Assessing Climate-Related Risks and Vulnerabilities

1. Assessing Climate-Related Risks and Vulnerabilities Nick Brooks nb@garama.co.uk • www.garama.co.uk

2. Presentation content • Characteristics of a risk/vulnerability assessment • Risk, vulnerability & resilience frameworks • Tools & approaches for risk/vulnerability assessment • Common elements across approaches • Data & information for risk/vulnerability assessment • Assessing current climate hazards • Assessing future climate hazards • Sources of climate information • Assessing the societal components of risk/vulnerability • Next steps • Key issues 2

3. Characteristics of a risk/vulnerability assessment • Requires specialist climate change knowledge & climate information/data • Involves detailed analysis of risks, not just identification (e.g. likelihood, level) • Includes identification & analysis of adaptation options • Likely to be carried out by external specialists (unless ‘light touch’ in-house) • Based on scoping & terms of reference by agency/project staff From Hammill and Tanner 2010b. Climate Risk Screening and Assessment Tools. Presentation given on 22-23 November 2010, Copenhagen. 3

4. Risk, vulnerability & resilience frameworks

5. Risk or vulnerability? • Risk frameworks based on natural hazards literature/approaches • Vulnerability frameworks popularised by IPCC since 2001 • IPCC now moving back towards risk frameworks (SREX 2012; AR5 2013) Vulnerability Risk Measure of likelihood of experiencing negative impacts, &/or of magnitude of those impacts = = function of ( function of ( Exposure , Hazard , Physical mani-festations of climate change & variability – ‘external’ stresses Sensitivity , Adaptive Capacity ) Vulnerability ) Underlying societal/ecological factors that determine ability to anticipate, plan for, cope with, recover from & adapt to evolving climate hazards, their impacts, & other stresses – internal Source of much confusion, but vulnerability & risk frameworks serve similar purposes – evaluating how much harm a system/population is likely to experience as result interaction of external stresses (hazards/exposure) with internal factors that act to amplify or reduce impacts of these stresses. To a large extent, resilience may be viewed as the inverse of vulnerability/sensitivity 5 See Brooks 2003; Fussel and Klein 2006; Fussel 2007; Wolf 2011; IPCC 2012

6. Vulnerability & risk 6 From World Vision 2013. Promoting Resilience in Development Programming. World Vision UK’s Approach.F

7. Resilience • Like vulnerability, can be viewed in terms of a set of attributes that make people and systems more or less resilient. • Other ways of defining, e.g. ecological definition of degree of stress that a system can accommodate without collapsing into a qualitatively different state that is controlled by a different set of processes – ability of system to maintaining its function and preserve its characteristics, or to ‘bounce back’ after a shock. • Development & poverty reduction contexts • Resilience requires more than simply enabling social systems to continue functioning as they were before a disturbance of shock. • e.g. DFID working definition: • “the ability of countries, governments, communities and households to manage change, by maintaining or transforming living standards in the face of shocks or stresses, while continuing to develop and without compromising their long-term prospects”

8. Risk, vulnerability, resilience & adaptation are context specific • Who or what adapts (is vulnerable/resilient)? • To what [hazards] do they need to adapt (be more resilient / less vulnerable)? • What is the result of reduced risk/vulnerability (increased resilience)? • E.g. reduced losses, damages, mortality; increased wellbeing, growth, etc • What are the timescales we are concerned with? • How important is adaptive capacity (compared with resilience / coping capacity)? • What processes are involved? IPCC Third Assessment Report (2001), Chapter 18, p.882 8

9. Context specificity in recent DFID resilience framework Who or what (1) is resilient, to what (2), with respect to what outcome/impact (4)? What are the factors that enable them to cope or prevent them from coping (3)?

10. Tools & approaches for risk & vulnerability assessment

11. Different approaches to assessment

12. 3. Combine information: risk = hazard x vulnerability 2. Map distribution of vulnerability Quantitative risk/vulnerability mapping1. Risk mapping for natural hazards in Columbia 1. Map distribution of hazard Population, economic activities, infrastructure 4. Prioritize high-risk areas / communities for assistance Source: www.itc.nl/ilwis/applications/application01.asp

13. Quantitative risk/vulnerability mapping2. IPCC-style: Vulnerability mapping of SE Asia 2. Sensitivity map 2. Adaptive capacity map 1. Climate hazard map (exposure) 4. Vulnerability map • Droughts, floods, cyclones, landslide exposure, 5m SLR inundation. • Population density & protected areas (%). • HDI, poverty, income inequality, electrification, irrigation, road density, communication • Average 1-3 & classify based on quartiles. Gives overall picture of vulnerability, useful for identifying ‘hot-spots’ where impacts of climate change may be most problematic. Yusuf, A. A. & Francisco, H. 2009. Climate Vulnerability Mapping for Southeast Asia. IDRC/SIDA.

14. A2 and B2 scenarios Hadley Centre model PRECIS2 regional model (China) Example - crop yields in China1 Emissions scenarios RCMs or downscaling Regional crop models System model Yields Impacts GCMs Impacts assessment (model-based) Typical sequence: • Uses range across high and low emissions scenarios • PRECIS model calibrated to Hadley Centre model, so 1st 2 steps implicit • Examines impacts with and without CO2 fertilisation effect • When CO2 fertilisation included, rice yields increase (decrease) under A2 (B2) • Wheat yields increase (decrease) with(out) CO2 fertlisation, but poorer quality • High spatial variability, differences between irrigated & rainfed 1Erda, L., Wei, X., Hui, J. et al. 2005. Climate change impacts on crop yield and quality with CO2 fertilisation in China. Philosophical Transactions of the Royal Society B 360: 2149-1743. 2 http://precis.metoffice.com/ 15

15. Model-based studies should yield range of impacts Box & whisker plots (left) and frequency distribution (right) of projected mean yield changes (%) for 257 ‘observations’ for Africa & South Asia, based on data for all crop types, all crop modelling approaches, all GCMs and (right) all time slices. From Knox et al. 2012.

16. Emissions scenarios Plausible changes System model Impacts Sensitivity studies GCMs CC impacts on Indian wheat1 • Under 2º C warming & 425 ppm CO2 • Yields reduced • Boundaries for 2.5 t/ha and 4.5 t/ha migrate (green to red in graphic) • Using INFOCROP crop growth model • Similar approach to impacts studies but with plausible/arbitrary data representing future climate, rather than GCM output • Illustrative rather than predictive 1Data and graphic from Climate Change Impacts on Agriculture in India: Defra Key Sheet 6, based on work undertaken by Naveen Kalra, Indian Agricultual Research Institute, Delhi. See also Agoumie (2003) for further example of sensitivity to warming of water resources in Morocco. 16

17. ? Scenario A Scenario D Scenario B Scenario C Scenario planning (to assess risks & impacts) Present day ? • What future risks/impacts are implied by climate change scenarios? • Projected climate hazards & other trends allow exploration of risks • May be quantitative/model-based, using climate model output or plausible values • May be qualitative/narrative based, using simple or detailed scenarios • Climate change scenarios allow planners and communities to • Identify potential impacts/implications of range of future conditions • Identify potential responses, adaptation options & associated needs ? ? 17 Cavanna & Abkula 2009 (IIED)

18. Participatory assessments Assess risk baselines (hazards, impacts, livelihoods, social vulnerability & its drivers) CRiSTAL tool uses such approaches for • Hazard and resource mapping • Understanding livelihoods and resources important to livelihoods • Natural, physical, financial, human, social, political resources - access & control • Relative importance of resources • Discussing experience of past climatic changes / trends • Understanding non-climate hazards & interaction with climate hazards • Identify impacts associated with climate hazards & most climate-sensitive resources • Identify & explore response strategies 18 See CRiSTAL User Manual: http://www.iisd.org/pdf/2012/cristal_user_manual_v5_2012.pdf - see resources (Tools)

19. Elements common to most assessments Analysis of current situation • Existing hazards that may affect intervention, system, population • Impacts/consequences of current/recent historical hazards (risks) • Existing drivers & patterns of vulnerability that mediate impacts • Options for vulnerability reduction in short-term H I V O Analysis of future situation • Potential future (evolution of) hazards over relevant timescales • Potential future impacts/consequences of hazards (risks) • Potential future drivers & patterns of vulnerability • Options for longer-term climate risk management / adaptation H I V O • Plan for implementing, monitoring & evaluating options (prelim.) 19

20. Analysis of current situation UKCIP example (business focus) - recent impacts, losses & mechanisms Other contexts might involve identification/mapping of vulnerable populations or areas, exposure to hazards, or impacts/losses, based on recent historical/observational data, to prioritise areas or populations for assistance. http://www.ukcip.org.uk/wizard/wizard-2/- see resources (Tools) 20

21. Analysis of future situation UKCIP example (business focus) - climate projections, other trends Or map how exposure may change based on climate and impacts projections (e.g. water resources). How might vulnerability of different areas or groups evolve? http://www.ukcip.org.uk/wizard/wizard-2/- see resources (Tools) 21

22. DATA & INFORMATION FOR RISK ASSESSMENT 22

23. Assessing the hazard component of risk: 1. Current hazards 23

24. Using observed data to assess existing hazards • Observational data from weather stations, national meteorological services • Historical gridded data or reanalysis data (requires expertise & resources) • World Bank Climate Variability Tool or similar • Relevant trends? E.g. warming, drying? Jan-Dec precipitation (mm/month) for Mwanza, Tanzania, from 1901-2000, based on station data, produced using World Bank Climate Variability Tool. http://iridl.ldeo.columbia.edu/maproom/Global/World_Bank/Climate_Variability/index.html 24

25. Detecting changes – climate indices • Simple indices include mean, max, min temperature, total annual or seasonal rainfall, average mm of rain per day • Should be available from met. services or existing datasets • Indices must be relevant to development contexts/impacts • E.g. onset of wet season, longest period without, rainfall intensity, may be much more important than total annual or seasonal rainfall • Relevant indices constructed from simple indices, e.g. as in IPCC: • Wettest consecutive 5 days –max. consecutive 5-day precipitation • Consecutive dry days – max. no. days with precip. <1mm • High temps. – seasonal/annual max. value of daily max/min temp. • Precip. From v. wet days - % precip. from days >95th percentile • See also indices such as Palmer Drought Severity Index • Power Dissipation Index for storms (see Emmanuele 2005) 25

26. Climate indices in IPCC 2014 - AR5 Ch.2 WG1 p.221 26

27. Using phenological indicators to track climate change Phenological indicators of warming in Germany, from IPCC 2007 AR4 WG2, Ch.1, Fig. 1.4b. http://www.ipcc.ch/publications_and_data/ar4/wg2/en/figure-1-4.html 27

28. Local/traditional/indigenous knowledge • Local tracking of changes where data scarce or absent • Complements scientific data where latter misses the details • Often has significant predictive value/success (Risiro et al. 2012) • Can be calibrated against meteorological data in some instances • e.g. famine correlation with rainfall deficit > 1.3 std dev. N. Nigeria (Tarhule & Woo 1997) Caveats • Erosion of knowledge makes forecasting less reliable (Risiro et al. 2012, Enock 2013) • Disappearance of biological indicators makes forecasting more difficult • Changing relationships between indicators & phenomena or interest (due to CC) • Easy, simple explanations (e.g. drought) may mask more complex situations See also: Nakashima, D.J., Galloway McLean, K., Thulstrup, H.D., Ramos Castillo, A. and Rubis, J.T. 2012. Weathering Uncertainty: Traditional Knowledge for Climate Change Assessment and Adaptation. Paris, UNESCO, and Darwin, UNU, 10 pp. 28

29. Assessing the hazard component of risk: 2. Future hazards 29

30. Using climate models to assess future hazards • Global & regional climate model output for broad characterisation • Downscaled model output for higher resolution projections • Limited availability, resource intensive, precision ≠ accuracy! • Need multiple projections to define range of possible conditions • Multiple simulations, preferably from multiple models • How much agreement across models / what is range? Projected mean annual temperature anomalies for Tanzania. UNDP Climate Change Country Profile: http://www.geog.ox.ac.uk/research/climate/projects/undp-cp/index.html?country=Pakistan&d1=Reports 30

31. Envelopes of uncertainty for climate projections Projected changes in monthly average rainfall, temperature, and number of days with heavy rainfall, for the Tanzania for 2020-39, from a range of different models. Graphics generated using World Bank Climate Change Knowledge Portal: http://sdwebx.worldbank.org/climateportal/index.cfm 27

32. Models predicting higher precip. Mean annual precipitation Model consensus doesn’t remove uncertainty • IPCC projections for East Africa: wetter conditions, high confidence, good model agreement (Christensen et al. 2007) • More recent work (Funk et al. 2008; Park Williams & Funk 2010): 15% decline in main growing season rainfall since 1980 in E & Southern Africa bordering Indian Ocean Graphic sources (clockwise from top left): IPCC AR4 WGI Ch.11 (Christensen et al.,2007, p.869); UNHCR/USAID/OCHA; Funk et al. 2008, PNAS vol. 105 no. 32: 11081-11086. 32

33. Scenarios for risk assessment • Construct scenarios spanning range of projections (‘envelope of uncertainty’) • E.g. hottest-driest, coolest-wettest, hottest-wettest, coolest-driest • What are implications of scenarios for activities of interest? • Are scenarios compatible with observed trends? • If not, what would be implication of continuation of these trends? • Are there other reasons for lack of confidence in model projections? • E.g. poor understanding of key physical processes, more recent studies • Use models to guide assessment but be aware of their limitations • Coarse spatial & temporal resolution - addressed by downscaling (resource intensive) • Limited representation of some extremes & aspects of variability • Possibility of systematic bias due to paramaterisation errors (limits to knowledge) • Non-model based approaches • How would systems/populations cope with repeat of historical extremes? • How would they cope with plausible (or even arbitrary) changes? See also: http://weadapt.org/knowledge-base/using-climate-information/guide-to-using-climate-information http://weadapt.org/knowledge-base/using-climate-information/introduction-to-climate-modelling 33

34. Sources of climate data/information 34

35. Documentary sources UNDP-Oxford Climate Change Country Profiles • Climate data analysis for 52 developing countries, nationally averaged data • Some years old now, based on old SRES climate scenarios • http://www.geog.ox.ac.uk/research/climate/projects/undp-cp/ • IPCC reports (http://www.ipcc.ch) • Regional and sectoral chapters (WGII) and regional projections (WG1) useful • Reference numerous individual studies that can be used in assessment National & other country documents • NAPAs, NAPs, National Communications (governments) • Country profiles (development agencies) • National & other vulnerability assessments Academic studies & technical reports • Research papers, peer-reviewed journals • Technical studies for specific sectors, projects, etc + National met. & hydro. services, online data portals 35

36. World Bank Climate Change Knowledge Portal http://sdwebx.worldbank.org/climateportal/index.cfm • Historical climate data & Historical Variability Tool (station data) • Future projections - averaged and downscaled for key variables (multi-model) • Climate Risk and Adaptation Country Profiles - hazards, baselines, vulnerabilities • Development & climate impacts • World Bank funded activities 36

37. Climate Information Platform http://cip.csag.uct.ac.za/webclient2/app/ • Hosted at University of Cape Town • Global map of observed sea-level trends with data for individual stations • Observed & projected (downscaled) climate data for Africa, by station • Plot ranges across 10 models, with 10th & 90th percentile values & extremes • Older version with SRES scenarios, new version with RCPs 37

38. Assessing the societal components of risk/vulnerability 38

39. Societal component of risk - what to measure? • Some general development-related factors such as poverty, isolation, poor access to resources & services, may tend to make people more vulnerability/sensitive to climate-related shocks & stresses (hazards) • However, drivers of vulnerability will tend to be highly context specific • To assess societal component of vulnerability/risk, need to identify what factors are important & develop ways of tracking these (e.g. indicators) • Factors/attributes that affect people’s ability to anticipate, avoid, plan for, cope with, recover from & adapt to stresses & shocks • Identify these during project development phase or early in implementation phase through participatory assessments • Can then be tracked using indicators of vulnerability, resilience, coping capacity, adaptive capacity, etc.

40. Indicators for vulnerability/resilience • Indicators may be: • Qualitative, e.g. based on beneficiary perceptions of how easily they can cope with particular stresses/shocks; how easily they can access certain resources that make them more resilient; whether their ability to cope with a particular shock or stress has increased or decreased; etc. • Quantitative – continuous, based on measurement of a continuous variable such as household income, time to recover from previous shocks, etc. • Quantitative – binary, based on assigning a beneficiary a score of 0 or 1 according to whether they meet a particular criterion. • Quantitative – categorical or score based, based on assigning a beneficiary a score (e.g. 0-3 or 0-5) representing a category or level of resilience (e.g. low, moderate, high). • Indicators may be disaggregated or combined into composite indices – to combine, standardise by converting to common scoring system or values scaled from, e.g. 0-1. Consider weightings and use of thresholds (e.g. vulnerable or not vulnerable).

41. Data collection • Secondary data from public sources, national agencies (if appropriate) • Primary data collection – sampling beneficiary populations • Panel surveys: track individuals/households over time, resource intensive • Random sampling: lower scope for contextual data, less resource intensive • Use of control/comparison groups: not always feasible/practical • Sampling strategy • Beginning & end of project; annually, ex post evaluation – has vulnerability been reduced; resilience, coping/adaptive capacity enhanced? • Can measured changes be attributed to interventions? See also CRiSTAL tool (Slide 17)

42. From assessment to action Design phase • Identification & evaluation of potential adaptation options (e.g. during detailed risk/vulnerability assessment or light-touch assessment) Implementation phase • Review, analysis, prioritisation & selection of options with stakeholders • Preparation of implementation plan with next steps, roles, timeline, resource needs, system for monitoring & evaluating success E.g. USAID 2007. Adapting to Climate Variability and Change: A Guidance Manual for Development Planning http://pdf.usaid.gov/pdf_docs/PNADJ990.pdf From Hammill and Tanner 2010a. 42

43. KEY QUESTIONS • How are the key concepts of risk, vulnerability, and/or resilience relevant in your work? • How might you develop a framework for risk/vulnerability assessment in your institutional/project/programme context? • How would you frame/scope a vulnerability/risk assessment (e.g. that you were commissioning for a specific intervention) • What questions need to be asked? • Who should do the assessment? • What type of assessment is appropriate? • What data sources & types of analysis are needed? 43