Climate Resilience Analysis Framework: Testing The Resilience Of .

The Climate Resilience Analysis FrameworkThis framework represents a general approach forassessing the resilience of existing natural and engineeredsystems, and supports the development of options forimproving system resilience. In this context, a system isdefined as the interaction of physical characteristics (e.g.natural characteristics and built infrastructure) with anyrelevant operating characteristics (e.g. operating rules) tofulfill one or several functions. The framework is illustrated in Figure 1 and contains thefollowing five elements: The system is the central concern of the analysis, andthus requires an assessment of how the systemshould perform. You can define ‘systemperformance’ in a number of ways, including binarysuccess/failure criteria or quantitative performancemeasures, as well as across multiple economic, socialand/or environmental measures;A climate ‘stress test’ is then applied to the system toassess the rate of system performance degradationand/or identify situations under which it can fail;Multiple ‘lines of evidence’ are then used tounderstand possible future climate changes. Lines ofevidence may include climate model projections (bycombining global climate models with dynamicaland/or statistical downscaling, or bias corrections),historical climatic changes, expert judgement and/oranalogues from paleo records;Performance of multiple alternative options forstrengthening system resilience (e.g. infrastructureaugmentation, land use planning, operations,demand management, etc.) can then be analyzedand compared; andDecision-analytic approaches are then used todetermine the preferred system managementoption. This analysis can proceed in multiple ways,depending on user preference and interpretation ofclimate uncertainty (e.g. probabilistically or throughscenarios).Throughout the process there are a number of points tocheck in with system and climate experts. Therefore, toensure that this framework has the ability to informdecision making it is advised that it is underpinned by astakeholder engagement strategy.What follows is a detailed description of how toimplement the five-step framework for a system. Youmay find that some iteration between Steps 2 to 4 isrequired to meet individual needs of your investigation.Figure 1: Elements of the climate resilience analysis framework (top) and framework process flow (bottom)Climate risk analysis framework 3

Step 1: Define the problem and system performancemeasuresIn this Step you will describe and analyze the systemunder consideration. The aim is to define the problem(s)you are attempting to solve in order to achievesustainable management of the system. Firstly, the systemdomain needs to be identified. This includes the system’sphysical and operational characteristics, as well as thesystem boundaries. The system domain forms the centralreference point for the remainder of the framework.Defining the system domain includes consulting decisionmakers and seeking expert knowledge from systemoperators (e.g. utilities, government agencies). Thesediscussions should be a creative exploration of the systemdomain, recognizing that how a system is defined (e.g.where the boundaries are set) can often stronglydetermine the assessment of system resilience or theavailability of options to improve performance.Performance measures can then be developed to quantifysystem performance. These can represent a range ofsocial, economic and environmental measures of thesystem. Performance measures represent the systemvalues important to stakeholders and often we need toconsider hidden costs, such as opportunity costs, andtrade-offs of various kinds. Measures can include averageperformance (e.g. average annual net profit) orprobability-based measures (e.g. probability of systemfailure).Typical questions to ask in this step include: What is the purpose of the system?How is the system defined, and what are the systemboundaries?How can system performance be measured? Arethere clear success/failure criteria or is performancerepresented using multiple measures?What non-climate factors should be considered aspart of understanding overall system performance(e.g. population growth, energy pricing, and systemoutages due to maintenance issues)?What alternative system management options maybe available and should be considered?You can consider alternative system management optionsat this point or return to this in Step 4. Depending on thesystem boundaries and options available to decisionmakers, alternative system management options mayinclude modification of operating rules, system reoptimization, infrastructure augmentation, economicsignaling (e.g. modification of resource prices) and so on.Following the identification of key performance measures,possible climate conditions that may affect systemperformance should be established. Potential climateconditions may include seasonality, extremes, annualtotals and timing of variables such as rainfall, temperatureand evapotranspiration.Questions to consider include: What are the specific weather and climaticconditions that could affect system performance?What climate conditions could present risks to thesystem?How have climate variables been used in thedecision making?What range of climatic change could the system beexpected to experience? Within what climatebounds should the system be evaluated?In light of this investigation, choose (or develop) a suitablenumerical system model based on the identifiedperformance measures and the set of climate conditions.The ‘system model’ in this context is a model thattranslates changes in the climate conditions into thesystem performance, and it may represent thecombination of multiple separate computer models (e.g.hydrology, reservoir operation and demand models).The system model is used to evaluate system performanceacross a wide range of scenarios and system managementoptions in Steps 2 to 4. Hence it is critical that the model isable to capture the system’s response to current andalternate climates, and is able to be easily adjusted torepresent alternate system management options. Aninitial sensitivity analysis is recommended to understandthe model’s intrinsic behavior.Outcomes: Problem scope defined,performance measures set & systemmodel developed4 Climate risk analysis framework

Step 2: Stress test the systemIn this Step you will evaluate how system performanceresponds to changes in the weather and climate variableproperties identified in Step 1.Begin with a preliminary investigation of the climate variableproperties identified in Step 1. This investigation also relies onthe chosen system model. Preliminary investigation questionsinclude: Options for visualizing system performance maps include: Binary pass/fail regions (if the system has distinctperformance thresholds)Heat maps/contours plots of the systemperformanceDo changes in the a priori identified climate variableproperties produce changes in system performance?Does the identified range of changes from Step 1encompass the changes projected by global climatemodels? CHECK IN POINTCheck in with your system and climate experts toconfirm whether the climate variable propertiesidentified in Step 1 are producing logical changesin system performance.Once the sensitivity of the system to the selected climatevariables is confirmed, generate perturbed time series thatcover the range of plausible climate conditions identified inStep 1. These sets of perturbed time series are used todrive the system model and are termed ‘scenarios’.Options for generating scenarios include: Simple scaling: multiplicative/additive changes toobserved climate time seriesStochastic generation: time series are generatedusing a stochastic generator with the requiredpropertiesNext use all the sets of perturbed climate time series todrive the system model and generate the systemperformance measures.Now you have collated the system performance for all setsof perturbed climate time series, visualize the systemperformance across the range of investigated changes as asystem performance map.Example system performance mapThe system’s response to the range of perturbed timeseries provides insight into the system’s sensitivity tochanges in the climate variables and the characteristics ofthose variables (e.g. averages, extremes). Use theoutcomes of this stress test for system diagnosis whereunacceptable performance is encountered. A thoroughunderstanding of the reasons for unacceptable systemperformance will be required in Step 4.CHECK OUT POINTIf you find your system is very resilient consultwith your stakeholders to check if you need toproceed to Steps 3 – 5.Outcomes: Quantitativeunderstanding of the system and itssensitivities to climate variationClimate risk analysis framework 5

Step 3: Climate projections and other lines of evidenceIn Step 2 you identified critical variables that affect thesystem’s vulnerability. Now you can place this systemunderstanding in context with climate projections fromglobal climate models and other lines of evidence. This isdone by superimposing the climatic changes projected byGCMs or other sources of climate information on to thesystem performance maps produced in Step 2.Use the system sensitivities and thresholds at which thesystem is pushed beyond acceptable operating conditionsuncovered in Step 2 as practical guidance on whatinformation on projected change should be considered.This can be done in consultation with climate experts.Now superimpose downscaled and bias-corrected GCMprojections onto the system performance maps. Thisprovides an indication of the plausibility of the climateconditions causing system failure.Other sources of climate information relevant to theproblem specifications can also be visualized, such as theobserved historical climate variability, the limits ofengineering specifications (e.g. 1-in-100 year floods fromwhich the system may have been originally designed),expert knowledge, and/or paleo-climate information.CHECK IN POINTThis is a good time to consult climate expertsabout possible change to key climatevariables. How well do the climate models represent theprocesses and climate variables your system ismost sensitive to? What other ‘lines of evidence’ would be usefulin light of the sensitivities uncovered in Step 2Selected climate projections shown as black dots on systemperformance mapCHECK IN POINTThis is a good time to consult the projectstakeholders to see what time periods (e.g. timeslices) are relevant to your system.A key consideration is how to interpret alternative ‘lines ofevidence’. Whereas Step 2 involved stress testing thesystem against climate time series (or ‘scenarios’) thatrepresent hypothetical alternative climate states, in thisstep we are concerned with identifying parts of theclimate space that are more or less likely, for examplethrough the use of projections from climate models. Thequestion of whether or not different lines of evidence canbe interpreted as probabilistic statements of the futurehas important implications on which decision making toolsare most appropriate (Step 5), and particularly whetherapproaches that account for ‘deep uncertainty’ areneeded.Outcomes: System performancemaps incorporating other lines ofevidence6 Climate risk analysis framework

Step 4: System management optionsAgain this should be a creative and explorative process inwhich it is important to think laterally. It is useful to keepthe problem definition developed in Step 1 in mind. It maybe necessary to iterate on the problem definition orperformance measures in Step 1 depending on thefindings of Steps 2 and 3.EXPLORATORY PROCESSExamples of questions useful for thisexploratory process include:This Step revisits the potential system managementoptions and analyses each option in turn.In Step 2 you gained a deeper understanding of thesensitivities and behaviour of the system. Here this deeperunderstanding and diagnosis of existing performancesensitivities becomes critical in evaluating systemmanagement options.SYSTEM MANAGEMENT OPTIONSRemember there may be a large number ofalternative options to achieve the same objective.Some examples of options include: modification of operating rules system re-optimisation infrastructure augmentationeconomic signaling (e.g. modification ofresource prices)In light of the investigation in Step 2 reflect on theidentified system management options identified in Step1. If necessary iterate with your stakeholders/systemoperators.CHECK IN POINT What alternative system managementoptions are available?Should a variation of operating rules beinvestigated?Should additional or alternativeperformance metrics be considered?How constrained is the system?Are any of the system managementoptions similar in the treatment of theidentified climate variables?At the end of this exploratory process you should have anagreed set of system management options that requirefurther evaluation.At this point, you will need to repeat Steps 2 to 3 for allagreed system management options. At the end of thisprocess you will have visualized the system performanceas system performance maps incorporating other lines ofevidence for each agreed system management option.Outcomes: System performancemaps incorporating other lines ofevidence for all considered systemmanagement optionsCheck in with your stakeholders regarding systemmanagement options.Climate risk analysis framework 7

Step 5: Decision analysisThe final outcome of this Step is to arrive at a preferredoption that can be implemented within the parametersidentified in Step 1.Based on the analysis carried out in Steps 2 to 4, thevarious system management options must be evaluated.This includes considering their feasibility, costs, benefitsand potentially political will to investigate alternativeinfrastructure investments. There may also be a need tobalance economic benefits with environmental and socialvalues.benefit analyses, and/or quantitative risk assessmentsmay be appropriate) or as scenarios (in which caserobustness approaches may be required). The scenariobased approach is becoming increasingly accepted giventhe recognition that climate models are unlikely toaccurately represent multiple key physical processes thatare likely to be relevant to a given system, requiring afocus on ‘what-if’ scenarios rather than ‘best estimates’ offuture climate.Use the system performance ‘stress test’ carried out inStep 2 alongside the alternate climate informationoverlaid in Step 3 to characterize system resilience. Thisway, it becomes more apparent which climate states willpresent the most challenges, and indicates how muchchange in climate can occur before the system is no longerable to provide the expected services. Combining this withthe alternative options identified in Step 4, the conditionsat which an option becomes preferable can bedetermined, enabling the development of adaptivepathways.Regardless of the approach taken to decision makingunder uncertainty, the analysis in Steps 1-5 provides aholistic view of key modes of system vulnerability, a‘multiple lines of evidence’ view of future climate, and adetailed exploration of alternative system design optionsthat may increase system resilience. The combination ofthis information can form the basis for a finalrecommendation, which may include the ‘do nothing’option, implementation of alternative systemmanagement options, or the identification of key triggerpoints at which action is required as part of an adaptivepathways approach.Questions that may assist in your decision analysis include: Under what conditions does a system managementoption become preferable? What are the trade-offs between system managementoptions? Can an adaptation pathway be developed?Decision-centric approaches can be tailored depending onwhether alternate climate futures are interpretedprobabilistically (in which case approaches such as cost-8 Climate risk analysis frameworkCHECK OUT POINTYou’re done.Outcomes: Assess system resilienceacross a range of systemmanagement options and determinefinal recommendations

GlossaryAdaptationThe process of adjustment to actual or expected futureclimate and its effects (IPCC, 2014).Bottom-up approachBottom-up climate assessment begins in the vulnerabilitydomain. It takes important system characteristics andlocal capacities into account before the sensitivity androbustness of possible adaptation options are testedagainst climate projections (e.g. GCM outputs).Climate changeClimate change refers to a change in the climate’s statethat can be identified (e.g. via statistical tests) and thatpersists for an extended period, typically decades orlonger (IPCC, 2014).Climate projectionClimate projections are typically derived using climatemodels and are the simulated response of the climatesystem to a scenario of concentrations of greenhousegases and aerosols or future emissions.ResilienceThe capacity of a system to cope with disturbance,hazardous event or trend, responding or reorganizing soas to still retain essentially the same function, structure,identity, and feedbacks as well as retaining capacities foradaptation (IPCC, 2014).Top-down approachTop-down approaches for climate impact assessmentbegin by downscaling climate model projections and thenusing these downscaled projections to drive variousmodels in order to develop expectations for changes inhydrology, vegetation, social systems, etc.VulnerabilityVulnerability is the degree to which a system, or elementof a system, may adversely react as a result of theoccurrence of a hazardous event. This concept impliessome risk combined with the system’s ability to cope andthe level of economic and/or social liability associatedwith an event’s occurrence.Deep uncertaintyThe “condition in which analysts do not know or theparties to a decision cannot agree upon (1) theappropriate models to describe interactions among asystem’s variables, (2) the probability distributions torepresent uncertainty about key parameters in themodels, and/or (3) how to value the desirability ofalternative outcomes” (Lempert et al., 2006).DownscalingThe process by which coarse GCM climate projections aretransformed into higher resolution climate information.Global climate model (GCM)GCMs are numerical representations of the global climatesystem

To evaluate system resilience, the climate resilience analysis framework was applied to the MAR scheme. The system captures stormwater from a 16 km2 residential and industrial catchment, which passes through two storage basins and wetlands for cleaning and sediment reduction. Four wells inject and extract water from the aquifer for reuse for .