafl finventory shows the cumufla- tfive effffects off many flandsflfide events over a perfiod off tens, hundreds or thousands off years. In a hfistorficaflfin- ventory, the age offthe flandsflfides fis not dfistfingufished, or fis gfiven fin refla- tfiveterms(fi.e. recent, ofld or very ofld). An event finventory shows flandsflfides caused by a sfingfletrfigger, such as an earthquake, rafinffaflfl event or snowmeflt event, and the date offthe flandsflfide
seasonaflfinventorfies can be prepared. A seasonaflfinventory showsflandsfides trfiggered by sfingfle or mufltfipfle events durfing a sfingfle season, or a ffew sea- sons, whereas mufltfitemporaflfinvento- rfies showflandsflfidestrfiggered by mufl- tfipfle events overflonger perfiods(years to decades).
Lands fl fide suscep t fib fi fl fi ty fis the p robab fi fl fi ty off spa t fia fl
occu r rence off s flope ffa fi flu res , g fiven a se t off
geo-env fi ronmen ta fl cond fi t fions . Lands fl fide haza rd fis the p robab fi fl fi ty tha t a flands fl fide off a g fiven
magn fi tude w fi fl fl occu r fin a g fiven pe r fiod and fin a
g fiven a rea .
Conventfionafl methods to prepare flandsflfidefinventory maps refly prfimar- fifly onthe vfisuaflfinterpretatfion off ste- reoscopfic aerfiafl photography, afided by fiefldsurveys. Newandemergfing technfiques, based on sateflflfite, afir- borne and terrestrfiafl remote sensfing technoflogfies, promfiseto ffacfiflfitatethe productfion offflandsflfide maps, reduc- fing the tfime and resources requfired ffor thefir compfiflatfion and systemat- fic update. These can be grouped fin three mafin categorfies, fincfludfing the anaflysfis off surfface morphoflogy, chfieff- fly expflofitfing very-hfigh-resoflutfion dfigfitafl eflevatfion modefls captured ffor exampfle by LfiDAR (flfight detectfion
thetfic aperture radar (SAR) fimages, andthe use off newtooflsto ffacfiflfitate fiefld mappfing.
Quaflfitatfive and quantfitatfive methods ffor assfignfing flandsflfide susceptfibfifl- fitycan becflassfifiedfinto five groups (Guzzettfi et afl., 1999): (1) geomor- phoflogficafl mappfing, based on the abfiflfity off an expert finvestfigator to evafluate and map the actuafl and po- tentfiafl sflope finstabfiflfity condfitfions; (2) anaflysfis off flandsflfide finventorfies, whfich attempts to predfict the ffuture flandsflfide spatfiafl occurrence ffrom the known dfistrfibutfion off past and pres- ent flandsflfides (typficaflfly, thfis fis ob- tafined by preparfing flandsflfide densfity maps); (3) heurfistfic or findex-based approaches, fin whfich finvestfigators rank and wefight the known finstabfifl- fity ffactors based on thefir assumed or expected fimportance fin causfing flandsflfides; (4) process-based meth- odsthatrefly onsfimpflfified physficaflfly basedflandsflfide modeflflfing schemesto anaflysethe stabfiflfity/finstabfiflfity condfi- tfions usfing sfimpfle flfimfit equfiflfibrfium modefls, such as the ‘finfinfite sflope stabfiflfity’ modefl, or more compflex ap- proaches; (5) statfistficaflfly based mod- eflflfing contfingent on the anaflysfis off the ffunctfionafl reflatfionshfips between known or finfferred finstabfiflfity ffactors and the past and present dfistrfibu- tfion offflandsflfides. Regardfless offthe method used,fitfisfimportantthatthe susceptfibfiflfity zonatfions are vaflfidated usfingfindependentflandsflfidefinfforma- tfion, andthattheflevefl off uncertafinty assocfiated wfith the zonatfion fis gfiven (Rossfi et afl., 2010).
CHAPTER 3 UNDERSTANDING DISASTER RISK: HAZARD RELATED RISKISSUES - SECTIONII
Landsflfide hazardfis more dfiffficufltto obtafin than flandsflfide susceptfibfiflfity, sfincefit requfiresthe assessment offthe temporafl ffrequency offflandsflfides and the magnfitude offthe expected ffafifl- ures(Guzzettfi et afl., 2005). Thetem- porafl ffrequency(orthe recurrence) off flandsflfides, or off flandsflfide-trfiggerfing events, can be estabflfished ffrom ar- chfivefinventorfies and ffrom mufltfitem- poraflflandsflfide mapscoverfingsufffi- cfientfly flong perfiods. Furthermore, where a flandsflfide record fis avafiflabfle, an approprfiate modeflflfing fframework needs to be adopted (Wfitt et afl., 2010). Aflternatfivefly, ffor meteoroflogfi- caflfly trfiggered flandsflfides, one can fin- ffer the ffrequency offflandsflfide events ffrom the ffrequency offthe trfiggerfing ffactors, ffor exampfle the ffrequency (or the return perfiod) offfintense or proflonged rafinffaflfl perfiods. The un- certafintyfinherentfinthe predfictfion off trfiggers that may resuflt fin flandsflfides adds to uncertafinty finherent fin the predfictfion off occurrence off fland- sflfides.
To determfine the magnfitude off an expectedflandsflfide,finvestfigators most commonfly revert to determfinfing the statfistfics off flandsflfide sfize (area or voflume). Accurate finfformatfion on flandsflfide area can be obtafined ffrom hfigh-quaflfity geomorphoflogficaflfinven- torfies. Determfinfing the voflume off a suffficfientfly flarge number off fland- sflfidesfis more probflematfic, and usuafl- flyfinvestfigators refly on empfirficafl refla- tfionshfips flfinkfing flandsflfide voflume to flandsflfide areas (Guzzettfi et afl., 2009; Larsen et afl., 2010; Catanfi et afl., 2016). Ffinaflfly, when determfinfing flandsflfide hazard as the jofint probabfiflfity off flandsflfide sfize(a proxy ffor magnfitude), the expected temporafl occurrence off flandsflfides(ffrequency) andthe expect-
ed spatfiafl occurrence (flandsflfide sus- ceptfibfiflfity), great care must be taken to estabflfish fiff, or to what extent, the three probabfiflfitfies are findependent. In many areas, gfiventhe avafiflabflefin- fformatfion and the flocafl settfings, thfis may be dfiffficuflt to prove (Guzzettfi et afl., 2005). We expectthatthe quantfi- tatfive assessment offflandsflfide hazard wfiflfl remafin a major scfientfific chafl- flengefinthe next decade.
Suchfidentfificatfion off areassuscepbtfi- bfletoflandsflfide hazardfis essentfiafl ffor theflandsflfide rfisk assessment and pos- sfibflefimpflementatfion off effffectfive dfis- aster rfisk reductfion strategfies. These strategfies (Dafi et afl., 2002) fincflude fland-use pflannfing, deveflopment con- trofl fland, the appflficatfion off bufifldfing codes wfith dfifffferent engfineerfing soflu- tfions, acceptance, and monfitorfing and earfly warnfing systems. Land pflannfing controfl reduces expected eflements at rfisk. Engfineerfing soflutfionfisthe most dfirect and costfly strategy ffor reducfing efitherthe probabfiflfity offflandsflfidfing or the probabfiflfity off spatfiaflfimpact off a flandsflfide. One approachfis correctfion off the underflyfing unstabfle sflope to controfl finfitfiatfion offflandsflfides (such as stabfiflfisatfion off sflope, drafinage, retafinfing waflfls or pflantfing), and the other fis controflflfing offthe flandsflfide movement (such as barrfiers/waflfls to reduce or redfirect the movement when aflandsflfide does occur). The ac- ceptancestrategy definesacceptabfle rfisk crfiterfia(Feflfl, 1994;Feflfl and Hart- fford, 1997); and the monfitorfing and warnfing system strategy reduces ex- pected eflements at rfisk by evacuatfion fin advance off ffafiflure.
3 .5 .5 Lands fl fide mon fitor fing and
ear fly warn fing
Thesesystemsrequfirea fineassess- ment off the socfioeconomfic fimpact off flandsflfides, whfich must be based on accurateflandsflfide mappfing, as weflfl as an understandfing offthefir causes. EWSs fforflandsflfides are based onthe reflfiabfle contfinuafl monfitorfing off refl- evant findficators (e.g. dfispflacements, rafinffaflfl, groundwater flevefl) that are assumed to be precursors to trfig- gerfing flandsflfides or reactfivatfions. When vaflues ffor these findficators ex- ceed predefined threshoflds, aflarms are transmfitted dfirectfly to a chafin off peopflefin charge off decfidfingtheflev- efl off warnfing and/or emergencythat must be transmfitted to the reflevant stakehoflders,ffoflflowfinga predefined process (Ffigure 3.27). In some cases, warnfings can aflso be automatficaflfly transmfitted. Usuaflfly, one to five aflert flevefls are used (Bflfikra, 2008; Intrfierfi et afl., 2013):the hfighestflevefl mayflead to emergency warnfingstothe popufla- tfion, evacuatfions orthe use off sfirens and floudspeaker messages fin severafl flanguagesto fforce peopfleto moveto a saffer pflace, asfinthe case offtsuna- mfisfinduced byflandsflfides.
An EWS needsto be set up wfith spe- cfific requfirements. Ffirst,the potentfiafl fimpacts must be defined based ona rfisk anaflysfisfinfformed by hazard map- pfing, fincfludfing the fimpact off gflobafl changes (Coromfinas et afl., 2014). In addfitfion, the causes and trfiggers off dfisasters must bethoroughfly anaflysed andthe deveflopment offflocafl copfing
gerfing condfitfions and/orto detectfing an ongofing event(Sättefle et afl., 2016). For exampfle, monfitorfing systems off debrfis-flow or shaflflowflandsflfide EWSs are usuaflfly based onthreshoflds off rafinffaflfl amount over a perfiod off tfime. These threshoflds are based on rafinffaflfl fintensfity-duratfion, cumuflat- ed event rafinffaflfl-duratfion (Guzzettfi et afl., 2008), or antecedent precfipfita- tfion(fincfludfing snow depth) measures and sofifl mofisture (Baum and Godt, 2010; Jakob et afl., 2012). An extended monfitorfing offthosefindficators usuafl- fly makes fit possfibfle, thereffore, to set regfionafl aflarms. Landsflfide types aflso constrafin the maxfimum flead tfime or tfime off reactfion affterthe aflarmtrans-
sensors that detect the debrfis-flow movements (Marchfi et afl., 2002) and automatficaflfly send a warnfing message to shortenthe reactfiontfime as much as possfibfle.
For sfite-specfific systems, dfispflace- ments measured by dfifffferent sensors and pore water pressure and/or pre- cfipfitatfion are usuaflfly used (Mfichoud et afl., 2013). Varfious sensors can be setto monfitor dfispflacements,fincflud- fing extensometers(cabfle orflaser) and crackmetersthat measurethe dfistanc- es between two pofints, and totafl sta- tfionsthat are aflso usedto provfide dfis- tances and 3D posfitfions usfingtargets posfitfioned on sfite. Moreover, GPSs the begfinnfing offthe 21st century be-
cause offthe progress made fin eflec- tronfics, communficatfion and com- puter programs ffor monfitorfing and fimagfing. In addfitfion, the finnovatfions fin sateflflfite technoflogfies and ground remote sensfing have greatfly fim- proved the capacfity off remote fimag- fing measurements versusfin sfitu pofint measurements (Toffanfi et afl., 2013). Impflementfing an EWS depends on the context, namefly (1) the type off flandsflfide(Hungr et afl., 2014),(2)the dfisaster scenarfios consfidered, (3) the degree off awareness offthe stakehofld- ers,fincfludfing popuflatfions, and(4)the aflflocated resources (e.g. budgetary, human).
(A)Iflflustratfion off the components off a modern EWS that does not show the energy sources and the two or threeflevefls off redundancy. (B) Fflow chart off the actfivfitfies off thefimpflementatfion and operatfion off an EWS (modfifffied ffromIntrfierfi et afl., 2012). The bflue boxfin (b)findficates the actfionflfinked to the monfitorfing system. Source: courtesy off authors
FIGURE 3.27
Weather statfion Crackmeter
GPS base statfion (Ffixed)
A
GB-InSAR Extenso
meter (cabfl
e dfispflacement)
B Geoflogficafl knowfl-
edgerfiskscenarfios Instaflflatfion offthe monfitorfingsystem
Data acqufisfitfion
Are data wfithfin specfifffied bounds?
Are anomaflfies due to finstrumentafl error?
Issue an aflarm or pre-aflarm Emergency pflans
Troubfleshootfing and mafintenance
YES YES
NO
NO MONITORING
RISK KNOWLEDGE
ANALYSIS AND FORECASTING
RESPONSE WARNING
Constant update Feedback
CHAPTER 3 UNDERSTANDING DISASTER RISK: HAZARD RELATED RISKISSUES - SECTIONII
are nowadays wfidefly used, whfich can gfive the reafl 3D posfitfion off a pofint (Gfiflfi et afl., 2000). Aflflthe abovetech- nfiques usuaflfly provfide data onfly at specfific pofintflocatfions;thus,severafl offthem must offten be set upfin a net- work to monfitor areafl defformatfions. Incflfinometers gfive defformatfions at depth aflong borehofles, provfidfing es- sentfiafl data on the changes fin depth offflandsflfide behavfiour(Bflfikra, 2008). For the flast ffew years, ground-based finterfferometrfic radar (GB-InSAR) has been used ffor the most crfitficafl flandsflfides(Casagflfi et afl., 2010; Bflfikra, 2012; Rouyet et afl., 2016). It provfides a map offthe dfistance changes, ffrom the GB-InSAR to the flandsflfide sur- fface, at a mfiflflfimetre scafle and wfith a tfime resoflutfion off a ffew mfinutes. Sateflflfite InSAR fimages are aflso used to monfitor flong-term dfispflacement trends, wfith resuflts befing strongfly dependent on the type offtreatment. In optfimafl cases, the tfime resoflutfion fis about 6 days, wfith mfiflflfimetre pre- cfisfion and metre spatfiafl resoflutfion (Berger et afl., 2012). Ffinaflfly, as fland- sflfides reactto waterfinfifltratfion, many finstruments are dedficatedto monfitor water: rafin gauges, pfiezometers,ther- mometers, barometers, mofisture con- tent sensors and other meteoroflogficafl data. Pore water pressure changes monfitored wfith pfiezometers usuaflfly have a good correflatfion wfith sflope movements(Mfichoud et afl., 2013). Behfind the fimpflementatfion off the monfitorfing part off EWSs fis the un- derstandfing off the flandsflfide mech- anfisms,thatfis,thefidentfificatfion off the mafin parameters controflflfing the movements offthe flandsflfide (Intrfierfi et afl., 2012 and 2013). For thfis pur- pose, the desfign off a flandsflfide con- ceptuafl modefl(LCM)fis ffundamentafl,
sfincefit wfiflfl gufidethetype andtheflo- catfion offthe sensorstofinstaflfl, andfit fis requfiredto fforecastflandsflfide ffafiflure scenarfios. The updatfing off an LCM must be contfinuafl durfing the whofle flfiffe off an EWS. In addfitfion,flandsflfide ffafiflures may trfigger other hazardous events fin a cascade effffect, such as tsunamfis or dam breaks,that haveto be consfidered fin the EWS. The rea- sons why an EWSfisfimpflemented are efitherthefidentfificatfion offan unac- ceptabfle rfiskflevefl or anfincreasefin, or abnormafl,flandsflfide actfivfity. Aflthough the LMC fimpflementatfion process provfides reasonsto fix approprfiate sensors that wfiflfl monfitor the most sfignfificant ffafiflurefinfitfiatfionfindficators, there are usuaflfly many practficafl con- strafints, such as topography, access, vfisfibfiflfity and avafiflabfle resources.
Lands fl fide mon fi to r fing and EWSs a re too fls to ffo recas t the po ten t fia fl occu r rence off d fisas te rs ,
thus con t r fibu t fing to the fimp flemen ta t fion off effec t five d fisas te r r fisk-
reduc t fion s t ra teg fies .
Ideaflfly, the first data ffrom a monfitor- fing system are used to caflfibrate and fix aflarm threshoflds usuaflfly based on dfispflacement veflocfitfies or acceflera- tfions, or pore water pressure or pre- cfipfitatfions(Cfloutfier et afl., 2015). Thfis approach can be supported by ffafiflure fforecast modefls, such asthe Fukuzo- no method, or by more compflex mod- efls(Crosta and Agflfiardfi, 2003; Feder- fico et afl., 2012). The aflarmthreshoflds
wfiflfl be used to trfigger chafins off ac- tfions that wfiflfl finvoflve dfifffferent flev- efls off peopfle dependfing on the aflert flevefl, ffromtechnficfians and expertsto offficersand poflfitficfians who wfiflfl be finvoflvedfinthe assessment offthe ab- normafl sfituatfions and who wfiflfl have to make decfisfions(Froese and More- no, 2014). Thfis starts ffromthefinfitfiafl check offthe sfituatfion andthe coher- ence offthe movement detectfion off the sensors(to avofid ffaflse aflarm), and fit can end wfith an evacuatfion decfisfion. It requfiresthatthe monfitorfing system fis reflfiabfle and fis thereffore redundant fin terms off sensors, communficatfion andthe stakehofldersfinvoflved. Pre-de- fined crfisfis unfits mustffoflflow decfisfion trees to propagate or stop the warn- fing at eachflevefl. Thfis aflso necessfitates the requfirement to verfiffy constantfly thatthe observedflandsflfide behavfiour fis stfiflfl ffoflflowfingthe expected course, whfich aflsofimpflfiesthatthethreshofld and aflarmflevefls can be reassessed by the crfisfis unfits.
The most fimportant actfions that can be prompted by EWS hfigh-aflertflevefls are evacuatfions and a rapfid set-up off protectfion measures. Theyfimpflythat aflfl stakehoflders,fincfludfingthe reflevant popuflatfion, must be preparedthrough educatfion and trafinfing to fimpflement the approprfiate response.
In addfitfion,the methods usedto emfit and communficate the emergency sfit- uatfion must be adapted to the flocafl popuflatfion cuflture. It must be stressed that aflfl stages offfimpflementatfion or operatfion must fincflude ffeedback to the other stages. Frequent ffeedback and updates are a key pofint. They must aflso fincflude the reapprafisafl off thefindfirecteffffects(cascade). A finafl probflem reflatesto communficatfionto
and actfions wfithfin the nofise off our
‘connected worfld’. It appears that onfly 38 % offthe EWSs have more than one communficatfion vectortofin- fform the popuflatfion (Mfichoud et afl., 2013).
3 .5 .6
Conc flus fions and key messages
Partnersh fip
Understandfing flandsflfide rfisk requfires a mufltfihazard approach, based on networkfing and partnershfip between dfifffferent scfientfific dfiscfipflfines, wfith transdfiscfipflfinary research that afims to fidentfiffy those socfioeconomfic and finstfitutfionafl eflements that requfire at- tentfionfinflandsflfide DRM.
Know fledge
Knowfledge off flandsflfide rfisk fis a mufltfidfiscfipflfinary task that requfires an understandfing off processes and mechanfisms, spatfiafl and tfime predfic- tfion, vuflnerabfiflfity assessment, mon- fitorfing and modeflflfing offthe effffects reflated to envfironmentafl and cflfimate change.
Innovat fion
The effffectfiveness off flandsflfide rfisk mfitfigatfion measures crfitficaflfly depends onscfientfificfinnovatfionandtechno- flogficafl deveflopment ffor rapfid map- pfing, monfitorfing and earfly warnfing.
