Modelling (mountain) flood risk and managing its uncertainties

Transcription

1 DICA seminar Modelling (mountain) flood risk and managing its uncertainties 14 March 2016

2 Abstract 2 Modelling (mountain) flood risk and managing its uncertainties Hydraulic engineers fight against flood risk. However, most of the effort within this battle has been devoted to the hazard component of the process, that is, the forcing action (see DICA seminar by prof. Menduni, 2015). Parallel to this, most of the scientific and technological progress has focussed on the forcing phenomena. However, no "risk" can be assessed if the expected damages are not evaluated; the latter are by far the least known part of the risk chain: data are scarce and low-quality, damage models are, consequently, poor. By means of a case study (Orvieto flood, 2013) this seminar is aimed at discussing possible approaches to the assessment of the complete risk chain and warning about the fragilities of available models.

3 Risk assessment: what for? 3 land use planning (say, construction permission) design mitigation works (river training) emergency management (contingency plans) insurance costs cost - benefit analysis

4 Structure of the seminar 4 Introduction presentation of case study the flood phenomenological chain the structure of risk flood risk Flood risk assessment hazard exposure + vulnerability damages Case study hazard exposure + vulnerability damages discussion flood risk assessment

5 Case study: presentation 5 Umbria 2012, town of Orvieto

6 Case study: presentation Rivers: Paglia and Chiani (confluence) Courtesy Francesco of Claudio Ballio Margottini, ISPRA

7 Introduction: the flood chain 7 atmosferic hydrological propagation along river flood damages

8 Introduction: the flood chain 8??? atmosferic rain: i(t) hydrological propagation along river flood damages

9 Introduction: the flood chain 9 atmosferic rain: i(t) discharge: Q(t) hydrological propagation along river flood damages

10 Introduction: the flood chain 10 atmosferic discharge: Q(t) water depth: h(t) hydrological propagation along river flood damages

11 Introduction: the flood chain 11 atmosferic hydrological water depth: h(t) propagation along river flood flooded area: A(t) damages

12 Introduction: the flood chain 12 atmosferic residential buildings hydrological propagation along river flood flooded area: A(t) damages

13 Introduction: the flood chain 13 commercial activities atmosferic hydrological agriculture propagation along river flood flooded area: A(t) damages

14 Introduction: the flood chain 14 atmosferic hydrological propagation along river flood infrastructures flooded area: A(t) damages

15 Introduction: the structure of risk 15 RISK = DAMAGE "+" PROBABILITY Damage = f ( x 1, x 2, x 3,, x N ) Hazard Exposure Vulnerability Probability variables x i may be considered / treated either as stochastic or deterministic How should we address this?

16 Introduction: the structure of risk RISK = DAMAGE "+" PROBABILITY Damage = f ( x 1, x 2, x 3,, x N ) For example: x 1 = rain quantity on a given day x 2 = river geometry 16 x 3 = residential value in a given area x i = x n = cars parked in a given area pdf pdf pdf pdf x 1 x 2 x 3 x n pdf D=0 Damage

17 Introduction: the structure of risk RISK = DAMAGE "+" PROBABILITY Damage = f ( x 1, x 2, x 3,, x N ) 17 Another example (bridge collapse): x 1 = flow depth and velocity x 2 = bridge geometry x 3 = resistance of materials x i = x n = traffic load pdf pdf pdf pdf x 1 x 2 x 3 x n pdf D=0 Damage

18 Introduction: the structure of risk 18 RISK = DAMAGE "+" PROBABILITY pdf Damage = f(threshold) pdf D=0 Damage Design: we basically consider only tails (to be avoided safety factors) Risk Assessment: we need the shape of the whole pdf D=0 Damage Safety factors are not of much help within risk assessment

19 Introduction: flood risk 19 Many s, models, parameters. Many steps, many uncertainties atmosferic Hazard Damage = f ( x 1, x 2, x 3,, x N ) Exposure Vulnerability causes effects hydrological propagation along river flood damages

20 Flood risk assessment: hazard 20 atmosferic Damage = f ( x 1, x 2, x 3,, x N ) river geometry town geometry vegetation bank and levee failures sediment transport origin / characteristics of uncertainties: natural system, complexity unsteadyness discharge: Q(t) hydrological propagation (1D) flood (2D) flooded area water depth damages water velocity contaminants

21 Flood risk assessment: hazard 21 atmosferic discharge: Q(t) hydrological propagation (1D) We consider our models to be relatively robust from the conceptual and mathematical perspectives. BUT: Available maps mostly come from 1D stationary calculations All inputs but discharge are treated as deterministic variables (often considered as null!) Uncertainties more or less subjectively embedded in safety factors. flood (2D) flooded area water depth damages water velocity contaminants

22 Case study: hazard Courtesy Francesco of Claudio Ballio Margottini, ISPRA

23 Case study: hazard 23 event discharge as for an estimated return period T=100 years

24 Case study: hazard Study mandated by the Orvieto Municipality to the professional consultant BETA Studio from Padova. Source: BETA Studio srl, S. Nicolò, PD, Analisi idraulica e idrologica della piena del fiume Paglia nelle zone di Orvieto Scalo e Ciconia e scenario di esondabilità: implicazioni per il sistema delle infrastrutture Francesco e Ballio dei trasporti, Ottobre 2013.

25 Reasons for discrepancies: singularities (1D vs.2d modelling) neglected / simplified parameters and (sediment transport) but: ex-post! Study mandated by the Orvieto Municipality to the professional consultant BETA Studio from Padova. Source: BETA Studio srl, S. Nicolò, PD, Analisi idraulica e idrologica della piena del fiume Paglia nelle zone di Orvieto Scalo e Ciconia e scenario di esondabilità: implicazioni per il sistema delle infrastrutture Francesco e Ballio dei trasporti, Ottobre 2013.

26 Flood risk assessment: exposure + vulnerability damages 26 atmosferic exposure: what is inside the affected area? vulnerability: how much of it will be damaged? hydrological Damage = f ( x 1, x 2, x 3,, x N ) Exposure Vulnerability physical damage value effects propagation (1D) flood (2D) damages

27 Flood risk assessment: exposure + vulnerability damages 27 flood (2D) flooded area water depth water velocity contaminants damage models residential buildings industrial & commercial activities agriculture sector infrastructures system parameters public items emergency costs people no more chain, cluster hazard (in)dependent? environmental and cultural herit.

28 Flooded area Calculation of DTM Flood risk assessment: exposure + vulnerability innundation depth (h) flooded area water depth water velocity contaminants system parameters flood (2D) PoliMi procedure Flood-IMPAT Integrated Meso-scale Procedure to Assess Territorial flood risk Flood profile CORINNE map Census block map Cultural/Env. heritage map Hazardous installations map Lifelines and strategic /pubblic facilities map Exposure map (spatial damage models distribution of high residential buildings Market value Building areas Net capital stock Production units EXPOSURE/VULNERABILITY Conversion table Overlapping/Intersection VULNERABILITY MAP (spatial distribution of land use classes) Monetary value estimation - settlements Monetary value estimationindustry agriculture Monetary value sector Net capital stock Monetary value Production estimation- monetary value) industrial units & commercial activities industry estimationwild areas Monetary value estimationhigh vulnerable infrastructures areas Population public items density vulnerable elements) Estimate of population at risk Monetary value emergency estimation- costs roads/railways HAZARD EXPOSURE MAP (spatial distribution of RISK FLOOD HAZARD MAP (spatial distribution of h) Monetary value (struct. & cont.) Monetary value (struct. & cont.) Monetary value (struct. & cont.) Damage curvessettlements Damage curvesindustry Damage curvesagriculture Direct tangible damage FLOOD HAZARD MAP (spatial distribution of h) Damage (struct. & cont.) Damage (struct. & cont.) Damage (struct. & cont.) FLOOD HAZARD MAP (spatial distribution of h) People at risk 28 Damage model (v=1) Indirect tangible damage Intangible damage FLOOD HAZARD MAP (spatial distribution of h).. Affected people people Monetary value estimationother lifelines Monetary value estimationpubblic/strategic facilities Monetary value estimationcultural/environmental and cultural H. goods Monetary value estimationhazardous installation Significance Combination TOTAL DAMAGE map.

29 Case study: damages 29 agriculture highway houses cars (parking) industries railway (station)

30 Flood risk assessment: damages - buildings 30 Location and dimension of buildings from Cadastre Value of building structure from Real Estate and Property Price Database (OMI) Value of contents: 7.5% 25.5% of the reference market value of building structures (low-cost buildings luxury buildings)

31 Flood risk assessment: damages - buildings 31 Indicator Description a) Location Hazard level depends on building location b) Type of use Contents value (and, consequently, damage) depends on buildings use (e.g. residential, commercial, public service, etc.) c) Level of maintenance Well maintained buildings better face the impacts of floods than crumbling d) Age Age is usually linked to the level of maintenance e) Materials Some materials (like concrete and masonry) are more resistant to the flood impacts than others (e.g. timber, plasterboard, etc.) f) Number of storeys The presence of more than one storey allows people to move contents to upper floors g) Presence of basement h) Number of openings at street level i) Height from street level l) Presence of vulnerable equipments Basements can be flooded also in case of minor events (small water depth) Openings at street level make water to easily enter the building If ground floor is higher than street level water is hindered to enter the building If vulnerable equipments are present in more flood prone floors than damage can be higher

32 Case study: damages - buildings 32 1,2 1,0 Standard Method (Ministerie van Verkeer en Waterstaat, Holland) DAMAGE [-] 0,8 0,6 0,4 0,2 low rise intermediate high rise M Observed damage 0.56 Estimated_SM 0.84 Estimated_FLEMO_ps , WATER DEPTH [m] 0,90 FLEMO_ps (Thieken et al., 2008) DAMAGE [-] 0,75 0,60 0,45 one family detached multifamily umbria data 0,30 0,15 0,00 0 0,5 1 1,5 2 2,5 3 WATER DEPTH [m]

33 Case study: damages - buildings 33 DAMAGE [-] 1,2 1,0 0,8 0,6 0,4 0,2 0,0 Standard Method (Ministerie van Verkeer en Waterstaat, Holland) low rise intermediate high rise WATER DEPTH [m] to be discussed: M "richness" Observed of the model: damage how many parameters 0.56 it considers Estimated_SM 0.84 scale: object, areal (averaged) Estimated_FLEMO_ps 0.42 uncertainties (referred to scale) 0,90 FLEMO_ps (Thieken et al., 2008) damage % h Unresolved features are embedded in the pdf. Calibration needed! DAMAGE [-] 0,75 0,60 0,45 0,30 0,15 0,00 one family detached multifamily umbria data 0 0,5 1 1,5 2 2,5 3 WATER DEPTH [m]

34 Flood risk assessment: damages industrial activities 34 Rhein atlas (International commission for the protection of the Rhine) structure content

35 Case study: damages industrial activities 35 The monetary value of industry land use class is calculated from the knowledge of (i) the net capital stock per production unit and (ii) the number of production units per census block, both supplied by ISTAT (subdivided by structures and contents). The net capital stock is differently calculated (at national level) for the several classes of industrial activities as defined by ATECO. Estimated damages Observed damages Damage to Businesses [M ] 26,5 28,6

36 Case study: damages, overall results 36 2% Umbria 2012, whole region 42% 11% 19% 23% 2% First aid Cultural Heritage Infrastructure Industry Agricolture Residential

37 Flood risk assessment: what for? (discussion) 37 There is no universal / general structure for (flood) risk assessment. Models, i.e.: their structure their scale their comprehesiveness their statistic vs. deterministic balance their accuracy (calibration / validation on data!) their can differ significantly from each other depending on the purpouse of the assessment. (what for?) back to the first slide

38 Risk assessment: what for? land use planning (say, construction permission) design mitigation works (river training) emergency management (contingency plans) insurance costs cost - benefit analysis Conclusions complexity of the problem lack of knowledge on the damage side role of uncertainty / stochastic variables the risk of (over)simplification no ONE solution (what for?)