FILEY CASE STUDY: A GUIDE TO BENEFIT APPRAISAL

Transcription

1 FILEY CASE STUDY: A GUIDE TO BENEFIT APPRAISAL Introduction This case study provides the user with a walkthrough of the costbenefit appraisal process. Cost-benefit appraisal (CBA) is used by authorities to justify spending money on flood schemes. The cost is the full price of a scheme and includes construction costs. The benefit is how much a scheme reduces flood risk and prevents a flood causing damage (damage avoided). Filey flood, July 2007 George Gingell Surface Water Flooding in Filey, North Yorkshire in 2007 affected hundreds of residential and a handful of commercial properties including an Ambulance Station. Depths of flooding were up to 1 to 1.5 metres, though many properties were flooded to just a few centimetres. Rescue and evacuation were an important part of Incident management. Scarborough Borough Council and their consultants discussed options, with a short list as follows: Do Nothing Do Minimum - Maintenance Do Minimum - Demountable & Temporary Solutions Do Something - Earth bund and storage option (without environmental enhancements) Do Something - Earth bund and storage option (with environmental enhancements) This benefit appraisal case study evaluates the benefits of the earth bund and storage option (without environmental enhancements). This guide identifies 7 steps to complete the CBA process: Step 1 - Receptor Data in the benefit area Step 2 - Modelling the flood for each return period Step 3 - Calculate the Annual Average Damages (AAD) Step 4 - Repeat Steps 2-3 for the other scenarios (Do minimum, Do nothing) Step 5 - Robustness analysis: visualisation allows auditing of errors in modelling Step 6 - Benefit Cost Summary Step 7 - Visualisation of results for transparent decision making 1

2 Step 1 - Receptor Data in the benefit area The first step is to identify the benefiting area and the receptors at risk from flooding. In this example three benefiting areas are identified within Filey, North Yorkshire: Wooldale, Seadale and Whafedale (Figure 1) Wooldale Seadale Wharfedale Figure 1 The three sub-areas of the Filey Case Study OpenStreetMap contributors This data is available under the Open Database License Within each area the receptors at risk are identified using the National Receptor Dataset (NRD), which georeferences all property, infrastructure and environmental assets (provided by The Environment Agency). In this example, we will focus solely on the direct impacts to residential property. Figure 2 illustrates the NRD data, visualised as a GIS output, where residential properties are categorised as a dwelling. Within the NRD, further detail is provided for each dwelling, such as flat, semi-detached etc. 2

3 Figure 2 An illustration of property points In some cases, information for the property point is missing (this is shown as the blue rectangles in Figure 2). Semi-automated tools such as Google Streetview make auditing of data and gap filling fast and efficient. The NRD does not contain the property threshold (the point at which flood damage begins, usually the door step or air brick where property has suspended floors). This needs to be estimated or measured during a field survey. Within the NRD, each type of residential property is given an MCM code corresponding to a depthdamage curve (figure 3). It is important to ensure that the correct curve is attributed to the property in question. MCM code Property Type Residential Sector Average 1,031 21,598 32,106 50, Detached ,433 45,144 72, Semi-detached 1,254 19,854 29,523 47, Terrace 1,089 18,161 26,933 42, Bungalow ,583 43,228 71, Flat ,645 28,965 41,289 Figure 3 A small sample of an MCM residential depth-damage curve These data are also available for all non residential data types for different charcacteristics of flooding (short and long duration/with and without flood warning). Other data is available on the MCM-Online for: Emergency services costs 3

4 Evacuation costs and temporary accommodation costs Damage to vehicles Traffic disruption Infrastructure Losses Agricultural Losses Recreation and ecosystem services This case study will follow the process for evaluating residential property damages and link the results to other receptor losses and damages contributing significant further loss or damage. One of the properties suffering minor flooding is 47 Unnamed Road off Scarborough Road. We will follow this property through the benefit appraisal process. #47 is a semi-detached property in the Seadale area; flooding here was less significant than in the two other neighbouring areas of Wooldale and Wharfedale. The corresponding depth-damage curve for this property would be the one highlighted in Figure 3. At this stage of the process, you will know the charateristics of the receptor (property) in the benefit area (type, property threshold and its associated depth-damage curve). The next stage is to identify the flood depth at each property for all return periods (RPs). Step 2 - Modelling the flood for each return period The 2007 flood was only one of a potential series of greater or lesser flood events. Mathematical modelling (in this case a product called InfoWorks) is used to simulate the flood depths for a range of these flood events. The threshold flood is where flooding commences and damages are zero. A range of sucessive floods allows the estimate of annual average damage for all floods over a long time period. Six such flood events were modelled for Filey. #47 High (0.6m) Low (0m) Figure 4 A 5-year flood with an annual probability of 20% 4

5 High (0.6m) Low (0m) Figure 5 A 10 year flood with an annual probability of 10% High (1.0m) Low (0m) Figure 6 A 30 year flood with an annual probability of 3.33% 5

6 High (1.4m) Low (0m) Figure 7 A 50 year flood with an annual probability of 2% High (1.5m) Low (0m Figure year flood with an annual probability of 1% 6

7 High (1.5m) Low (0m) Figure 9 A 200 year flood with an annual probability of 0.5% Step 3 Calculate the Annual Average Damages (AAD) For every property including #47 we now have: A property type. A threshold level where flood water enters the property. A flood depth for each of the modelled return periods, and as a result; A damage value for every property for every return period. Every input and calculation is stored against every property in the Environment Agency Standard Spreadsheet (Figure 10). At this stage, the first data checking process can begin. For example, an error can be found for the #12 Unnamed Road property, where the 30 year damage is greater than the 50 year damage. Figure 10 Individual event property damage for each residential property under Do Nothing Scenario using EA standard spreadsheets (Seadale sub area) 7

8 Property #47 has zero damages until the 200 year flood event. This may not be an error, as it may not have flooded in Unravelling such errors will be done later as part of a robustness analysis. Damages are capped per property when the annual average damage accrued over the lifetime of any improvement scheme are greater than the market price of the property (taking average regional prices per property type). No property can have more damage than the total value of the property. Within the same EA Standard Spreadsheet, the annual average damages (AAD) is automatically calculated. Present Value of Damage is 1.2 million using the fixed Treasury Discount rate and a 100 year time Horizon (Figure 11). Summary Annual Average Damage Sheet Nr. Client/Authority Scarborough Borough Council Project name Option: Filey Flood Alleviation Strategy: Seadale only Do nothing Project reference - Base date for estimates (year 0) Aug-11 First year of damage: 0 Prepared (date) 22/09/2011 Scaling factor (e.g. m, k, ) Last year of period: 99 Printed 12/05/2014 Discount rate 3.5% PV factor for mid-year 0: Prepared by JC Checked by Applicable year (if time varying) Checked date Average waiting time (yrs) between events/frequency per year Total PV Infinity Damage category Damage Residential property , , , , ,274 1,150, ,651 Ind/commercial (direct) ,578 7,578 8,588 3,464 Ind/comm (indirect) - - Traffic related - - Emergency services - - 2,858 8,289 19,942 25,405 45,128 64,907 41,166 Extra Heating - 3,025 12,705 37,510 58,685 64,130 69,575 75, ,044 Alt. Accommodation 3,348 6,078 12,156 25,546 52,326 85, ,276 91,820 Road Damage 2,376 12,277 51,390 85, , , , Total damage - 8,749 84, , , ,367 1,127,762 1,560,223 Area (damagexfrequency) - 1,312 4,685 11,411 5,357 6,213 4,560 6,720 Total area, as above 40,259 PV Factor, as above Present value (assuming no change in damage or event frequency) 1,200,225 1,200,225 Notes Area calculations assume drop to zero at maximum frequency. Default value for the highest possible damage assumes continuation of gradient for last two points, an alternative value can be entered, if appropriate. One form should be completed for each option, including 'without project', and for each representative year if profile changes during scheme life (e.g. sea-level rise) Residential property, Industrial / commercial (direct), and Other damages are itemised in Asset AAD sheet and automatically linked to this sheet Figure 11 - Calculation of Annual Average Damages (Seadale) and Present value of Damages with other receptor damage/disruption included The Loss Probability Curve plots the total damage for each return period. The area below this curve represents the annual average damages ( 40,259) (Figures 11 and 12). The curve visualises a good fit. If the curve is not asymptotic to both axes then the annual average damages are more than likely to be overestimated. 8

9 Project: Filey Flood Alleviation Strategy: Seadale only Option: Do nothing Damage Frequency Damage Frequency Figure 12 The loss-probability curve for Seadale, do nothing option. Step 4 - Repeat Steps 2-3 for the other scenarios (Do minimum, Do nothing) The model is then run to establish the residual damages with each option as it is unlikely that any option will eliminate all flooding completely. For schemes with 100 year or more standard of protection it is unlikely that residual damages will be more than 5% to 10% of Do Nothing damages. For simplicity in the case of Filey they are assumed to be zero. Step 5 - Robustness analysis: visualisation allows auditing of errors in modelling The model is adjusted to eliminate obvious anomalies, picked up through visualisation of the results. The scenario illustrated in Figure 13 cannot happen in nature as the 30 year depth appears shallower than 10 year depth so model refinement was necessary. 9

10 Figure 13 The initial modelling run showing erroneous results Indicates where 30- Year RP flood depths shallower than 10-year depths. Checking the modelled damages against reality is difficult if not impossible as by no means all modelled flood events have occurred, usually just one as in 2007 case. Using insurance claims and historic data can assist to some degree. Though insurance data uses financial replacement costs whilst FDGiA damages do not include VAT (a transfer payment) or betterment (it is assumed that all damaged goods are approximately half-way through their lives.) Figure 14 - Selected insurance data for the 2007 flood The following sources can be applied to ground truth the veracity of the modelling: 10

11 Historic flood maps Insurance data Web sources (flicker, youtube) Anecdotal evidence Step 6 - Benefit Cost Summary Once all errors have been eradicated and final results obtained, the cost-benefit assessment can be completed using the EA Standard Spreadsheet. The cost of each Option (total scheme: Wharfedale, Wooldale and Seadale) needs also to be included. Costs include capital, maintenance and operational costs over a 100 year period (adhering to Government long termism for major infrastructure schemes) discounted using Treasury discount rates. A 60% optimism bias is usually added at initial appraisal stage to reduce the difference between appraisal and outturn costs after the scheme is complete. The Project Summary Sheet (Figure 15) automatically calculates the decision-making criteria (net present value, cost-benefit ratio and incremental cost-benefit ratio). Property #47 is part of the Seadale sub area and Option 3c costs were divided equally between the 3 sub areas. The cost-benefit summary indicates that on its own using FDGiA rules the Seadale scheme is not cost beneficial for any of the options with a 0.9:1 ratio for Option 3c. 11

12 Project Summary Sheet Client/Authority Prepared (date) 22/09/2011 Scarborough Borough Council Printed 12/05/2014 Project name Prepared by JC Filey Flood Alleviation Strategy: Seadale only Checked by Project reference Checked date Base date for estimates (year 0) Aug-2011 Scaling factor (e.g. m, k, ) (used for all costs, losses and benefits) Year Discount Rate 3.5% 3.00% 2.50% Optimism bias adjustment factor 60% Costs and benefits of options Costs and benefits Option number Option 1a Option 2a Option 2b Option 3cg Option 3dg Option name Do Nothing Do Minimum - Maintenance Do Minimum - Demountable & Temporay Solutions Earth bund and storage option (without env.enhanceme nts) Earth bund and storage option (with env.enhanceme nts) COSTS: PV capital costs ,000 0 PV operation and maintenance costs 0 1,308,281 1,308,281 0 PV other 0 16,333 32,667 0 Optimism bias adjustment 0 794, ,568 0 PV negative costs (e.g. sales) 0 98, ,973 0 PV contributions Total PV Costs excluding contributions 0 2,021,383 2,029,543 1,300,000 0 Total PV Costs taking contributions into account 0 2,021,383 2,029,543 1,300,000 0 BENEFITS: PV monetised flood damages 1,200,225 1,200,225 1,200,225 - PV monetised flood damages avoided 0 0 1,200,225 1,200,225 PV monetised erosion damages PV monetised erosion damages avoided (protected) Total monetised PV damages 1,200,225 1,200,225 1,200, Total monetised PV benefits 0 0 1,200,225 1,200,225 PV damages (from scoring and weighting) PV damages avoided/benefits (from scoring and weighting) PV benefits from ecosystem services Total PV damages 1,200,225 1,200,225 1,200, Total PV benefits 0 0 1,200,225 1,200,225 DECISION-MAKING CRITERIA: excluding contributions Based on total PV benefits ( in cludes benefits from scoring and weighting and ecosystem services) Net Present Value NPV -2,021,383-2,029,543-99,775 1,200,225 Average benefit/cost ratio BCR Incremental benefit/cost ratio IBCR Highest bcr Figure 15 Project Summary Sheet showing all Options for the Seadale case 12

13 Step 7 - Visualisation of results for transparent decision making It can sometimes be useful to convert a spreadsheet of figures into a format which is more visually interesting. This can be represented in graphs, but two alternative methods are heat maps (Figure 16) or damage stacks (Figure 17). The darker colours in this heat map indicate higher damages, so we can see that the sub-area of Seadale shows the least damage per square metre and Wooldale the highest. Wooldale Seadale Damage per square metre 550 Wharfedale Figure 16 - Heat Maps of damage density for the 3 sub areas 13

14 The damage stack map below shows the flood extent and the concentration of damages. Simply, the higher the building stack, the higher the damages. Figure 17 - Visualisation of the Do Nothing option, showing present value of damages for the three subareas. Source: Black & Veatch GIS Team,

15 Additional Considerations: Partnership Funding (Wharfedale, Wooldale and Seadale combined) The Defra Partnership Funding (PF) calculator determines the contribution to flood and coastal erosion scheme costs from Flood Defence Grand in Aid (FDGiA) (funding from government) and partner contributions (stakeholder funding from local authorities, business, community organisations, residents etc.) This calculation is based on a series of Outcome Measures (OM). OM1 relates to the total benefits achievable by the flood scheme. These include avoidance of damages to housing, business, agriculture, local government, communications, infrastructure, utilities and public health. OM2 relates to households moving from one category of flood risk to a lower category. Households must be at direct risk of flood damage and have been built or converted into housing before January 2012 to be counted. The greater the level of social deprivation, as measured by the Office for National Statistics (ONS), and the greater the movement of risk before and after a scheme has been implemented the higher the overall FDGiA contributions will be. Any shortfall has to be made up by contributions from partners. FDGiA Calculator, based on interim funding arrangements announced 23rd May 2011 epublications Catalogue Product Code - FLHO0511BTXS-E-E Project Name/ref: Filey Flood Alleviation Strategy Study - Option 3c (ALL SCHEME) Key Input cells Summary: prospect of FDGiA funding Calculated cells "FDGIA Contribution": 10,511,974 "Raw OM Score": % Scheme Benefit to Cost Ratio: to 1 Cost saving and/or external contribution required: - Effective return to taxpayer: to 1 Less scheme contributions secured: Effective return to area: n/a to 1 "Adjusted OM Score": % Result: Potential candidate for FDGiA funding dependant upon funding availability FDGiA required for next phase(s): 2,826, Scheme details Who will maintain asset? PV Whole-Life Costs: PV Whole-Life Benefits: Cash cost of next phase(s): Duration of Benefits: Average flood damages: Construction phase? LA million million million 100 years 30,000 per household Yes - costs for approval include construction 2. Qualifying benefits under Outcome Measure 2: households better protected against flood risk Number of households in: Before After Change due to scheme 20% most deprived areas % most deprived areas % least deprived areas At: Moderate Significant Very Moderate Significant Very Moderate Significant Very risk risk significant risk risk significant risk risk significant risk risk risk Annual damages avoided, compared with a household at low risk ,350 Change in household damages, in: Per year Over lifetime of scheme Qual. benefits (discounted) 20% most deprived areas - - OM2 (20%) % most deprived areas - 206,250-20,625,000 OM2 (21-40%) 6,159,319 60% least deprived areas - 1,273, ,395,000 OM2 (60%) 38,044,432 Figure 18 The Defra FDGiA calculator showing details of the Filey study In the Filey case, the movement of flood risk to properties from very significant to moderate or low risk is so significant that, when combined with high numbers of these properties within the 15

16 21% to 40% most socially deprived in England (Figure 19), that no partnership funding is required. Within EA investment limits, total funding could be made available through Central Government Flood Defence Grant in Aid. Figure 19 Change in household risk in Filey before and after the scheme Visualising Social Deprivation data Examples are provided for social deprivation and visualising contributors to support transparent decision for partnership funding. Lower social deprivation ranking mean greater contributions when applying Partnership Funding payment for outcomes methodology. Figure 20 illustrates the levels of social deprivation for Filey. Figure 20 - Social Deprivation Data from Office of National Statistics Super Output Areas 16

17 It is important to remember that total flood risk will never be removed completely, but Figure 21 shows how levels of risk can change with the introduction of flood protection measures. For example, the properties coloured orange have all moved from a very significant level or risk to a low level of risk. The more properties with reduced risk following a scheme the better chance for scheme justification under FDGiA. Very Significant to Moderate risk Very Significant to Low risk Significant to Moderate risk Figure 21 Visualising movement between risk zones improves outcome measure score under partnership funding. Flood Hazard Research Centre,