WRMP19 Water Resource Zone Problem Characterisation for Selection of Appropriate Planning Horizon and Assessment Method Statement

Transcription

1 WRMP19 Water Resource Zone Problem Characterisation for Selection of Appropriate Planning Horizon and Assessment Method Statement Overview Water resource zone (WRZ) problem characterisation is carried out to guide water resource planners toward the appropriate decision support tool method of assessment for the size and complexity of their planning problem. Analysis of the size and complexity of the planning problem also guides planners to the appropriate length of planning horizon, and therefore both assessment method and length of planning horizon are outcomes of the problem characterisation within this document. UKWIR WRMP 2019 Methods Decision Making Process: Guidance (2016) 1, hereafter referred to as the Guidance, provides a decision making framework for both defining the water resources planning problem and selecting the most appropriate assessment method to address it using the full array of feasible techniques. Defra s Guiding Principles for water resources planning 2 makes clear that while 25 years is the statutory minimum time horizon for water resources planning, there is strong encouragement to take a longer term view where it is appropriate, particularly in view of longer term pressures, uncertainties and the time it takes to develop some infrastructure. A longer horizon should be justified where appropriate. In light of the complex water resource planning problem in the South East England and the ongoing pressures associated with population growth and the impacts of climate change, Thames Water (TW) commissioned NERA to develop a framework for assessing the most appropriate time horizon for water resource planning 3. NERA were part of the expert team who developed the UKWIR 2016 Guidance. There are four sections within this methodology statement: 1. Characterising the water resources planning problem in TW s supply area 2. Appropriate time horizon over which to address the planning problem in TW s supply area 3. Appropriate decision support methods to assess TW s planning problem over the selected time horizon 4. Problem characterisation in the wider context of South East England 1 UKWIR (2016) WRMP 2019 Methods Decision Making Process: Guidance Report Ref. No. 16/WR/02/10 2 Department of Environment, Food and Rural Affairs (Defra) Guiding principles for water resources planning May NERA How Should the Appropriate Horizon for Integrated Water Resource Planning be Ascertained Thames Water September 2016

2 1. Characterising the water resources planning problem for TW s supply area Following the UKWIR guidance (2016), problem characterisation has been carried out for each of TW s water resource zones. A Water Resource Zone (WRZ) is defined as the largest possible zone in which all resources, including external transfers, can be shared and, hence, the zone in which all customers will experience the same risk of supply failure from a resource shortfall (WRPG 2014). As such, a WRZ is the appropriate geographical area within which problem characterisation should be carried out. TW has six defined WRZs: London, Guildford, Henley, Kennet Valley, Slough Wycombe Aylesbury (SWA) and Swindon and Oxfordshire (SWOX), shown in Figure 1. Figure 1 Water Resource Zones in South East England TW has a number of existing water transfers with neighbouring water companies and within its WRZs (Table 1). The majority of the transfers are historic, in perpetuity agreements. Most are relatively small and not large enough to affect the integrity of TW s WRZs. 2

3 Table 1 Existing Transfers to and from Thames Water WRZs (2016) Existing Transfer Capacity Current Agreement Ml/d From (WRZ) To (WRZ) Type Capacity Ml/d Duration Normal Year Dry Year SWA, TW SWOX, TW Raw 5 Available 2.08 LON, TW CEN, AW Treated 27 Perpetual 12.1 LON, TW CEN, AW Raw 10 Perpetual 2.00 ESW LON, TW Raw Until LON, TW ESW Raw Perpetual GUI, TW SOUTH, AW Treated Perpetual 2.3 The planning problem characterisation for each of TW s WRZ can therefore be carried out. This exercise has been undertaken following the methodology set out in the UKWIR Guidance (2016). Essentially the planner works through a series of questions relating to the size and complexity of the planning problem in each WRZ. The questions and associated scoring are set out in Tables 2 (How big is the problem?) to 5 (How difficult is it to solve?) with a summary of the results of the analysis for TW s individual WRZs presented. Table 2 Strategic Risk Water Resource Zone Customer Service could be significantly affected by current or future supply side risks, without investment How big is the problem? Strategic WRMP Risks (Score 0-2 each) Customer Service could be significantly affected by current or future demand side risks, without investment Investment programme likely to be unacceptably costly or contain contentious options Strategic Risk Score London SWOX SWA Kennet Guildford Henley

4 Table 3 Supply Complexity Water Resource Zone How complex is it to solve? (1) Supply Side Complexity (Score 0-2 each) Concerns about near/ medium term Concerns about Concerns about step changes to near term supply? future supply supply (Reliable/ resilient (climate change/ (sustainability to drought) water quality) reductions) Concern DO may fail to represent resilience Supply Complexity Score London SWOX SWA Kennet Guildford Henley Table 4 Demand Complexity Water Resource Zone Changes in current or near-term demand? How Complex is it to solve? (2) Demand Side Complexity (Score 0-2 each) Forecast uncertainty? Demand versus critical drought timing critical? Demand Complexity Score London SWOX SWA Kennet Guildford Henley Table 5 Investment Complexity Water Resource Zone Does uncertainty around capital expenditure affect the investment decision? How complex is it to solve? (3) Investment Programme Complexity (Score 0-2 each) Do factors such as lead time and promotability affect the decision? Can wider nonmonetisable considerations be properly considered? Is the investment programme sensitive to assumptions about the utilisation of new resources? Investment Complexity Score London SWOX SWA Kennet Guildford Henley

5 The scores presented in Tables 2-5 have been combined into the problem characterisation heat map, as advised in the Guidance, to give an indication of the complexity per WRZ. This is presented in Table 6. Table 6 Thames Water Problem Characterisation results by WRZ Problem Characterisation Strategic risk score (Table 2) Complexity factors score (Tables 3,4 & 5) Low <7 Medium 7-11 Henley Kennet Guildford SWA High (11+) SWOX London This analysis demonstrates that London and SWOX zones have a water resource planning problem of high, significant concern, while the rest of Thames Water s region has planning problems of low concern. 2. Determination of the appropriate time horizon over which to address the planning problem in TW s supply area The time horizon should be chosen so that events beyond the horizon end would be unlikely to affect the decisions about what to do initially (NERA, 2016). 4 Where there is no relevant deficit, or sufficient robust, low-cost options which can be quickly implemented, then the statutory minimum 25 year planning horizon is considered to be sufficient. However, where there is a large potential deficit, and options have long lead times and long asset lives, extension of the planning horizon may be necessary to ensure equitable, comparable assessment; this need must be weighed against the decreased reliability of forecasts over a longer time horizon. One of the key limiting factors for extension of planning horizon is the impact of the discount rate on investment in the distant future. 4 NERA How Should the Appropriate Horizon for Integrated Water Resource Planning be Ascertained? Thames Water 5 September 2016

6 NERA (2016) state that events beyond horizons of 100 years are most unlikely to influence the initial steps, and therefore a planning horizon beyond this limit is unlikely to be justified 5. In order to assess the correct planning horizon for a complex problem, NERA advocate use of a stepwise approach for extending the 25-year planning horizon in 5 year timesteps, by a flow chart of questions. The flow chart and associated questions are shown in Figure 2. Figure 2 Framework for horizon setting (NERA 2016) The questions can be translated into a score and a worked example for the London WRZ is given in Table 7 relating to the minimum statutory 25 year planning period and its potential extension to 30 years. 5 NERA 2016 Executive Summary 6

7 Table 7 scoring the potential for expanding the London WRZ planning horizon from 25 to 30 years Question n/a Yes No Increase horizon by 5 years Are discounted net costs within the 5 year extension substantial? (Yes) Are asset lives considerably longer than the current horizon? (Yes) Can the options to close the supply demand imbalance be considered complex options requiring detailed evaluation? (Yes) Is there strong concern about events beyond the current horizon? (Yes) Are forecast scenarios for a 5 year extension available? (Yes) Total score 0 (Yes) The score indicates whether the current planning horizon is appropriate, with 4 or above signifying yes, whereas a score of 0-2 indicates that extension by five years is both feasible and necessary. A score of 3 indicates an extension could be considered if it is considered to enhance evaluation of the planning problem. The requirement and feasibility of extension is then assessed for the following five years, and iterated until further extension becomes unnecessary or infeasible. The reasoning for the scoring has been worked through in 5 year timesteps for each WRZ in Appendix A, and summarised in Table 8 from 25 to 100 years. The appropriate horizon can be selected from the range of years showing the highest score for each zone. Table 8 Summary of scoring for extending planning horizon Water Resource Zone Is the current planning horizon appropriate? Potential Planning Horizon (years) Appropriate Horizon London SWOX SWA Kennet Guildford Henley The analysis demonstrates that a planning horizon of between 65 and 80 years would be most appropriate for London. SWOX is also likely to benefit from a similar planning horizon, mainly due to the nature of the strategic resource options available which can be shared between SWOX and London WRZs. The remaining TW zones have no need for planning horizon extension beyond 25 years given the size of the forecast supply demand imbalance is small in relation to availability of resource options. 7

8 As the baseline supply-demand gap is currently forecast to continue to expand beyond 100 years, the limiting factor on selecting the planning horizon for London and SWOX is uncertainty. As such it is appropriate to select the longest time horizon within the appropriate range, and an 80 year horizon has been adopted for both London and SWOX. An 80 year planning horizon aligns with that chosen by the Environment Agency when deriving its strategy of flood protection for London. The economic and social consequences of water supply failure in London would be equally as catastrophic as those associated with flood inundation and as such it is appropriate to work to a similar planning horizon when deriving the strategy for future water supply. 3. Appropriate methods to assess TW s planning problem over the selected horizon The problem characterisation in Table 6 demonstrates that both London and SWOX WRZs have a problem of high, significant concern. The Guidance therefore recommends the use of extended or complex risk-based techniques for thorough analysis of the planning problem, as can be seen in Figure, which is colour coded to match the problem characterisation matrix (Table 6). Figure 3 Decision making methods and tools for problems of different complexity 6 6 UKWIR (2016) WRMP 2019 Methods Decision Making Process: Guidance, Figure 3, page 30 8

9 The other WRZs have low complexity problems and analysis of Guildford, Henley, Kennet Valley and SWA WRZs can therefore be carried out over a 25 year period using the current EBSD approach where the NPV of the investment programme is optimised. In order to ensure that a best value investment programme is identified for these zones TW has enhanced its WRMP14 EBSD model. The EBSD+ model is a step-wise regression approach which includes best value optimisation of the investment programme against a series of metrics including initially cost, and then subsequently environmental performance, resilience, deliverability, sustainability and acceptability. However, assessment of the planning problem for London and SWOX requires the use of methods to better assess the complexity of potential solutions and increasing uncertainty of forecasts, especially over the extended planning horizon. Defining the problem The problem characterisation analysis summarised in Table 6 demonstrates that both the London and SWOX WRZs have large, highly complex supply demand problems, driven by the existing intensity of water resource utilisation in these areas, which is enhanced by ongoing population growth and decreasing water resource availability driven by climate change and environmental requirements. The UKWIR 2016 Guidance indicates that for complex problems, with complex potential solutions, and a high degree of uncertainty within the future forecasts, it will be beneficial to look beyond the most likely design scenario (Dry Year Annual Average or Critical Period), and sample within the uncertainty range for that scenario, to develop a range of additional risk scenarios to be specifically tested, rather than relying on headroom allowance being sufficient. Given the wide variance associated with long-term forecasts of population, climate change and environmental requirements, this approach is a robust way of testing the resilience of investment portfolios against specific, plausible future scenarios. The risk of a selected new option not performing as expected can also be included in such scenario modelling. The Info-gap, Real Options analysis, Adaptive Pathways and Multi-criteria Search with scheduling approaches are all decision support tools that can be used to define a multitude of scenarios for testing. Finding a potential solution Solutions can be developed for any scenario using aggregated methods, where distribution input is weighed against deployable output and a value for both is pre-determined for the scenario. Alternatively system simulation methods can be used, where either supply, demand or both may be simulated within a model under the influence of boundary conditions such as rainfall, temperature, water efficiency media campaigns or tariff changes. Evaluating a potential solution Robust decision making (RDM) can be used to develop and quantify the best parameters for solution of a complex problem, such as cost, environmental impact, resilience and customer acceptability. There is a wide range of uncertainties and potential future scenarios when planning strategic resources over a long time horizon. The complex planning problem in both the SWOX and London WRZs can be described as non-linear given the conjunctive nature of the water resource system and robust decision making combined with system simulation is considered more appropriate method for analysis of non-linear problems than application of aggregated methods as it provides a robust 9

10 approach of demonstrating best value against many uncertain futures. The strengths and weaknesses of the different decision support tools and the circumstances when their use is considered most appropriate for application is set out in detail in UKWIR s 2016 Guidance report. Finding the best solution Both aggregated and system simulation methods can be coupled with an optimiser to find portfolios of solutions to address a water resources problem, from single-parameter optimal solutions such as least cost, through modelling to generate alternatives (MGA) which finds near-optimal solutions for a single parameter, through to full optimisation techniques which seek trade-off solutions to best satisfy multiple parameters simultaneously, multi-criteria search (MCS). Method Selection TW is developing a range of scenarios to fully sample the complex planning problem in London and SWOX, and plans to use both aggregated and system simulation methods to develop potential solutions which will be evaluated using a range of performance metrics. The metrics are cost, environmental performance, resilience, sustainability, adaptability, deliverability and acceptability. Both aggregated and system simulation models are coupled to optimisers which use MGA or MCS to find a range of potential solutions. The output of models and support tools will provide information to aid and support decision making. They will not in themselves provide the decision. TW, with engagement of expert advisors, will consider and evaluate the output from the decision support tools to determine the preferred investment strategy. 4. Problem characterisation in the wider context of South East England In the context of Water Resources South East (WRSE), problem characterisation for TW identifies that the London WRZ has a high strategic need and high complexity affecting identification of the best solution. When London is considered in conjunction with neighbouring water company zones, the problem expands beyond the immediate TW WRZ boundaries. Existing, in perpetuity, water transfers exist from London to Essex and Suffolk Water (91 Ml/d in normal years, temporarily reduced to 71 Ml/d in drought years until April 2035) and Affinity Water (14.1 Ml/d). Further new transfers are proposed and are being assessed as part of the strategic water resource planning by the WRSE group of six water companies: Affinity Water, Portsmouth Water, South East Water, Southern Water, Sutton and East Surrey Water, and Thames Water (Table 9). 10

11 Table 9 Potential additional transfers between WRSE water companies Potential Additional Transfers with TW Capacity (Ml/d) From (WRZ) To (WRZ) Company Import/Export Duration? Maximum Minimum LON, TW Central Affinity Export LON, TW Sutton Sutton and East Surrey Export 30 5 LON, TW Kent Southern Water Export SWA, TW RZ4 South East Water Export 25 5 GUI, TW South Affinity Water Export RZ2, SEW GUI TW Import 15 1 The majority of the proposed transfers affecting TW are exports, with a potential additional 175 Ml/d demand from London should they be identified as part of the strategic need for the South East. The significance of the problem within the London WRZ could therefore be extrapolated to the recipient transfer zones (Figure ) were the transfer volume significant. Figure 4 Thames WRZs planning problem significance and potential transfers to neighbouring zones from London 11

12 Figure demonstrates that while problem characterisation should be carried out at WRZ level, the problems apparent in one zone may well be transferred to another and as such, selected planning methods for connected or potentially connected neighbouring zones should ideally be as closely aligned as possible to best analyse transfer capabilities and shared resource planning where a problem is significant and widespread. This is the remit of the WRSE group, which also collaborates with Water Resources East (WRE) to tackle the water resource issue across the whole SE region. Specifically with respect to WRSE, the extended modelling approaches being implemented utilises the aggregated EBSD InfoGap methodology and the system simulation RDM. This is not unlike the approaches being utilised by TW in its WRMP19 programme appraisal, with use of both aggregated and system simulation decision support tools to evaluate a complex planning problem. Summary TW has a problem of significant concern and complexity in London and SWOX WRZs. For these zones, an 80 year planning horizon has been selected, both using the method advised by NERA to take into account problem continuity, asset life and investment horizons, and by consideration of planning horizons of neighbouring zones and potential water trading partners. Complex assessment methods are being developed for London and SWOX WRZs which reflect the significance and complexity of the planning problem. The London WRZ also needs to be considered in the context of Water Resources South East as investment solutions have potential to contribute to the wider water resources need identified in the South East. 12

13 Appendix A: Planning Horizon There are several reasons to consider extending the planning horizon for a zone, and several reasons why it may be unnecessary or infeasible. NERA has condensed these into a series of five questions: Question Scoring of answers Are discounted net costs within the 5 year extension substantial? Yes 0 No 1 Are asset lives considerably longer than the current horizon? Yes 0 No 1 Can the options to close the supply demand imbalance be considered complex options requiring detailed evaluation? Yes 0 No 1 Is there strong concern about events beyond the current horizon? Yes 0 No 1 Are forecast scenarios for a 5 year extension available? Yes 0 No 1 Discounted Net Costs Costs have been annuitized for all types of large water resource option as per the method developed by NERA, for immediate construction start. The percentage annuitized investment beyond a series of planning horizons is shown in Figure. Figure A1: Annuitized investment of large water resource options across different planning horizons 13

14 As such it can be seen that while discounted costs of large options beyond a 25 year planning horizon are significant and could affect investment, beyond 60 years they are small, and beyond 80 years insubstantial. Asset Life The asset lives of the large options considered by TW are: Option Type Capex Category Asset Life (years) Reuse plant Treatment works (civils) 60 Transfer pipeline Raw water transport: Tunnels & conduits 80 Reservoir Raw water abstraction: Storage reservoirs & lakes 250 Asset lives for all are considerably longer than a 25 year planning horizon. However, beyond 60 years the asset life for the majority of the options are not considerably longer, and past 80 years only the reservoir options would have significant further life. Sufficient low-cost rapid solutions Within the London and SWOX WRZs, there are insufficient simple low-cost options to meet the supply-demand gap over a 25-year planning horizon. For the remaining TW WRZs, the available small supply options from WRMP14 have been compared to the 25-year supply demand gap. WRZ DYCP Small Options (Ml/d) Baseline DYCP Supply- Demand gap increase over 25 years (Ml/d) Horizon for which small options sufficient SWA years Kennet years Guildford years Henley years While this does not take into consideration the significant increases in population forecast since the publication of WRMP14, or the effects of the planned demand management programmes, it does indicate the capability of rapidly available low-cost solutions to supply the Thames Valley WRZs. Concern for events beyond the current horizon Several UK water companies together with a number in the WRSE and WRE regions have stated the intention to plan for a longer period than 25 years, which indicates a high level of concern across the region. The Water UK long-term planning study national model also uses a 50 year time horizon to 2065, and the schematic in Figure shows the distribution of known planning horizons. 14

15 Figure A2 Horizon end points for different water resources plans Southern Water is planning for a 50 year horizon and it is likely that several other individual companies which form part of the WRSE group may also do so. WRSE itself plans to assess the problem to 2080, a sixty year horizon, while WRE intends to develop its model to 2100 (eighty year horizon), as is Anglian Water. Generation of reasonable forecasts The forecasting of supply or demand into the future depends upon the reliability and extent of the base data forecasts which support them (population, housing, climate change, etc.). Population forecasting is being carried out by Professor Phil Rees of Leeds University, who considers that forecasting of water demand should at least be carried out for the life expectancy of someone born today, i.e. upward of 80 years. He also stated that at a national level, the ability to forecast population growth up to 100 years ahead has improved, although noted that at sub-regional levels net migration flows can introduce substantial uncertainty. Housing Growth is forecast by the Greater London Authority (GLA) to Climate forecasting was carried out for the 2015 Climate Change Risk Assessment, in which contributing factors for the levels of risk in water resource planning are projected to A feasible cut-off for forecasting supply and demand could be set at 2100 to mirror the availability of baseline population and climate data projections. Zonal Planning Horizon Extension The following pages show the results of assessing each zone by the above criteria. 15

16 Question Scoring of answers Are discounted net costs within the 5 year extension substantial? Yes= 0 No= 1 Are asset lives considerably longer than the current horizon? Yes= 0 No= 1 Can the options to close the supply demand imbalance be considered complex options requiring detailed evaluation? Yes= 0 No= 1 Is there strong concern about events beyond the current horizon? Yes= 0 No= 1 Are forecast scenarios for a 5 year extension available? Yes= 0 No= 1 London Comment Costs beyond Y Y Y Y Y Y Y Y N N N N N N N N Asset Life Y Y Y Y Y Y Y Y N N N N N N N N Complex solutions Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Concern beyond Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Forecasts Y Y Y Y Y Y Y Y Y Y Y Y N N N N Score SWOX Comment Costs beyond Y Y Y Y Y Y Y Y N N N N N N N N Asset Life Y Y Y Y Y Y Y Y N N N N N N N N Complex solutions Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Concern beyond Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Forecasts Y Y Y Y Y Y Y Y Y Y Y Y N N N N Score SWA Comment Costs beyond N N N N N N N N N N N N N N N N No large options Asset Life N N N N N N N N N N N N N N N N No large options Complex solutions N N N N N N N N N N N N N N N N Concern beyond N N N N N N N N N N N N N N N N Forecasts Y Y Y Y Y Y Y Y Y Y Y Y N N N N

17 Score Kennet Valley Comment Costs beyond N N N N N N N N N N N N N N N N No large options Asset Life N N N N N N N N N N N N N N N N No large options Complex solutions N N N N N N N N N N N N N N N N Concern beyond N N N N N N N N N N N N N N N N Forecasts Y Y Y Y Y Y Y Y Y Y Y Y N N N N Score Guildford Comment Costs beyond N N N N N N N N N N N N N N N N No large options Asset Life N N N N N N N N N N N N N N N N No large options Complex solutions N N N N N N N N N N N N N N N N Concern beyond N N N N N N N N N N N N N N N N Forecasts Y Y Y Y Y Y Y Y Y Y Y Y N N N N Score Henley Comment Costs beyond N N N N N N N N N N N N N N N N No large options Asset Life N N N N N N N N N N N N N N N N No large options Rapid solutions N N N N N N N N N N N N N N N N Concern beyond N N N N N N N N N N N N N N N N Forecasts Y Y Y Y Y Y Y Y Y Y Y Y N N N N Score This scoring is summarised in Table 7. 17