The Natural Hazard Loss Datascape

Transcription

1 The Natural Hazard Loss Datascape Susan L. Cutter IRDR ICoE Vulnerability and Resilience Metrics University of South Carolina Daniele Ehrlich European Union Joint Research Center ropa.eu IRDR Science Committee and DATA Project Co-Chairs

2 What are disaster loss data? Quantifiable measures that account for the human impact of disasters. Quantifiable counts of direct human losses people who died, were physically injured, declared missing from the event. number of people needing immediate assistance for sheltering (homeless) and those who were temporarily forced to leave their homes (evacuated) or who were forced to leave for longer time periods (displaced). Direct, tangible monetary impact - damages to buildings, infrastructure, agriculture, and the natural environment Indirect, economic losses - damages to services, lifelines, usually measured in different societal sectors, banking, insurance, industrial Monetary losses paid out by the insurance industry.

3 Why do we care about disaster loss data? 1. Provide a national inventory of losses for a baseline to determine our risks and vulnerabilities 2. Track pattern of losses to see where losses are greatest and why 3. Provide scientific basis for allocation of resources to higher needs areas 4. Provide new tools for better emergency preparedness and planning for disasters 5. Help in developing policies and programs for loss reduction from natural hazards 6. Track efficiency of disaster relief funding and mitigation money 7. Assess the effectiveness of hazard reduction policies and programs

4 Loss data inform policy making in different ways JRC, 2013, as reported in GAR 2015

5 What is the problem? Extensive loss data are collected and stored by different organizations, but the thoroughness and accuracy of the data varies from country to country, between existing global databases, and among local entities. Government agencies, private companies, and other organizations may collect and manage data related to their own areas of interest using their own standards and procedures, without significant collaboration with other groups. This result in gaps in the data, inconsistent overlaps, and biases that ultimately affect the quality of research conducted and policies made based on the data.

6 Existing Issues Lack of comparability across global databases and between global and nations, regional, or local Data gaps time period covered, region covered Loss estimates highly variable especially for low impact, high frequency events Inconsistent loss indicators Inconsistent attribution of losses to event origin or classification, and geographic location

7 How are the losses attributed by individual peril? Where are the losses attributed?

8 IRDR Data Project Co-Chairs: Susan L. Cutter, University South Carolina, USA Daniele Ehrlich, EU Joint Research Center Ispra, Italy Sisi Zlatanova, Delft University Delft, The Netherlands Project Members: Regina Below, CRED (University of Louvain) Lucia Bevere, Sigma (SwissRe) Bob Chen, CIESIN (Columbia University) Daniele Ehrlich, EU Joint Research Centre Jan Eichner, NatCatService (MunichRe) Julio Serje, DesInventar (UNISDR) Ramesh Sharma, UNDP Adam Smith, NCDC/NOAA Wei-Sen Li, National S&T Center for Disaster Reduction Maria Patek, Austrian Government Sisi Zlatanova, Delft University Frederic Zanetta, IFRC Ricardo Zapata Marti, Eclac Cepal Francis Ghesquiere, World Bank

9 DATA Project Goals Identify quality of existing data and data needs for improving integrated disaster risk management Bring together loss data stakeholders to identify common issues and develop synergies Develop (standards)guidelines/protocols to minimize data uncertainty Define losses and create transparent methodologies for assessing them Advocate for loss data at sub-national geographies Educate users on database biases and data interpretation

10 Harmonizing Peril Classifications Issues Databases classify events differently (e.g. Landslides could be geophysical (due to seismic shaking)or hydrological (due to excessive rainfall causing slope failure) events depending on the initiating process Perils may have different meanings in different languages irdr-peril-classification-and-hazard-glossary/ Value: A consistent peril classification allows data users to compare losses from landslides for example in database A with those recorded in database B, knowing that differences reflect variations in loss estimation methods, not different definitions of landslides or how they were classified.

11 Framework Designed to apply to multiple types of databases global, national, subnational Includes only the most common perils that cause measureable damage Has three levels representing classification from the most general (Family) to the Main Event, to the most specific (Peril) Family Main Event Peril

12 Family Geophysical Hydrological Meteorological Climatological Biological Extraterrestrial Structure Geophysical a hazard originating from solid earth Hydrological a hazard caused by the occurrence, movement, and distribution of surface and subsurface freshwater and saltwater Meteorological a hazard caused by short-lived, microto meso-scale extreme weather and atmospheric conditions that last from minutes to days Climatological a hazard caused by long-lived, meso- to macro-scale atmospheric processes ranging from intraseasonal to multi-decadal climate variability Biological a hazard caused by the exposure of living organisms and their toxic substances (e.g. venom, mold) or vector-born diseases that they may carry Extraterrestrial a hazard caused by asteroids, comets, and meteoroids, as they pass near or strike Earth; and changes in interplanetary conditions affecting Earth s magnetosphere, ionosphere, and thermosphere

13 Structure Main Event Family Geophysical Hydrological Meteorological Climatological Biological Extraterrestrial Earthquake Mass Movement Volcanic Activity Drought Glacial Lake Outburst Wildfire Flood Landslide Wave Action Convective Storms Extratropical Storms Extreme Temperature Fog Tropical Storms Animal Incident Epidemics Insect Infestation Airburst Impact Space Weather

14 Peril Ground Movement Fire following EQ Liquefaction Landslide following EQ Ash Fall Lahar Lava Flow Pyroclastic Flow Tsunami Coastal Flood Flash Flood Ice Jam Flood Riverine Flood Avalanche: Snow, Debris Debris/Mud Flow/Rockfall Expansive Soil Sinkhole Coastal Erosion Rogue Wave Seiche Structure Derecho Hail Lightning Rain Snow/Ice Tornado Wind Sandstorm/Dust storm Winter Storm/Blizzard Cold Wave Frost/Freeze Heat Wave Storm Surge Dessication/Subsidence Forest Fire Land fire: Brush, Bush, Pasture Viral Bacterial Parasitic Fungal Prion Shockwave Energetic Particles Geomagnetic Storm Radio Disturbance

15 Can aggregate or disaggregate, and link specific perils to more than one Main Event The associations between classes is not a one-to-one correspondence DATA Project, DATA Project,

16 Definitions

17 Significance Permits comparability across databases despite different goals and objects of individual databases Enables top-down or bottom-up structure which meets needs of global to local databases

18 Implementation: Munich Re

19 Implementation: EM-DAT

20 Implementation: DESINVENTAR net/definitions.html

21 Implementation: SHELDUS SHELDUS (18 Hazards) Earthquake Tsunami/Seiche Volcano Avalanche Coastal Flood Landslide Fog Heat Hurricane/Tropical Storm Lightning Severe Thunderstorm Tornado Wind Drought Winter Weather Wildfire

22 Planned Implementation: European Union Countries to IRDR Perils

23 Human and economic impact indicators For the 3 global databases (EM-Dat, MunichRe, Swiss Re) and the more than 55 national disaster loss databases: No consistent set of impact measures across all databases Inconsistent definitions of those loss accounting variables Problem: Not really know the true impact of events because the indicators vary widely depending on which database is used Policies based on inconsistent or incomplete data with much uncertainty Example: Mortality Deaths only? Deaths plus missing persons? Immediate deaths or delayed deaths? Direct or indirect deaths?

24 Guidance for impact indicators Provides minimum set of desired indicators across all loss databases Glossary of definitions to maintain consistency in counting

25 Framework for impact indicators Primary those should be included in all loss databases Secondary additional qualifiers of the primary variables, or secondary information useful in response and recovery Tertiary more detailed information on each of the primary and secondary indicators

26 Example: Exposed Number of people who permanently or temporarily reside in the hazard area before or during the event. Critical indicator as it determines what percentage of the total population is at risk; provides the denominator on magnitude estimations of human impacts Example: an event causing 1,000 deaths Number of deaths per exposed population = 1,000/4,000 or 25% of the exposed population died Number of deaths per capita= 1,000/10,000 or 10% of the total population Area A Area B Exposed Population Area A population =10,000 Area B population = 4,000

27 Coastal storm surge in New York City Number of deaths, injuries, etc. /exposed population Vs Number of deaths, injuries, etc./capita

28 Consultations on indicators

29 Organizations involved at present Global CRED/EMDAT Munich-Re Swiss-Re Regional USo.Carolina UNISDR, UNDP (Sheldus) (Desinventar ) Other CIESIN ECLAC (DALA) World Bank (PDNA) NCDC/NOAA European Union Taiwan Austria World POLICIES Indicators LOSS DATABASES Store and manage METHODOLOGIES Gather Loss data

30 Loss data reporting Post 2015 Framework for DRR Global statistics Continent A Accounting Comparability Continent B Local, Regional, National Country Country Country Country

31 Next steps Analysis of Post 2015 Framework DRR indicators Promote IRDR indicators and perils at national and sub-national scales Help implement classification by provide expertise and assistance Assist nations in development loss accounting databases IRDR DATA group is committed to the harmonization and improvement of local to global loss accounting and using the best science available to do so.

32 Conclusions Learn what indicators will be required from post 2015 DRR framework and Sustainable Development Goals Refine our loss data guidelines to be able to generate indicators and information required for DRR and SDG Need to all work together to make sure that we produce consistent attribution of losses, indicators consistent in time and space, indicators that are comparable, and that the data we collect can be aggregated to regional and continental scales.

33 For more information projects/data/ Susan L. Cutter Daniele Ehrlich europa.eu