South West Water Water Resources Management Plan June 2014

SWW Water Resources Management Plan 2015 – 2040

SWW Water Resources Management Plan 2015 – 2040

Contents

Introduction

Summary of final Water Resources Management Plan Page no. 1 General information on plan content and development 1.1 Planning period 1.1 1.2 Water Resource Zones 1.1 1.2.1 Introduction 1.1 1.2.2 Colliford WRZ 1.3 1.2.3 Roadford WRZ 1.5 1.2.4 Wimbleball WRZ 1.7 1.3 Planning scenario 1.9 1.4 Links to other plans, Government Policy and aspirations 1.9 1.4.1 Ofwat Business Plan 1.9 1.4.2 Strategic Environmental Assessment and Habitats Regulations 1.10 1.4.3 Government Policy and aspirations 1.10 1.5 National security and commercial confidentiality 1.10 1.6 Level of service 1.11 1.7 Climate change 1.12

2 Water supply 2.1 Deployable output 2.1 2.1.1 General 2.1 2.1.2 DO / WAFU: Colliford WRZ 2.3 2.1.3 DO / WAFU: Roadford WRZ 2.3 2.1.4 DO / WAFU: Wimbleball WRZ 2.4 2.2 Confidence rating for our source DOs 2.5 2.3 Level of service and DO assessment 2.6 2.4 Reductions in DO – sustainability changes 2.7 2.4.1 Changes identified in WRP09 2.7 2.4.2 Licence renewals due between 2015 and 2020 2.8 2.4.3 Changes identified for this Plan 2.8 2.5 Abstraction Incentive Mechanism 2.13

SWW Water Resources Management Plan 2015 – 2040

2.6 Reductions in DO – Outage 2.13 2.6.1 General 2.13 2.6.2 Planned outage 2.13 2.6.3 Unplanned outage 2.14 2.6.4 Specific borehole-related outages 2.14 2.6.5 Other potential outages 2.14 2.6.6 Total outage allowance for each WRZ 2.14 2.7 Raw and potable water transfers and bulk supplies 2.15 2.8 Distribution and treatment works operational use and losses 2.15 2.9 Greenhouse Gas Emissions 2.16

3 Water demand 3.1 Introduction 3.1 3.1.1 Scenarios modelled 3.1 3.1.2 Demand in the base year 3.1 3.1.3 Metering policy 3.3 3.2 Demography 3.4 3.2.1 Our region 3.4 3.2.2 Demographic forecasts 3.4 3.2.3 Population 3.5 3.2.4 Housing 3.6 3.2.5 Average household size 3.8 3.3 Household consumption 3.9 3.3.1 Our approach to forecasting household consumption 3.9 3.3.2 Historic PCC 3.10 3.3.3 Normalising base year demand 3.12 3.3.4 Dry year demand in the base year 3.15 3.3.5 Summary of base year PCC adjustment factors 3.16 3.3.6 PCC forecasts 3.17 3.3.7 The effect of metering on household demand 3.20 3.3.8 The effect of climate change on household demand 3.21 3.3.9 Household consumption forecast 3.21 3.3.10 Forecasting PCC in the future 3.22 3.4 Non-household consumption 3.23 3.4.1 Our approach to forecasting non-household consumption 3.23 3.4.2 Historic consumption 3.24

SWW Water Resources Management Plan 2015 – 2040

3.4.3 Base year demand in normal and dry year scenarios 3.24 3.4.4 Forecasts of measured non-household demand 3.24 3.4.5 Forecasts of unmeasured non-household demand 3.24 3.4.6 The effect of climate change on non-household demand 3.25 3.4.7 Overall non-household consumption forecast 3.25 3.4.8 Measured non-household consumption by industry sector 3.26 3.5 Leakage 3.27 3.5.1 Determining base year leakage 3.27 3.5.2 Determining the sustainable economic level of leakage 3.28 3.5.3 Leakage forecast 3.30 3.5.4 Meeting our leakage target 3.31 3.6 Other components of demand 3.33 3.6.1 Water taken unbilled 3.33 3.6.2 Distribution system operational use 3.33 3.6.3 Overall forecast of other components 3.34 3.7 Total demand 3.34 3.7.1 Summary of forecast demand 3.34 3.7.2 Water efficiency activity 3.35 3.7.3 Profile of annual demand 3.36 3.8 Derivation of weighted average demand forecasts 3.37

4 Climate change 4.1 General 4.1 4.2 Supply 4.1 4.2.1 Climate change vulnerability 4.1 4.2.2 Assessment of the impacts of climate change on river flows 4.1 4.2.3 Assessment of the impacts of climate change on groundwater 4.2 resources 4.2.4 Assessment of the impacts of climate change on DO 4.2 4.3 Demand 4.3 4.4 Impact on supply demand balance 4.3

5 Target headroom 5.1 Method 5.1 5.1.1 Target headroom 5.1 5.1.2 Calculation of target headroom 5.1

SWW Water Resources Management Plan 2015 – 2040

5.1.3 Available headroom 5.2 5.1.4 Security of Supply Index 5.2 5.2 Target headroom 5.2 5.2.1 Approach to target headroom determination 5.2 5.2.2 Reduction in uncertainty since WRP09 5.3 5.2.3 Headroom components relevant to the current Plan 5.3 5.2.4 Target headroom and sustainability reductions 5.3 5.2.5 Target headroom and the level of uncertainty 5.3 5.2.6 Target headroom and the impact of climate change 5.4

6 Water resources strategy 6.1 Baseline supply demand balance 6.1 6.1.1 Colliford WRZ 6.1 6.1.2 Roadford WRZ 6.1 6.1.3 Wimbleball WRZ 6.2 6.2 Options appraisal 6.3 6.3 Final supply demand balance 6.4 6.3.1 Actual level of service 6.4 6.3.2 Impact of level of service on deployable output 6.4 6.3.3 Greenhouse Gas Emissions 6.4 Scenario testing 6.4 6.4.1 Sensitivity of final Plan to headroom risk percentile 6.5 6.4.2 Sensitivity of final Plan to changes in demand 6.7

7 Choices 7.1 Introduction 7.1 7.2 Government policy priorities and expectations 7.1 7.3 The long-term perspective 7.2 7.3.1 An extended planning horizon 7.2 7.3.2 Long-term resilience and flexibility 7.4 7.4 Water scarcity and environmental damage 7.5 7.4.1 Environmental liaison 7.5 7.4.2 Catchment management 7.6 7.5 Further interconnection and water trading 7.8 7.5.1 Conjunctive use and interconnection 7.8 7.5.2 Water trading 7.9

SWW Water Resources Management Plan 2015 – 2040

7.6 Reducing the demand for water 7.13 7.6.1 Water metering 7.13 7.6.2 Per capita consumption 7.13 7.6.3 Leakage 7.14 7.7 The future 7.16

8 Glossary of terms used in the WRMP

Appendices Appendix A Headroom uncertainty calculations Appendix B Demand profiles Appendix C Outage determination Appendix D Micro-component per capita consumption forecasts Appendix E Demographic forecasts Appendix F Climate change Appendix G Summary of strategic environmental assessment Appendix H Water resource zones Appendix I Consultation Appendix J Business Planning and Price Review Process

SWW Water Resources Management Plan 2015 – 2040

Introduction

In this Plan, we set out our strategy to ensure that all customers have a secure supply of water through until the year 2039/40. We submitted our draft Water Resources Management Plan to Defra in March 2013. There followed a period of public consultation which ended in August 2013. A total of six organisations sent comments on the plan to Defra. We then produced our Statement of Response which showed how we had modified our Plan, where appropriate, to take into account the comments and any updated information. We also provided further information to Defra in January 2014. These were examined by Defra which, in May 2014, gave us permission to go ahead and publish this Final WRMP. Developing the strategy involves estimating the future demand for water, calculating how much water is available from current sources and then looking at the need for options to either reduce demand or increase supply. In planning so far ahead, we have to make many assumptions and inevitably there will be a lot of uncertainty surrounding these. For example, how much will the population of the region change? What will be the impact of climate change? We have dealt with this uncertainty by including some headroom between supply and demand and also by ensuring that the strategy is sufficiently resilient to cope with changing circumstances.

Our customers’ views have been of paramount importance to us in the development of our water resources strategy. Therefore, over the last two years, we have carried out an extensive programme of research and engagement activity with our customers covering all areas of the Company’s business. This research covers issues such as water availability, metering, leakage, catchment management etc and is fully described in a document which provides supporting information to our Business Plan 2015-20. The document was published on 2 December 2013 and can be viewed using the following link http://www.southwestwater.co.uk/businessplan2015-20

We have not had a hosepipe ban in our region since 1996 and we are continuing to plan on the basis that demand restrictions will only be imposed in exceptional circumstances.

We are very keen to ensure that our strategy is sustainable and we are therefore continuing to promote the efficient use of water to our customers, leakage control and the optimum use of existing sources in order to continue to defer further resource development.

We believe that the strategy described in this Plan is robust, sustainable and meets the needs of our customers and the region.

SWW Water Resources Management Plan 2015 – 2040

Summary of Water Resources Management Plan (WRMP)

1 This Plan presents our supply demand projections to 2039/40 which we have completed in accordance with the Environment Agency guidelines1. The position of this Plan within the overall planning and price review process can be seen in the Document Plan structure shown in Appendix J of our WRMP.

2 Our customers have told us that they do not wish to pay more for a higher level of service, but neither do they want the existing level of service reduced. Therefore we do not propose to change the existing level of service.

3 Over the last two years, we have carried out an extensive programme of research and engagement activity with our customers to ascertain their views on matters covered by this WRMP.

4 Our demand projections have been based upon the use of a micro-component analysis approach to forecast per capita consumption for the domestic customer base.

5 We have taken climate change into account in producing our demand projections.

6 Our demand projections are based on our current metering policy and evidence of the effect of metering on water use.

7 Our demand projections also take into account our routine water efficiency work that we carry out that enables us to continue to meet the Ofwat water efficiency targets.

8 The supply appraisal follows Environment Agency guidelines2, and includes a full assessment of Water Available For Use (WAFU) in each of our Water Resource Zones (WRZs).

9 The calculation of WAFU has taken climate change into account in accordance with the Environment Agency Guidelines.

10 The target headroom between supply and demand has been based on the recommended methodology of headroom uncertainty3.

11 Our supply demand balance assumes that we will continue to keep leakage at or below the agreed target of 84 Ml/d until 2019/20, after which we will reduce leakage to 64 Ml/d over the planning period. Our current target of 84 Ml/d remains below the Sustainable Economic Level of Leakage.

12 Our Plan shows a reduction in distribution input over the planning period in all three WRZs.

13 Our domestic per capita consumption is already well below the national average and is expected to reach the Government’s aspirational level of 130 litres/person/day by 2016.

1 Environment Agency, “Water resources planning guideline – interim update”, October 2012 2 Environment Agency, “Reassessment of Water Company Yields”, February 1997 3 UKWIR, “An Improved Method for Assessing Headroom”, Report Ref No 02/WR/13/2, 2002

SWW Water Resources Management Plan 2015 – 2040

14 Only one additional sustainability option which has an impact on WAFU has been identified by the Environment Agency. The impact of this option is to reduce WAFU by approximately 2 Ml/d in the Roadford WRZ.

15 Our final supply demand balance shows that a surplus is available in all three WRZs throughout the planning period covered by this Plan. There is therefore no need for a new reservoir or other additional resource development in the region during the period covered by this Plan.

16 We have considered the long-term perspective and believe that this surplus exists in all WRZs until at least 2050.

17 One option to use this surplus is to make it available to other water users and we have therefore identified locations where water is available and stated the amounts available and the costs.

18 We are very keen to “do the right thing” by helping to achieve government policy objectives. After 2019/20, we are therefore proposing to reduce leakage from its current level of 84 Ml/d to 64 Ml/d over the planning period. This reduction is driven by customer preference and Government aspirations.

19 We are committed to continuing to optimise our use of water resources through their conjunctive use and intend to further develop links between WRZs. We are also undertaking a number of projects to increase the resilience of public water supply.

19 In this Plan we describe our major catchment management programme which will deliver benefits for both customers and the environment.

SWW Water Resources Management Plan 2015 – 2040

1 General information on plan content and development

1.1 Planning period

Our Water Resources Management Plan (WRMP) covers the period up until 2039/40 and has a base year of 2012/13. The 2012/13 figures shown in the associated tables are based on outturn data from the Company Annual Performance Review 2013 (formerly June Return).

1.2 Water Resource Zones

1.2.1 Introduction

Water resource planning is based on Water Resource Zones (WRZs) which are defined as:

“the largest possible zone in which all resources, including external transfers, can be shared and hence the zone in which all customers experience the same risk of supply failure from a resource shortfall”1

We defined our WRZs in accordance with the Water resources planning guideline2. We use three WRZs – Colliford, Roadford and Wimbleball - for planning and managing our water resources. The three WRZs are aggregations of Water Into Supply (WIS) zones and are shown in Figure 1.1.

Figure 1.1: South West Water WRZs and WIS Zones

Wimbleball WRZ

Colliford WRZ

Roadford WRZ

1 UKWIR/Environment Agency, “Definition of Key Terms for Water Resources Practitioners”, 1997 2 Environment Agency, “Water resources planning guideline”, August 2013

Page 1.1

SWW Water Resources Management Plan 2015 – 2040

At an early stage in the WRMP process we shared with the Environment Agency information about our WRZs; a copy of our report to them is shown in Appendix H. Appendix H also includes confirmation from the Environment Agency that our approach is appropriate.

Our WRZs have not changed since the publication of our last Water Resources Plan3. We currently have no plans to review the definition of our WRZs.

All of our WRZs are conjunctive use systems as defined in WR274.

A complete list of our sources within each WRZ is given in the WRP1a tables.

Sections 1.2.2 to 1.2.4 below give a brief outline of Colliford, Roadford and Wimbleball WRZs. Appendix H provides more details of our WRZs, including information about imports and exports between them.

3 South West Water, “Water Resources Plan 2010-2035”, November 2009 4 UKWIR, Project WR27 “Water Resources Planning Tools 2012”, 2012

Page 1.2

SWW Water Resources Management Plan 2015 – 2040

1.2.2 Colliford WRZ

The Colliford WRZ covers most of Cornwall except the north east of the County.

We use Colliford Reservoir conjunctively with local reservoirs, two disused former china clay pits and river intakes to form Colliford WRZ. These sources are supplemented by a bulk transfer from Roadford WRZ of up to the order of 3 Ml/d. The storage of Colliford Reservoir can also be supplemented by pumped transfers from Restormel.

Colliford Reservoir is both a river regulation and a direct supply reservoir and supports supplies in three ways:

. releases to the River Fowey for abstraction and treatment at Restormel Water Treatment Works (WTW) . pumping water direct to De Lank and Lowermoor WTWs . supplying water, via a gravity pipeline, direct to St Cleer WTW.

A schematic of the key components is shown in Figure 1.2 below. Figures 1.3 and 1.4 show Colliford and Siblyback Reservoirs respectively.

Figure 1.2: Key components of Colliford WRZ

N

Crowdy Launceston Stannon

Colliford Park Siblyback

Bodmin Newquay Porth

Truro

Stithians Key Water Treatment Works Penzance Reservoir College Town/City Argal Falmouth Drift River Indicative pipeline Pumped Storage

Not to scale

Page 1.3

SWW Water Resources Management Plan 2015 – 2040

Figure 1.3: Colliford Reservoir

Figure 1.4: Siblyback Reservoir

Page 1.4

SWW Water Resources Management Plan 2015 – 2040

1.2.3 Roadford WRZ

The Roadford WRZ covers a large part of Devon, from Plymouth, the South Hams and Torbay in the south to Bideford and Barnstaple in the north. It also includes parts of north east Cornwall. The area is served primarily by Roadford Reservoir operating conjunctively with other impounding reservoirs, river intakes and other sources.

The most important single source in the area is Roadford Reservoir on the River Wolf, a tributary of the River Tamar. We use Roadford to augment the River Tamar for abstraction downstream at Gunnislake and also for direct supply to parts of North Devon (via Northcombe WTW).

A schematic of the key components is shown in Figure 1.5 below. Figures 1.6, 1.7 and 1.8 show Roadford, Burrator and Meldon Reservoirs respectively.

Figure 1.5: Key components of Roadford WRZ

Key Water Treatment Works N Reservoir Ilfracombe Town/City Wistlandpound River Indicative pipeline

Barnstaple Bideford

Torrington

Upper Tamar Lake Tiverton

Bude

Okehampton Exeter Roadford Meldon

Fernworthy R.Lyd KTT Launceston Exmouth Newton Venford Burrator Abbot

Avon

Torbay Plymouth

Not to scale

Page 1.5

SWW Water Resources Management Plan 2015 – 2040

Figure 1.6: Roadford Reservoir

Figure 1.7: Burrator Reservoir

Page 1.6

SWW Water Resources Management Plan 2015 – 2040

Figure 1.8: Meldon Reservoir

1.2.4 Wimbleball WRZ

Wimbleball Reservoir was constructed by South West Water Authority, the predecessor organisation of South West Water, with part of the financing costs being paid by Wessex Water Authority (WWA). We use the reservoir principally for making augmentation releases to the River Exe for subsequent abstraction near Tiverton and Exeter. These releases support abstractions from the natural flow of the River Exe. Wessex Water uses the reservoir for direct supply.

The Wimbleball WRZ is also dependent on the significant groundwater resources of East Devon.

A schematic of the key components is shown in Figure 1.9 below. Figure 1.10 shows Wimbleball Reservoir.

Page 1.7

SWW Water Resources Management Plan 2015 – 2040

Figure 1.9: Key components of Wimbleball WRZ

Key N Water Treatment Works

Reservoir Wimbleball To Wessex Town/City Water Otter Valley groundwater sources

River

Indicative pipeline

Pumped storage

Tiverton

Crediton Honiton Exeter Lyme Regis

Sidmouth

Exmouth Not to scale

Page 1.8

SWW Water Resources Management Plan 2015 – 2040

Figure 1.10: Wimbleball Reservoir

1.3 Planning scenario

We have produced forecasts for both supply and demand that could occur in a dry year. We also consider and calculate supply and demand in more normal conditions but the Water resources planning guideline5 does not require us to publish them in this report. However we have used these normal year forecasts in our calculation of weighted average demand which is used by Ofwat in its review of price limits.

More detailed information on our demand forecasts is given in Section 3.

None of our three WRZs is dependent only on groundwater, run of river abstraction or limited storage, nor are they particularly sensitive to peak demands, but we do carry out detailed modelling of the water resources systems, which implicitly consider these peaks. Therefore we do not report separately on the „critical period‟.

1.4 Links to other plans, Government Policy and aspirations

Our final WRMP is not produced in isolation but is influenced by and linked to a number of different plans. These are summarised in Sections 1.4.1 to 1.4.3.

1.4.1 Ofwat Business Plan

Ofwat will use our WRMP as a basis when assessing our investment and business needs.

5 Ibid. 2

Page 1.9

SWW Water Resources Management Plan 2015 – 2040

1.4.2 Strategic Environmental Assessment and Habitats Regulations

Government expects a water company to produce a WRMP that is informed by a Strategic Environmental Assessment (SEA).

The fundamental aim of the SEA process is to provide for a high level of protection of the environment, and to contribute to the integration of environmental considerations into the preparation of plans and programmes with a view to promoting sustainable development.

We consider that the strategic nature of a WRMP and its role in setting the framework for future water resources development, and management in the region means that it should be subject to a SEA under the Environmental Assessment of Plans and Programmes Regulations 2004.

We commissioned Hyder Consulting (UK) Limited to undertake an SEA to inform our previous Water Resources Plan (WRP)6.

Within this final WRMP we have not made significant changes since our previous WRMP nor have we proposed any options, therefore there is no requirement for any specific further work in connection with an SEA as nothing has changed. A summary of the SEA undertaken in 2009 is given in Appendix G.

Regarding the Habitats Regulations, our current abstraction licences were fully reviewed as part of the Review of Consents work undertaken by the Environment Agency. We have included within Section 2 of our plan current information available regarding the proposed changes in our abstraction licences.

Within this final WRMP we are not proposing any further options and therefore there is no requirement for any further work in connection with the Habitats Regulations.

1.4.3 Government Policy and aspirations

Our Plan has taken account of government policy as set out in the guiding principles 7 for developing a water resources management plan . Further details are given in Section 7.

1.5 National security and commercial confidentiality

Nothing has been excluded from our Plan on the grounds of national security and commercial confidentiality

6 Ibid. 3 7 Environment Agency, Ofwat, Defra and the Welsh Government, “Water resources planning guideline: The guiding principles for developing a water resources management plan”, June 2012

Page 1.10

SWW Water Resources Management Plan 2015 – 2040

1.6 Level of service

Our policy is to try to avoid imposing demand restrictions, such as hosepipe bans8, or seeking drought orders and in this we have been successful for the last seventeen years. However, for the purposes of planning we use the following level of service:

(i) Customer level of service:

. a major publicity campaign requesting voluntary savings of water not more than once in every 10 years on average

. a hosepipe ban8 for water resources reasons not more than once in every 20 years on average, giving a 5% reduction in demands

. 6-month maximum duration of hosepipe ban8

. a ban on the non-essential use of water not more than once every 40 years on average, giving a further 5% reduction in demands

. 4-month maximum duration of ban on non-essential use of water

. the use of rota cuts or standpipes is regarded as unacceptable for water resource planning purposes

(ii) Environmental level of service:

. the use of drought orders or drought permits reducing compensation or prescribed flows not more than once every 20 years on average

Our customer research showed that there was no support for deterioration in level of service and no desire to pay for an enhanced level of service. Therefore we have applied the above level of service throughout the period covered by this plan, ie up to and including 2039/40.

We have no plans to change our current planned level of service.

Our level of service in this plan is consistent with those in our Drought Plan.

During periods of exceptionally dry weather, we continually monitor the situation and adopt a flexible approach to how our sources are operated to accommodate any changing or unusual circumstances. By adopting such an approach, we are able to maximise the level of service that can be delivered.

Further details of the relationship between the level of service and our assessment of Deployable Output (DO) are covered in Section 2. Information regarding how we reliably supply the Company planned level of service and our estimated actual level of service is given in Section 6.

8 Now encompassed in legislation regarding temporary water use restrictions

Page 1.11

SWW Water Resources Management Plan 2015 – 2040

1.7 Climate change

Our Plan includes an assessment of climate change and its effects on the supply demand balance. This is covered in Section 4.

Page 1.12

SWW Water Resources Management Plan 2015 – 2040

2 Water supply

2.1 Deployable output

2.1.1 General

We have assessed Deployable Output (DO) using our detailed knowledge of our water resources systems and our water resources modelling tools, which include a computer model called MISER1. The MISER model is an optimisation model and is based on a water balance of the whole of our area. The model is a key tool in analysing and planning water resources availability and is also used in operational decision making.

We have calculated DOs using historic recorded flow records post 1957 to 2011 for Wimbleball Water Resource Zone (WRZ) and Roadford WRZ, and post 1962 to 2011 for the Colliford WRZ. These are the earliest periods when reliable flow records are available. These flow records include a variety of serious droughts eg 1959, 1975/76, 1978, 1984, 1989 and 1995.

We have also worked with the Environment Agency on available rainfall records prior to this period as well as extended flow sequences previously derived by the Agency, to investigate if the dry conditions experienced within the period of the reliable flow record are representative of dry conditions experienced over a longer time period. The work indicated that the area does not seem to have experienced any droughts more significantly severe than those represented in the more recent flow records. The work has also concluded that using the current historic period of flow records is reasonably representative of any longer theoretical flow sequences that are available.

We have shared this technical information with the Environment Agency who have confirmed that, given the difficulties with the current availability of historic data in our area and our current supply demand balance situation, at present DO should continue to be based on the historic flow sequences.

Our current approach is also supported by previous work and approaches including:

ƒ Environment Agency report2 which refers to the use of flow sequences from the early part of the century and states ‘In some cases it may not be possible to extend river flow records with an appropriate level of accuracy’.

ƒ An internal working Environment Agency (SW) report (June 2007) entitled Extension of Naturalised Flow Sequences Methodology Review and Recommendations3 .

1 MISER is produced by Tynemarch 2 Environment Agency, ”Reassessment of Water Company Yields”, February 1997 3 Environment Agency (SW), “Extension of Naturalised Flow Sequences Methodology Review and Recommendations”, June 2007. Unpublished reference

Page 2.1

SWW Water Resources Management Plan 2015 – 2040

ƒ In 1991 we commissioned the Institute of Hydrology to derive synthetic flow sequences for the Colliford WRZ using all available flow and rainfall records. Of the 21 flow sequences that were needed only one could be extended back as far as 1939 by using rainfall records. None of the sequences could be extended back to include the drought sequence of the early 1930s.

Furthermore, at present, we consider that the use of long-term flow sequences of dubious quality could lead to very misleading conclusions.

In our calculation of DO, we include the following key constraints: hydrological yield, licensed quantities, key pumping equipment, well/aquifer properties, raw water main capacities, key treatment capacities and constraints and level of service.

The demand patterns used in our assessment of DO are detailed in Section 3 of this Plan.

Our assessment of DO includes an allowance for climate change as detailed in Section 4 of this Plan.

Any areas of uncertainty such as meter accuracy are included in our headroom analysis as detailed in Section 5.

There are no significant changes to our values for emergency storage from those used in our 2009 Water Resources Plan4 (WRP09), although the emergency storage for Colliford has slightly reduced from 3000 Ml. Our current assessment for the emergency storage values in the three largest reservoirs is given below:

ƒ Colliford - 2854 Ml ƒ Roadford - 5370 M ƒ Wimbleball - 2132 Ml

There are no significant changes in our DO figures from those in our WRP09, although there are slight differences in the DO and Water Available For Use (WAFU) figures presented in the accompanying tables as a result of the use of the updated climate change scenarios. Where flows have had to be naturalised or estimated, we have shared our methodology for calculating historic reservoir inflows and river flows with the Environment Agency. During the pre-consultation period, the Environment Agency has raised no concerns over this or our previous estimates of DO.

The DOs shown in the WRP1a tables do not take account of the various recognised losses within the systems, such as WTW losses. We have shown these separately in the tables and taken account of them within our calculation of WAFU.

4 South West Water, “Water Resources Plan 2010-2035”, November 2009

Page 2.2

SWW Water Resources Management Plan 2015 – 2040

2.1.2 DO / WAFU: Colliford WRZ

2.1.2.1 Exports and imports

Our water resources model incorporates the very small export from Colliford WRZ to Roadford WRZ in the Bude area and the import from Roadford WRZ to Colliford WRZ in the Saltash area.

2.1.2.2 Derivation of the Colliford pumping curve

We have produced a Colliford Pumping Curve to maximise pumping to Colliford and other transfers into Colliford within the constraints of the licences, which ensures that the reservoir is as full as possible before the start of the summer drawdown.

2.1.2.3 Colliford fisheries bank

We have made an allowance in these calculations for releases from the Colliford Fisheries Bank in accordance with the provisions of the Colliford and Siblyback Reservoirs Operating Agreement.

2.1.2.4 Critical event(s) that define DO

We used our water resources model to simulate the system through the complete record of flow sequences. We found that similar severe drawdowns occurred in Colliford Reservoir in several years, including 1976 and 1984.

We chose 1976 as the design drought because Colliford does not refill in 1976 and for a number of years after.

The DO is determined by infrastructure constraints (including treatment) over the peak week demand period and detailed technical information has been shared with the Agency. The link to level of service is described in Section 2.3 below. .

2.1.3 DO / WAFU: Roadford WRZ

2.1.3.1 Exports and imports

Our water resources model incorporates the imports and exports for the Roadford WRZ, which include:

ƒ Saltash transfer from Roadford WRZ to Colliford WRZ

ƒ imports / exports from Wimbleball WRZ to Roadford WRZ near Exeter

ƒ Tiverton to North Devon transfer from Wimbleball WRZ to Roadford WRZ.

Page 2.3

SWW Water Resources Management Plan 2015 – 2040

2.1.3.2 Roadford fisheries bank

We have made an allowance in these calculations for releases from the Roadford Fisheries Bank in accordance with the provisions of the Roadford and Burrator Reservoirs Operating Agreement.

2.1.3.3 Critical event(s) that define DO

We used our water resources model to simulate the system through the complete record of flow sequences. We found that the most severe drawdown occurred in Roadford Reservoir during the 1976 drought and chose this as the design drought.

The DO is determined by infrastructure constraints (including treatment) over the peak week demand period and detailed technical information has been shared with the Agency. The link to level of service is described in Section 2.3 below.

2.1.4 DO / WAFU: Wimbleball WRZ

2.1.4.1 Conjunctive use of groundwater sources in the Wimbleball WRZ

Our MISER water resources model uses monthly output profiles derived from DO figures for the groundwater sources which were re-assessed in 2012. This approach has been supported by the Environment Agency for the all of our Water Resources Management Plans post 1999.

2.1.4.2 Exports and imports

Our water resources model incorporates the imports and exports from the Wimbleball WRZ, which include:

ƒ the Wessex Water abstractions from Wimbleball Reservoir for direct piped transfer to Maundown WTW

ƒ the treated water transfers between the Roadford WRZ and the Wimbleball WRZ in the Exeter area

ƒ the treated water transfers into the Roadford WRZ in the Tiverton area

ƒ the small exports of treated water to Wessex Water.

2.1.4.3 Derivation of the Wimbleball pumping curve

We have produced the Wimbleball pumping curve to ensure that the reservoir is full on 1st April before the start of the summer drawdown.

2.1.4.4 Wimbleball fisheries and conservation water bank

We have made an allowance of 900 Ml per annum for the Wimbleball fisheries and conservation water bank (in accordance with Clause 22 on Licence No 14/45/02/2388) in all calculations.

Page 2.4

SWW Water Resources Management Plan 2015 – 2040

2.1.4.5 Critical event(s) that define DO

We used our water resources model to simulate the system through the complete record of flow sequences. We found that similar severe drawdowns occurred in Wimbleball Reservoir in the 1976 and 1990 droughts.

We have selected 1976 as the design drought which is the same as the design drought in the Roadford WRZ, and is appropriate because of the linkage between Roadford WRZ and Wimbleball WRZ. Using identical design droughts in both Roadford WRZ and Wimbleball WRZ also simplifies representation of the imports and exports between the WRZs.

The DO is determined by infrastructure constraints (including treatment) over the peak week demand period and detailed technical information has been shared with the Agency. The link to level of service is described in Section 2.3 below.

2.2 Confidence rating for our source DOs

The Water resources planning guideline5 recommends water companies determine the confidence rating for its source DO‘s as proposed by WR276. A copy of the confidence matrix is shown below.

Figure 2.1: Level of confidence for DO confidence labelling7

Within the above table, we have interpreted the word “hydrological” to include all available hydrological data ie the use of both rainfall and river flow data.

5 Environment Agency, “Water resources planning guideline – interim update”, October 2012 6 UKWIR, Project WR27 “Water Resources Planning Tools 2012”, 2012 7 Ibid. 5

Page 2.5

SWW Water Resources Management Plan 2015 – 2040

Based on the table above, for all of our resource zones we believe the DO confidence label AB is appropriate. We have based our decision on the reliability of our flow data; assessment of historic rainfall data and detailed knowledge and understanding of our water resources systems. We are unable to give a higher confidence rating because of the lack of long term robust river flow, groundwater or rainfall records in our area.

2.3 Level of service and DO assessment

We have derived values for our baseline DOs in our baseline supply demand balance assuming our planned level of service as described in Section 1.6 of this Plan.

We have also assessed our baseline DO (without climate change) for the additional following level of service scenarios:

(i) No restrictions – defined as the constant rate of supply that can be maintained from the WRZ throughout the entire period of assessment, with no customer restrictions or other drought actions applied.

(ii) A reference scenario level of service as defined in the Water resources planning guideline8 . This is the rate of supply that can be maintained from the WRZ throughout the entire period of assessment when the system is operated to meet a specified level of service. For the reference level of service these are for temporary customer use restrictions of 1 in 10 years and non-essential use restrictions of 1 in 40 years. No rota cuts or standpipes should be used within the period of record.

A summary of the differences in the assumptions between the SWW level of service along with the reference level of service is shown below.

Table 2.1: Summary of differences between SWW and the reference level of service

Reference SWW

level of service level of service Temporary customer use Average frequency of 1 Not more than 1 in 20 restrictions in 10 years years on average Non-essential use Average frequency of 1 Not more than 1 in 40 restrictions in 40 years years on average

As can be seen in the above table, the levels of service are of a similar nature, with the differences being the frequency of the restrictions.

However, as detailed in Table 2.1 above, in all our WRZs, our DO is determined by strategic infrastructure constraints over the peak demand periods and therefore because peak demand periods can obviously occur outside the most critical dry periods, changing the level of service has no influence on our DO.

8 Ibid. 5

Page 2.6

SWW Water Resources Management Plan 2015 – 2040

We have held detailed technical discussions and shared technical information with the Agency on the above as well as showing how we meet our planned level of service for all our WRZs.

Within the Water resources planning guideline9, a table was recommended to help understand the relationship between DO and level of service. We understand this table is primarily for companies with a supply demand deficit. As we are in a supply demand surplus throughout the planning period, presenting the information in this way is potentially misleading and we have shared this approach with the Agency.

2.4 Reductions in DO – sustainability changes

2.4.1 Changes identified in WRP0910

The Habitats Directive is a European-wide law designed to protect internationally important habitats and species. The Directive has established a network of protected sites across Europe, called Natura 2000 sites, consisting of Special Areas of Conservation (SACs) and Special Protection Areas (SPAs).

As part of implementing the Habitats Directive, the Environment Agency has a duty to consider the impact of new and existing permissions on the special interests of these protected sites. This includes the Environment Agency looking at all the permissions it has issued, on, close to, or linked to Natura 2000 sites.

In preparation of our WRP09, the Environment Agency (South West) identified a number of abstraction licences which are in hydrological connection with the Dartmoor SAC. The Dartmoor SAC is designated for wet and dry heaths, blanket bog, alkaline fen, sessile oakwoods, southern damselfly, otter and Atlantic salmon.

The following South West Water public water supply licences were considered by the Environment Agency:

ƒ Red Lake, River Erme & Left Lake ƒ Venford Reservoir ƒ River Swincombe ƒ Littlehempston Intake ƒ Littlehempston ƒ Cowsic River, Blackbrook River, Devonport Leat & West Dart River ƒ Kennick, Tottiford & Trenchford Reservoirs ƒ Fernworthy Reservoir ƒ River Yealm/Broadall Lake/Ford Lake ƒ Lopwell Dam

9 Ibid. 5 10 Ibid. 3

Page 2.7

SWW Water Resources Management Plan 2015 – 2040

Whilst the proposed changes in the abstraction licences have been identified and confirmed at most of the above sites, the details of the potential changes at a few sites are still outstanding.

For the purposes of this Plan, we have therefore included the Environment Agency’s previously provided estimate of the impact on our base DO for the Roadford WRZ of 4.8 Ml/d for these sustainability reductions.

The Environment Agency identified improvements which are required at some of the intakes to assist fish passage and fish screening and these schemes are progressing. There is no impact on DO of these schemes.

2.4.2 Licence renewals due between 2015 and 2020

In the Wimbleball WRZ two key groundwater time-limited licences covering six boreholes in the Otter Valley will require renewal in 2017. In recent years we have not fully utilised these licences and the Environment Agency wishes to confirm that there will be no adverse environmental impact should we increase our level of abstraction. An environmental impact assessment for fisheries is currently underway and its conclusions will be submitted as part of supporting information for the licence renewal applications.

Also in the Wimbleball WRZ, three time-limited licences covering three boreholes at our Greatwell wellfield will require renewal in 2017. In line with a presumption of renewal of time-limited licences and following recent discussions with the Environment Agency, it is not envisaged that there will be any reductions to licensed rates that would affect the DO of these groundwater sources.

In the Colliford WRZ, time-limited abstraction licences for our Park Lake and Stannon Lake sources are due for renewal in 2016. Both sources are currently subject to a programme of investigation into their environmental impact which will inform the renewal process. As of 2014 significant environmental monitoring and analysis have already taken place. As required by WRMP planning guidelines it has been assumed in this Plan that both licences with be renewed.

2.4.3 Changes identified for this Plan

2.4.3.1 Changes identified by the Environment Agency

Some further water company abstractions across the country are also thought to have a detrimental effect on the environment and changes to the abstraction licence conditions or other mitigation measures could be required. These changes may be required to protect international or national designated conservation sites (Habitats Directive, Sites of Special Scientific Interest (SSSI) or Biodiversity 2020 sites), to protect locally important sites (undesignated sites) or to deliver Water Framework Directive (WFD) objectives.

The Environment Agency formally notified water companies about what changes could be required to their abstraction licences to make them more sustainable and how to take account of these within our WRMP. We have included information provided to us by the Environment Agency in August 2013.

Page 2.8

SWW Water Resources Management Plan 2015 – 2040

In our area, the table below shows the following sites which have been identified by the Environment Agency. All of the schemes below have been categorized by the Environment Agency as “confirmed” or “likely” and identified as “statutory” sustainability reductions as oppose to “non-statutory”. We have agreed the names of the schemes with the Environment Agency.

For completeness, where appropriate, we have included our assessment of the estimated impact on our baseline DO Water. Further information on this is given below.

Calculations to date confirmed that the above schemes are cost effective and were included in our Business Plan submission to Ofwat.

All of the sustainability changes being considered within the plan below have been identified as a result of work by both the Environment Agency and ourselves, and are in connection with the WFD.

Other further potential actions to address potential WFD issues, such as the way a fisheries bank allowance is used, have been identified by the Environment Agency and ourselves and details are being discussed with the Environment Agency. These schemes have not been included in the Table 2.2 as they cause no impact on DO and hence have no impact on this plan. They are also able to progress without the need to have been included in our Business Plan submission to Ofwat.

Page 2.9

SWW Water Resources Management Plan 2015 – 2040

Table 2.2: Options under consideration to address potential WFD issues

Estimated reduction in WRZ Scheme name Scheme Description baseline DO (Ml/d) Colliford Colliford and St Neot The preferred options are: (a) Annual stocking of salmon fry of 0.00 local provenance to the St Neot by hatchery restocking

(b) Annual improvement of 2 instream 0.00 habitat sites on the St Neot from

2015 to 2020

(c) Hydro-ecological investigation of 0.00 use of pumped storage pipeline and fisheries water bank releases to provide optimal salmonid production within the St Neot post Colliford impoundment Roadford Avon Gravel augmentation for spawning 0.00 areas, including monitoring of the success of gravel augmentation. Kennick Trenchford (a) To introduce a compensation flow 2.16 and Tottiford (KTT) (1) of Q95 (assumed to be 2.16 Ml/day) downstream of Kennick Trenchford and Tottiford (b) Monitoring / investigation to assess 0.00 the success of introducing a compensation flow, in terms of fish production, the possible need for a fisheries bank and gravel augmentation Fernworthy (a) Options appraisal to consider 0.00 further the introduction of a fisheries bank. (b) Gravel augmentation for spawning areas, including monitoring of the 0.00 success of gravel augmentation Wimbleball Haddeo – Wimbleball Gravel augmentation for salmonid 0.00 system (2) spawning areas, along with associated monitoring.

Notes:

(1) As referenced by the site names Beadon Brook; Tottiford Reservoir; Trenchford Reservoir and Kennick Reservoir in the spreadsheet provided by the Environment Agency in August 2013.

(2) As referenced by the site names in connection with the Wimbleball System in the spreadsheet provided by the Environment Agency in August 2013.

Page 2.10

SWW Water Resources Management Plan 2015 – 2040

2.4.3.2 How we have taken account of the sustainability changes

The Water resources planning guideline11 advises that companies should include both confirmed and likely sustainability reductions in the appropriate resource zones in the baseline supply demand balance. As can be seen in the above table, the schemes which we are required to include in our baseline supply demand balance are the KTT and Fernworthy schemes.

(i) Kennick Trenchford and Tottiford (KTT) - Introduction of compensation flow

Introduction of a compensation flow results in a constant loss of water from reservoir storage and hence impacts on the DO of the reservoir. There is currently no compensation flow released downstream of the reservoirs. This estimate of the impact on DO has been shared with the Environment Agency.

(ii) Fernworthy

As discussed above, it is not possible to quantify the potential impact on the DO at present, but the Environment Agency have advised it is not believed to be significant.

Although there are uncertainties with sustainability changes, we have not included them in our headroom calculations as advised by the Water resources planning guideline12.

The Environment Agency has confirmed that there are no sustainability reductions under the Restoring Sustainable Abstractions (RSA) Programme that we need to take account of in our forecasts of DO.

2.4.3.3 Scenario analysis of sustainability changes

We do not believe that with the information currently available, this WRMP is sensitive to sustainability changes. We have shared this information with our local Environment Agency team who have confirmed our decision that there is no appropriate scenario analysis for the sustainability changes in our plan.

11 Ibid. 4 12 Ibid. 4

Page 2.11

SWW Water Resources Management Plan 2015 – 2040

2.4.3.4 Implementation dates of proposed sustainability changes

The implementation dates we are currently proposing are shown in the table below.

Table 2.3: Our proposed implementation dates for sustainability change

Proposed WRZ Scheme name implementation date Colliford Colliford and St Neot As soon as practical Roadford Avon As soon as practical Kennick Trenchford and Tottiford (KTT) (a) introduction of 2016/17 compensation flow (b) Monitoring / 2016/17 investigations work Fernworthy (a) Consideration of 2017/18 Fisheries Bank (b) Gravel 2016/17 augmentation Wimbleball Haddeo – Wimbleball As soon as practical system

We have shared and discussed the above implementation dates with the Environment Agency, who believe our approach is reasonable.

2.4.3.5 Voluntary changes proposed by the Company

We have not proposed any further voluntary reductions in DO for environmental benefit within this plan.

Page 2.12

SWW Water Resources Management Plan 2015 – 2040

2.5 Abstraction Incentive Mechanism

As advised by the Water Resources Planning Guideline13, we have not factored in any implications of the Abstraction Incentive Mechanism (AIM) into our forecast of DO.

However, as advised by Ofwat, we have no abstraction sites that were required to be included with AIM.

2.6 Reductions in DO - Outage

Temporary short-term losses in DO that we might experience are dealt with as outage. A summary and conclusions of our work on outage is given below, with further details in Appendix C.

2.6.1 General

The methodology we have used for the calculation of the outage allowance is similar to that which we used for our previous WRP, which follows best practice14.

We calculated outage using a detailed analysis of historic records and a thorough understanding of water treatment and distribution systems. The outage we calculated for each WRZ is based on the effect of outages at individual sources/WTW. Further details of outage calculations are provided in Appendix C.

We considered the following types of outage:

ƒ Planned outages

ƒ Unplanned outages which occur as a result of a WTW’s inability to cope with the raw water quality it receives; and/or insufficient provision of standby equipment to cover for breakdowns and maintenance; and/or power and system failures.

In addition we consider three specific types of outage concerning borehole sources in the Wimbleball WRZ. These outages were assessed using a Monte Carlo simulation.

2.6.2 Planned outage

In our region, planned resting or shutting down of sources/WTW for say capital maintenance, tends to be for only short periods of time. In addition, short complete or partial shutdowns for maintenance are typically planned for when strategic sources are not being deployed eg during the winter period. Hence, planned outage has no significant impact on WAFU and therefore for the purposes of this report there is no requirement for a planned outage allowance.

13 Ibid. 4 14 UKWIR, “Outage Allowances for Water Resources Planning”, 1995

Page 2.13

SWW Water Resources Management Plan 2015 – 2040

2.6.3 Unplanned outage

We investigated potential unplanned outage by analysing historical source/WTW outputs and identifying when outputs had been lower than expected. When this had occurred, we collated information to confirm why the source/WTW had not produced the expected levels of output and then assessed and quantified the impact of the lower source/WTW output on the strategic WAFU. We used the results from this analysis to provide an unplanned outage allowance. Further details and historic data are given in Appendix C.

2.6.4 Specific borehole-related outages

Analysis of outage events in the Wimbleball WRZ revealed that in recent years borehole sources were subject to potential outage as a consequence of three specific types of failure. These were:

1. Dotton boreholes 1 and 3 removed from service because of high turbidity associated with flooding of the River Otter.

2. Greatwell boreholes 1, 2 and 3 suffering unacceptably high turbidity levels on start up.

3. High failure rate of borehole pumps experienced in AMP5.

Operational issues caused by these events were felt significant enough to warrant specific analysis as part of the outage assessment. The consequences of such outages were assessed by Monte Carlo simulation as described in the Outage methodology15.

2.6.5 Other potential outages

In our region, other losses of output tend to be associated more with long term issues than those of a temporary nature and hence we have factored them into the calculation of DO rather than allowing for them as outage.

2.6.6 Total outage allowance for each WRZ

In the Colliford and Roadford WRZ, the calculated outages are very small, less than 1 Ml/d (see Appendix C). Therefore we have assumed de minimus values for outage (as in our previous WRP16).

In the Wimbleball WRZ, we have identified two types of outage: unplanned outage and outage as a result of three specific types of borehole-related events. Whilst the former has been assessed using analysis of operational data which indicated a de minimus outage value of 1 Ml/d was appropriate, the borehole-related outage events have been investigated using Monte Carlo simulation, which indicated that an average of 4 Ml/d could be lost from a combination of these particular outage events.

15 Ibid. 14 16 Ibid. 3

Page 2.14

SWW Water Resources Management Plan 2015 – 2040

The cumulative outage allowance for each WRZ is summarised below:

Table 2.4: Cumulative outage

Outage allowance WRZ (Ml/d)

Colliford - general unplanned 1.0 Roadford - general unplanned 1.0 Wimbleball - general unplanned 1.0

Wimbleball - borehole-specific outages 4.0 Total 7.0

Although underlying levels of general unplanned outage do not justify specific measures to reduce outage further, the specific types of outage associated with groundwater sources are being addressed:

ƒ A replacement borehole for Dotton No. 1 is being considered as part of the planning process which would be located and designed such that river flooding will not affect water quality

ƒ Greatwell boreholes 1, 2 and 3 are currently being investigated and remediated with the intention of reducing turbidity issues on start-up

ƒ Borehole pumps are being more closely monitored for performance and pump quality concerns are also being addressed.

2.7 Raw and potable water transfers and bulk supplies

We have a very small export of treated water to Wessex Water, which we have shown as planned transfers for each time step across the planning period. The volumes are within the existing physical and operational transfer capacities and have been agreed with Wessex Water. The transfers are of a similar order in both dry and wet years and represent the existing maximum transfer capacity which is limited by pipeline capacity.

2.8 Distribution and treatment works operational use and losses

We have calculated our treatment works losses within each WRZ for a dry year and show these values in our tables. It should be noted that in wetter years these values can be higher for operational and water quality reasons.

Page 2.15

SWW Water Resources Management Plan 2015 – 2040

2.9 Greenhouse Gas Emissions

Regarding our current operations, we calculate the greenhouse gas emissions for each of our sites every year and provide an aggregated operational greenhouse gas measure for whole company activity to Ofwat each year as part of the Company Annual Performance Reporting (CAPR) requirements. The greenhouse gas measure for the water supply side of the business is 48,095 tCO2e.

We have a high degree of confidence in our greenhouse gas calculations and measurement. Our CAPR greenhouse gas emissions data is audited by external auditors prior to being sent to Ofwat and audited again under separate conditions as part of our carbon certification under CEMARS (the Certified Emissions Measurement and Reduction Scheme).

Page 2.16

SWW Water Resources Management Plan 2015 – 2040

3 Water demand 3.1 Introduction

This section of our Plan describes how we expect the demand for water to change over the period to 2039/40. Many of the definitions and methodologies used in our forecasts are described in the reports „Demand forecasting methodology‟1 and „Forecasting water demand components‟2 3.1.1 Scenarios modelled

When ensuring that we have the ability to meet the demand for water we considered dry years, as it is during these that the pressure on our resources is at its greatest. Therefore the supply demand analysis on which this Plan is based uses forecasts of demand under a dry year scenario. We have not included any restrictions in usage that may be required during a drought, as it is important to understand the unconstrained demand. We have also produced a forecast of demand for more normal years and this is used for the calculation of our weighted average demands.

As explained in Section 1 our water resources system is not constrained by a critical period, so we have modelled only annual average scenarios in our supply demand appraisal.

As we forecast that the supply of water will exceed the demand for the entire planning period, no resource development will be required, and our final planning scenario is identical to the baseline scenario. As we do not predict a deficit, no utilisation forecast is required. 3.1.2 Demand in the base year

Before producing forecasts of future demand, it is important to have robust estimates of water consumption in the base year of the Plan (2012/13).

Under the regulatory regime in place in England and Wales, demand is considered in terms of eight components, which together comprise Distribution Input (DI):

. Measured household consumption . Unmeasured household consumption . Measured non-household consumption . Unmeasured non-household consumption . Leakage . Distribution system operational use . Water taken legally unbilled . Water taken illegally unbilled

1 National Rivers Authority and UKWIR, “Demand forecasting methodology”, Ref 95/WR/01/1, 1995 2 Environment Agency and UKWIR, “Forecasting water demand components: Best practice manual”, Ref 97/WR/07/1, 1997

Page 3.1

SWW Water Resources Management Plan 2015 – 2040

Each year as part of our Company Performance Review (previously as part of our June Return to Ofwat), we make estimates of each of these components and compare them to the amount of water we output from our treatment works. The difference between the sum of the estimated components and the output of our treatment works is called the Water Balance Gap (WBG), and this must be accounted for in order to produce robust estimates.

To account for the WBG and reconcile our estimates of demand with works outputs, we use the well established Maximum Likelihood Estimation (MLE) methodology. MLE is a statistical technique which redistributes the WBG to the components of demand, with more of the gap being assigned to the large, less certain components. It is these reconciled estimates that are used as the basis of our Plan. Our WBG for 2012/13 was 13.49 Ml/d and this has been redistributed as shown in Table 3.1.

Table 3.1: Reconciliation of demand components in the base year

WBG Reconciled Estimate Demand component adjustment estimate (Ml/d) (Ml/d) (Ml/d) Measured household consumption 123.89 1.65 125.54 Unmeasured household consumption 87.89 3.51 91.40 Measured non-household consumption 81.70 1.09 82.79 Unmeasured non-household consumption 7.34 0.49 7.83 Leakage 82.04 2.18 84.22 Distribution system operational use 2.67 0.18 2.85 Water taken legally unbilled 14.30 0.95 15.25 Water taken illegally unbilled 4.55 1.21 5.76 Sum of components 404.38 11.26 415.65 Distribution input 417.87 -2.23 415.65 Note that values in this table may not sum exactly due to rounding.

The base year for our draft WRMP was 2011/12 as this was the latest year for which outturn data was available. As we now have 2012/13 outturn data available we have been able to update our forecasts to use this as the base year. The 2012/13 DI used in this Plan, which is based on the new outturn data, is around 1% higher than that predicted in the draft Plan. This difference is detailed in Table 3.2; it is not significant to the Plan.

Table 3.2: Comparison of 2012/13 DI used in the draft and final Plans

2012/13 distribution input Normal year Dry year Draft Plan forecast (Ml/d) 412.82 428.44 Final Plan outturn-based (Ml/d) 417.81 433.89 Difference +1.2% +1.3%

Page 3.2

SWW Water Resources Management Plan 2015 – 2040

3.1.3 Metering policy

We currently have one of the highest levels of metering in the industry, with around 75% of our household customers and 92% of non-household customers paying by metered billing.

3.1.3.1 Household metering policy

Since 2000 our unmeasured household customers have had the option of switching to pay according to the amount of water that they use, without being charged to make this change. This has helped the level of metering to increase rapidly from less than a quarter in 2000, to its current level of about three-quarters of households. The option to switch to metered billing remains popular, with over 13,000 households switching during 2012/13.

Under regulations published by the Secretary of State for the Environment we have the right to install meters at household properties with high discretionary use. We have exercised this right since 1990 when we asked sprinkler and swimming pool owners to register with us, resulting in meters being installed at 5,700 properties. We continue to install meters at properties having sprinklers or swimming pools but with the majority of such properties now assumed to be metered, the number of customers being metered for this reason is now very small.

With a continuation of the existing metering programme forecast to lead to almost 90% metering in 2039/40 and no supply demand deficit predicted, we are not intending to change our metering policy. Our demand forecasts are therefore based on a continuation of our current metering programme.

Continuing our proposed metering programme, consisting of optional metering along with a small number of selective meters installed because of high discretionary use, is forecast to require a capital expenditure of £12 million over the 2015/16 to 2019/20 period, with metering operating costs of £1.5 million per year. The process of selectively metering a property due to high discretionary use is the same as that for metering an optant property, and therefore there is no difference in either the capital cost of installing the meter or the operational expenditure required to read it. We have shared this further cost information with our regulators.

It is not always possible to install a meter at a property, and where this is the case the customer is put on an assessed charge. Ofwat‟s assessment is that it is too difficult or uneconomic to install meters at around 10% of properties3, hence our assumption that we will only achieve around 90% metering in 2039/40.

3.1.3.2 Non-household metering policy

Our non-household customers have to the option to switch to metered billing, although there is a charge for them to do so.

The proportion of non-households at which it is too difficult or uneconomic to install meters is lower than it is for household properties. This is illustrated by our current non-household metering level of around 92%, which we forecast will reach 95% by 2040.

3 Ofwat, “Exploring the costs and benefits of faster, more systematic water metering in England and Wales”, October 2011

Page 3.3

SWW Water Resources Management Plan 2015 – 2040

3.2 Demography 3.2.1 Our region

We supply Cornwall, Devon and small parts of Somerset and Dorset with water, an area which is largely rural with much of the population living in small communities. Close to a third of the population live in the three major urban areas; Plymouth, Exeter and Torbay, which are all located in Devon. 3.2.2 Demographic forecasts

Our forecast of population and housing growth up to 2039/40 has been provided by Experian, whose report detailing the production of these forecasts is shown in Appendix E. The projections have been produced in accordance with the Water resources planning guideline4 and the Methods of Estimating Population and Household Projections5 report and are based on Census 2011 data.

The forecasts have been updated since our draft WRMP to include the latest data available from the 2011 Census. The updated demographic data includes new forecast trends from the Office of National Statistics (ONS) which fully reflect the results of the 2011 Census, and the latest information from local authorities.

Experian have provided three sets of demographic forecasts:

. Local authority plan-based projections – These used information from local authorities on the development they expected to occur in their areas.

. Trend-based – These were based on the most recent available official demographic statistics from the ONS.

. Most likely –These forecasts are Experian‟s best view of the likely growth.

The economic recession which started in 2008 resulted in reduced numbers of houses being built. Prior to the recession there were between 6,000 and 7,000 new connections to our water supply network each year, but in 2012/13 there were less than 5,000. Population growth has proved more resilient to the economic conditions and was not significantly affected by the recession.

The local authority plan-based projections show a rapid recovery in house-building numbers, rising to over 9,200 in 2016, which is around one-third higher than was being achieved prior to the recession. This increase in housing growth is accompanied by a similar rise in population growth to a rate much higher than has been seen in the past. We have therefore concluded that the plan-based forecasts are unrealistically optimistic, with very little chance of them being met.

The most-likely estimates are influenced the local authority plan-based forecasts and while they show less optimistic build rates, the 2016 figure is still around 20% higher than in 2008. This build rate is again unrealistically optimistic.

4 Environment Agency, “Water resources planning guideline – interim update”, October 2012 5 Environment Agency, “Methods of Estimating Population and Household Projections: Update 2012”, June 2012

Page 3.4

SWW Water Resources Management Plan 2015 – 2040

The trend based forecasts show build rate quickly returning to a level close to that observed prior to the recession, with population increase continuing at a similar level to recent years. We consider that these forecasts are both realistic, and of the three sets provided by Experian are the most likely to describe actual growth rates. We have therefore used the trend-based forecasts in our Plan.

Through contact with developers who have consulted us as part of the planning process, we were aware of developments which are very likely to progress but which are not yet reflected in official statistics, and so were not included in the forecasts. Where we knew of such developments, we made changes to Experian‟s data to reflect this, concentrating the forecast growth in the area to that site. This engagement with developers along with regular contact with local authorities allows us to keep a clear view of where development will occur and will continue to play an important part in our long-term planning.

The forecasts were provided at the Lower Super Output Area (LSOA) level. By cross referencing with our billing system, in which we have each property assigned to its LSOA and its respective WIS zone, we are able to assign population to the appropriate WIS zone, summing these to obtain WRZ populations. 3.2.3 Population

A report produced for us by the School of Geography at the University of Leeds identified some categories of population that were not covered by ONS population estimates, and which are important for us to consider. These categories were EU accession country migrants, visitors overstaying their permitted time in the Country, those entering the Country clandestinely and victims of human trafficking. The University of Leeds‟ medium estimate of this additional population in our region was 15,464, so we have added this to Experian‟s figures.

The resident population of the region was estimated to be 1.709 million in 2012 and this is forecast to grow to 2.045 million by 2039. However some of this population will be connected to private water supplies and will not be reliant on our supply. Local Authorities have a responsibility to monitor private water supplies, so have information on the number of properties connected to them. By contacting the authorities to obtain summary data we have estimated that 1.3% of this total regional population is not served by us. Our forecast of growth in the resident population in the area that we supply with water is shown in Figure 3.1.

For comparison Figure 3.1 also shows the population forecasts we used in our draft Plan. The latest ONS forecasts show stronger population growth rates, resulting in an increase of 110,000 (5.9%) in the forecast 2039 population over that used in our draft Plan.

Page 3.5

SWW Water Resources Management Plan 2015 – 2040

Figure 3.1: Growth of the resident population in the area we supply with water

2.2

2.0

1.8

1.6

1.4 Population(millions) Draft Plan forecast Final Plan forecast 1.2 Historic

1.0

1979 1984 1989 1994 1999 2004 2009 2014 2019 2024 2029 2034 2039

We estimate that currently 5.4% of the population connected to our water supply reside in non-household properties. The non-household properties having resident population fall into two categories:

. Communal – such as barracks, nursing homes, boarding schools, etc.

. Domestic agricultural properties – farms which are classified as non-household, but which are also homes.

We have used Experian‟s estimates of the communal population which were provided at LSOA level, allowing us to assign this population to the appropriate WIS zone and hence to its parent WRZ. Domestic agricultural properties are identified on our billing system, and we have estimated the population resident in them by assuming they have the same average occupancy rate as measured and unmeasured households. 3.2.4 Housing

With the economic situation being difficult over the last few years the number of new housing completions has fallen from previous levels. While the economy is showing signs of recovery we expect that it will take some time for the construction sector to recover to a position of strength, so the number of new connections to the water supply network is likely to remain suppressed for the immediate future. Our forecasts suggest that new connections will remain around 5,000 until 2014/15 before recovering to over 6,300 in 2015/16.

Page 3.6

SWW Water Resources Management Plan 2015 – 2040

To ensure consistency of this Plan with other returns to our regulators, we use the definition of households historically used by Ofwat. This definition is slightly different to that used by the ONS, and this difference can be highlighted by considering the example of a block of flats served by a single meter. This is considered to be a single non-household property by the Ofwat definition, but as a number of households by the ONS. Because of this difference in definition, the numbers of households used in this Plan do not match those provided by Experian using ONS data.

To overcome the difference in definitions, we first took the base year property numbers from our billing system using Ofwat definitions. As all new properties are now metered individually, we were then able to take Experian‟s year-to-year increases in household numbers and apply these to our base year numbers.

As all properties currently in our billing system are assigned to a WIS zone, and forecasts have been provided at LSOA level, it is simple to group both existing properties and forecast growth to WRZs.

Figure 3.2 shows our forecast of the number of household properties connected to our supply system. In 2012/13 there were 716,000 household properties connected to our clean water network and this is forecast to reach 882,000 in 2039/40.

Also shown in Figure 3.2 is a comparison with the forecasts we used in our draft Plan. Including the most recent data in our forecasts has increased the forecast number of connected properties by 16,000 in 2039/40.

Figure 3.2: Stacked line chart showing the number of household properties connected to our supply system

1,000 Measured

Unmeasured 800 Draft plan total

600

400

200

0

Connectedhousehold properties (thousands)

1999 2004 2009 2014 2019 2024 2029 2034 2039

Page 3.7

SWW Water Resources Management Plan 2015 – 2040

At any one time, a number of the properties connected to our supply system are not billed because they are unoccupied. We have obtained the number of these void properties in the base year from our billing system, and in 2012/13 0.9% of metered households were void compared to 2.7% of unmeasured households, an overall household void rate of 1.4%. We have assumed a continuation of the measured and unmeasured void rates, but as meter opting and new connections add to the measured customer base, the overall household void rate is projected to fall slightly to 1.2% in 2039/40. 3.2.5 Average household size

In recent decades the Average Household Size (AHS) has fallen; nationally it has dropped from 3.1 people per household in 1961 to 2.4 in 2012. We expect this trend to continue, predicting that AHS in the region we serve will drop slightly from its current value of 2.3 people per household to 2.2 in 2039/40.

To estimate AHS in measured and unmeasured properties in the base year, we used data obtained from our household consumption monitor. Each year we ask members of the measured and unmeasured surveys for the number of people resident in their household, and use this information to calculate averages for these categories. The surveys have been designed to be representative of the wider customer base, so it is reasonable to base our AHS estimates on these data. Using these AHS estimates in combination with property numbers from our billing system gave an estimate of the measured and unmeasured populations, which we then reconciled against the ONS regional estimate by applying a correction factor. For the base year a correction of 0.4% was required to be applied to the survey AHSs to match the ONS estimate. AHS is calculated for our region as a whole rather than for each WRZ individually, as the area we serve does not differ enough demographically to justify individual estimates.

In producing these forecasts we have assumed that the AHS in new build properties is the same as the overall measured household AHS.

Forecasts of the AHSs for the different population categories are shown in Figure 3.3 below. The overall AHS is based on Experian‟s demographic data. The AHS of meter optant properties is currently very close to the overall AHS and is expected to rise as it becomes financially advantageous for larger households to switch to metered billing. In the second half of the planning period however there are very few optants as few properties remain unmeasured, so the trend is influenced more by the general reduction in AHS. The AHS of unmeasured properties initially rises as the smaller of these households migrate to the metered category, again the small number of optants in later years results in the trend following that of the overall AHS. As meter penetration is already high, and we forecast that it will reach almost 90% by 2039/40, measured AHS is similar to the overall level.

Page 3.8

SWW Water Resources Management Plan 2015 – 2040

Figure 3.3: Forecast change in average household size

3.5

3.0

2.5

2.0

1.5 Measured 1.0 Average Average Household Size Unmeasured

0.5 Meter optant Overall

0.0

2012/13 2017/18 2022/23 2027/28 2032/33 2037/38

3.3 Household consumption 3.3.1 Our approach to forecasting household consumption

The first step in estimating how much water customers will use in the future is to understand their current consumption. In gaining an understanding of this there are a couple of key aspects that need to be investigated:

. How much water do consumers use around their homes, and what do they use this water for? . How does the amount of water they use vary according to the weather?

Once these aspects are understood, they can be used as the basis for estimating how water use is likely to change in the future. The different steps we took to understand current water use and to predict it in the future are detailed in the sections below.

Different groups of customers use different amounts of water, for example measured customers on average use less than unmeasured ones. To ensure our forecasts were robust we grouped households into four distinct categories, each of which is likely to exhibit a different pattern of consumption:

. Existing measured – Properties that were metered in the base year of the Plan. A property in this category will remain in it for the duration of the planning period.

. Unmeasured – Properties that remain unmetered. Due to the optional metering programme that it is assumed will run for the duration of the Plan, members of this group will migrate to the meter optant category, and it will reduce in size.

Page 3.9

SWW Water Resources Management Plan 2015 – 2040

. Meter optants – In the base year there are no properties in this group, as all meter optants up to this time are included in the „existing measured‟ category. When a household switches to metered billing it joins this group, where it remains until the end of the Plan. As unmeasured households with lower consumptions are more likely to save money by switching to metered billing, these meter optants will tend to have lower consumption than the unmeasured average.

. New build – As with meter optants, there are no properties in this group in the base year. New build houses are more likely to have more water efficient fixtures and appliances, therefore their average consumption is likely to be lower than the „existing measured‟ average.

The most widely used measure of how much water household consumers use is Per Capita Consumption (PCC) and our forecasts were based on this measure. Once robust PCC forecasts were produced, we combined these with the population data to obtain total household demand. Our population data is derived from estimates of AHS (described above) and property data from our billing system, and because each property is assigned to a WIS zone, this allows consumption to be estimated at both WIS and WRZ levels.

We have used regional PCC estimates rather than ones for individual WRZs as the region we supply does not differ sufficiently in demographic and water consumption terms to make separate estimates worthwhile. 3.3.2 Historic PCC

One of the most useful sources of information in understanding current consumption is historic data and we have made extensive use of such information in preparing this Plan. Unmeasured household PCC has been obtained from our unmeasured consumption monitor, whilst measured data comes from our billing system.

Our household consumption monitors are very important to our understanding of customer consumption, and we will continue to operate these to allow us to collect data for the next planning period.

Historic average PCC (excluding the effect of meter under-registration and plumbing losses not recorded by a meter) is shown in Figure 3.4; we only have measured data from 2000 as prior to this there were very few households that paid according to the amount of water they used.

Page 3.10

SWW Water Resources Management Plan 2015 – 2040

Figure 3.4: Historic PCC for measured and unmeasured households

170 Unmeasured

Measured 160

150

140

130

120

110 Consumption (litres per person per day)

100

1982/83 1987/88 1992/93 1997/98 2002/03 2007/08 2012/13 1977/78 3.3.2.1 Estimation of historic unmeasured PCC

We have run an unmeasured household consumption monitor since the 1970s. This was set-up to include around 1,000 properties that were selected to be representative of the unmeasured customer base in our region. Given that the area we serve does not differ greatly in terms of socio-economic or geographic factors and consumption patterns are similar throughout, it was not necessary to stratify the sample by resource zone. Properties in the survey had a meter fitted, which we read twice a year, but they continue to receive an unmeasured bill. Over time many of the original sample decided to leave the survey or opted to switch to metered billing, requiring us to periodically recruit more properties.

As well as recording the consumption of member households, each year we ask member households for occupancy data, allowing us to calculate PCC from the meter readings. In addition, every few years we send survey members a questionnaire about how they use water around in their homes, covering issues such as personal washing, appliance ownership and garden watering.

Prior to 2012/13, members of our unmeasured household consumption monitor with high consumption were excluded from the estimation of PCC as this was assumed to be the result of a leak at the property rather than being real consumption. Given that the number of customers remaining unmeasured is now low, and that some of these customers are likely to have chosen to remain unmeasured because of high consumption, we checked survey properties that used a lot of water for leaks. These checks revealed that some of the properties did not have leaks, with the high consumption resulting from actual use, and therefore that they should be included in the calculation of unmeasured PCC. The result of including these properties in the estimation of unmeasured PCC was to increase it by 3.0%.

Page 3.11

SWW Water Resources Management Plan 2015 – 2040

Given that high-consumption properties were not included in the estimation of unmeasured PCC before 2012/13, we have also calculated the PCC without them as this allows a better comparison with earlier data. The 2012/13 unmeasured PCC shown in Figure 3.4 above and used in the modelling described below is that which excludes these properties.

3.3.2.2 Estimation of historic measured PCC

Over recent years the proportion of household customers paying measured bills led us to start a measured household consumption monitor, which we now operate in parallel with the long-running unmeasured one. As for the unmeasured survey, we selected a sample of 1,000 properties and send member households water use questionnaires every few years.

While the data we obtain from the measured household consumption monitor is very useful in understanding the way in which customers use water, it is not the best source from which to obtain average measured PCC. Instead we use our billing data for this purpose, as this enables us to account for the consumption of the entire measured household population rather than the limited sample that are members of the consumption monitor. We know the total consumption of all the measured households from meter readings, and by dividing this by the estimated population of these properties we obtain average PCC. Our estimate of population used in this calculation comes from combining official ONS population figures with measured household occupancy rates obtained from the consumption monitor. 3.3.3 Normalising base year demand

The base year for the Plan is 2012/13, so we have based our analysis of PCC on the data we collected in that year. However this was an exceptionally wet year, so the first step in producing our forecasts was to estimate what the base year demand would have been had the weather been „normal‟.

We approached this using a statistical modelling approach, attempting to find a robust linear regression model that enabled us to describe historic PCC using a number of explanatory variables. By entering long-term average (LTA) weather values into such a model, we could estimate base year PCC had the weather been normal.

We produced two separate models, one for the historic average PCC of measured households and one for unmeasured ones, as it is likely that the different factors affect these two types of households to different degrees. The data we used in these models did not include allowances for meter under-registration or plumbing losses that were not recorded by a water meter.

The explanatory variables which we included in our modelling are described in Table 3.3.

Page 3.12

SWW Water Resources Management Plan 2015 – 2040

Table 3.3: The potential explanatory variables included in the modelling of PCC

Explanatory Notes variable Rainfall Rainfall has an important influence on the amount of water people use. We used South West England monthly areal rainfall from monthly hydrological summaries published by the Centre for Ecology and Hydrology (CEH). These monthly totals have been aggregated in a number of different ways including annual total, Apr-Aug, Apr- Sep, Mar-Aug, Mar-Sep and Jul-Sep. Sunshine hours As with rainfall, the amount of sunshine is likely to play an important role in water used. We used monthly sunshine hours data for a Camborne weather station which we obtained from the Met Office. These monthly totals have been aggregated in the same ways as the rainfall data. Year Including the year in the analysis may allow a way of accounting for factors that affect annual average PCC, but which are very difficult to quantify. For example, increased water efficiency in water-using appliances and a greater awareness of environmental issues are both likely to have reduced water consumption over the years. Average The AHS can be used to account for the demographic change that Household Size has occurred in both the measured and unmeasured population, both (AHS) as a result of general demographic changes and changes resulting from households switching from unmeasured to measured billing. Average customer Increasing bills are likely to lead to measured customers being more bill careful with the amount of water that they use and this may be important in modelling PCC.

We performed this analysis using the linear regression functions of the IBM SPSS Statistics software package, assessing many models for their suitability in describing the observed PCC data, and identifying the best ones.

3.3.3.1 Exclusion of 2012/13 data

Since completing our draft Plan, we have obtained weather data for 2012/13 that we could include in our analysis of how weather affects both measured and unmeasured PCC. However the summer of 2012 was very wet and the number of sunshine hours during the summer was very low. These two factors meant that when 2012/13 data were included in the modelling the year always appeared as an extreme outlier, and a suitable model fit could not be found.

The models that we developed for our draft Plan were linear regression models, which assume that the relationship between the explanatory variables and PCC is linear. Given that the weather variables we modelled for the draft Plan were all fairly close to LTAs the assumption that they have a linear relationship to PCC was reasonable. However the rainfall and amount of sunshine in the summer of 2012 were so far outside of normal conditions that it is not reasonable to assume a continuation of the linear trend between these factors and PCC. This is the reason that 2012/13 appeared as an outlier in the modelling.

Page 3.13

SWW Water Resources Management Plan 2015 – 2040

We therefore decided not to include the 2012/13 data in our models of how weather variables affect PCC, and continued with those we derived for the draft Plan. Even without the 2012/13 data included, these models, which are described below, allowed us to estimate normalised PCC in the base year.

3.3.3.2 Unmeasured PCC modelling

Figure 3.4 shows that up until the late 1990s unmeasured PCC followed a reasonably linear, upward trend, but after that the trend is much less clear. When we fitted regression models to this data, the models were strongly influenced by this linear trend so they did not fit the more recent data. As the purpose of this modelling is to normalise 2012/13 data it is very important the model developed fits the recent data well, therefore we decided to exclude the earlier data from the modelling.

We found that we obtained models with the greatest explanatory power when we used the 2002/03 to 2011/12 data, this ten-year period providing the best balance of excluding earlier data without removing too much information.

The model we found that explained most variance in unmeasured PCC used the total rainfall between April and September and the total sunshine hours between March to August (both expressed as a percentage of LTA) as explanatory variables. These two weather variables do not show significant collinearity, with a Pearson correlation coefficient of -0.14. The model explained 76% of the variance in unmeasured PCC (p=0.007). A summary of this model is given in Table 3.4.

Table 3.4: Summary of the unmeasured PCC linear regression model

Explanatory Estimate of Standard T-statistic Significance variable coefficient error Constant 99.025 25.805 3.837 - Rainfall (Apr-Sep) -0.166 0.049 -3.401 0.011 Sunshine hours 0.695 0.247 2.810 0.026 (Mar-Aug)

The model coefficients obtained act in the expected directions, with higher rainfall reducing PCC, but increased sunshine hours increasing it.

Surprisingly given the exceptionally wet summer with much less summer sunshine, this model suggested that had 2012/13 been a normal year the unmeasured PCC would have been 2.0% lower. However this result must be considered in parallel with the obtained PCC data itself, which is shown in Figure 3.4 above. The unmeasured PCC in 2012/13 determined from our consumption monitor was higher than it has been for several years. This is most likely the result of greater uncertainty with the number of unmeasured customers falling, and with it the number of properties in our survey which in 2012/13 stood at 674. Some evidence of this increasing uncertainty can be seen in the chart, which shows greater fluctuations in unmeasured PCC in recent years than was present in the past. This result therefore suggests that some of the increase in unmeasured PCC in 2012/13 is the result of the uncertainty in its estimation, and the 2% reduction removes the effect of this uncertainty to reveal the underlying trend.

Page 3.14

SWW Water Resources Management Plan 2015 – 2040

We are addressing this increasing uncertainty by installing more survey meters at unmeasured properties. During 2013 we installed a further 80 survey meters, and are continuing this programme of meter installation over the next few years.

3.3.3.3 Measured PCC modelling

When we fitted regression models to the measured data shown in Figure 3.4, none satisfactorily described PCC in the first two years (2000/01 and 2001/02), which always appeared as outliers. As there were low numbers of measured households in these first two years it may be that the PCC obtained had a large margin of error, or perhaps that these first meter optants behaved in a slightly different way to later optants. Because of this we decided to exclude these years from the modelling, resulting in a measured model that was based on the same ten-year period between 2002/03 and 2011/12 as the unmeasured one.

Unlike the unmeasured one, the model we found to explain most variance in measured PCC didn‟t include rainfall, the sunshine hours between April and August being the only weather related variable. A further difference from the unmeasured model is the inclusion of the year as a significant component, suggesting a reducing trend in measured PCC over the past 10 years. This model explained 98% of the variance in measured PCC (p<0.0005). A summary of this model is given in Table 3.5.

Table 3.5: Summary of the measured PCC linear regression model

Explanatory Estimate of Standard T-statistic Significance variable coefficient error Constant 5585.714 380.071 14.696 - Year -2.748 0.188 -14.598 <0.0005 April to August 0.522 0.105 4.959 0.002 sunshine hours

As with the unmeasured model, increased sunshine hours result in higher PCCs, while the negative value of the year variable indicates a reducing trend. The scale of this trend is such that measured PCC is now around 20% lower than it was a decade ago. The model suggested that had 2012/13 been a normal year the measured PCC would have been 0.5% higher. 3.3.4 Dry year demand in the base year

The household demand in a dry year is very important for us to understand, as it allows us to model the way in which our water resources system might cope during a drought. To enable us to study this dry year scenario we needed to estimate how much higher PCC would be in a hot, dry year than it would be in a normal one.

Page 3.15

SWW Water Resources Management Plan 2015 – 2040

The last hot and dry year in the South West was 19956, when only around 8% of our household customers had water meters. Since that time the number of customers paying a measured bill has risen steeply and now stands at over 75%. As there has not been a dry year since the level of metering has reached significant levels, we do not have evidence of how much water measured households will use in such a year. Therefore we inferred the increase in demand using other available sources.

In order to estimate dry year PCC, we used our PCC normalisation models, described above, to predict what PCC would have been in 2012/13 given the weather experienced in each year of the weather record (which ran from 1979 for the weather data used in the measured model and 1990 for the unmeasured). We then applied a Box-Cox transformation to these PCCs to give a normal distribution, and used this to estimate a demand for a 1 in 20 dry year. We chose a 1 in 20 dry year as this matches our temporary water use restrictions level of service.

This approach gave a figure of 8.6% for the increase in measured PCC between a normal and a dry year, and 8.7% for the unmeasured PCC. However, applying these factors gives the increase in household demand had 2012/13 been a dry year as being greater than the increase actually seen in 1995. Given the increase in the level of metering and greater water efficiency in 2012 compared to 1995, this result is unrealistic.

The reason for this can be traced to the data on which models were based, which came from a ten-year period between 2002/03 and 2011/12 containing no significantly dry years. While this doesn‟t pose a problem for normalising the 2012/13 PCCs, using this approach does assume that the linear relationships between PCC and weather are still valid when extrapolated beyond normal conditions. The unrealistic result obtained suggests that this assumption doesn‟t hold and that the models over-estimate dry year PCC.

We also used the models to derive 1 in 10 dry year PCCs, resulting in normal to dry year factors of 6.1% and 6.8% for measured and unmeasured customers respectively. Although these factors were derived for a 1 in 10 dry year, given the way in which the models over-estimate PCC under dry conditions, the results will represent a year drier than 1 in 10. Applying these factors to PCC results in a consumption increase in a dry year similar to that observed in 1995, suggesting that they are approximately correct. Given the lack of other evidence available on the dry year consumption of a largely measured population (who have significantly reduced their PCC over the preceding decade), we decided to use these factors to define a dry year. This lack of evidence will only be addressed when a dry year similar to that of 1995 is experienced. 3.3.5 Summary of base year PCC adjustment factors

A summary of the factors we have estimated to calculate the expected dry and normal year PCCs in the base year of the Plan, and the resultant PCCs, are shown in Table 3.6.

6 Although 2003 was a hot year, July was wet and this suppressed demand for that month, and to some extent August as well.

Page 3.16

SWW Water Resources Management Plan 2015 – 2040

The unmeasured PCC adjustment factors shown in this table have been calculated from our estimate of 2012/13 PCC excluding the effect of the high consuming properties mentioned in section 3.3.2.1. This is because we are deriving these factors from historic data which did not include such properties. However, the base year unmeasured PCC in Table 3.6 does include these properties, as applying the derived factors to this figure allows us to account for high consumption properties in our forecasts.

Table 3.6: Summary of base year PCC adjustment factors. The PCCs shown do not include the effect of meter under-registration or plumbing losses that were not recorded by a water meter.

Measured Unmeasured Adjustment Adjustment PCC PCC from base from base (l/hd/day) (l/hd/day) year year Base year - 117.1 - 164.9 Normal year +0.5% 117.7 -2.0% 161.6 Dry year +6.7% 124.9 +4.6% 172.5

3.3.6 PCC forecasts

We have used micro-component analysis produced for us by RPS Group Plc to estimate the way in which household water consumption will change in the future. The RPS report detailing the preparation of these forecasts can be found in Appendix D. The analysis was undertaken in accordance with the Water resources planning guideline7 and made extensive use of the UKWIR report A good practice manual and roadmap for customer household consumption forecasting8. As consumption patterns do not differ widely over the region we serve, forecasts have been produced at a company rather than WRZ level.

The first step in producing the PCC forecasts was to identify how customers currently use water in their homes in a normal year, and how this differs in a dry year. To do this we used several sources of information including Defra‟s Market Transformation Programme and questionnaire responses obtained from members of our household consumption monitor. We estimated how current existing measured, unmeasured, meter optant and new connection PCC was split between the six micro-components recommended in the Water resources planning guideline, detailed in Table 3.7.

7 Ibid. 4 8 UKWIR, “Customer behaviour and water use: A good practice manual and roadmap for household consumption monitoring”, Ref 12/CU/02/11, June 2012

Page 3.17

SWW Water Resources Management Plan 2015 – 2040

Table 3.7: The micro-components considered in producing PCC forecasts

Micro-component Description WC flushing - Machine washing only Clothes washing The small volume used for hand-washing is included in the miscellaneous internal component Shower Personal washing Bath Washbasin Dishwashing machine Dishwashing Washing-up by hand Garden watering (by hosepipe, sprinkler and watering can) Filling/topping-up of ponds and water features External use Pressure washers Recreational use (swimming pools, hot tubs and paddling pools) Including but not exclusively: Hand-washing of clothes Household cleaning Plumbing losses Miscellaneous internal use Running of taps whilst waiting for hot water Water softeners Waste disposal units Watering of indoor plants Use for pets

After estimating the split between these micro-components, we used a range of sources to forecast how the usage for each of them was likely to change in the future. Full details of the way in which we have done this are given in Appendix D.

Our PCC forecasts assume a continuation in existing metering, water efficiency and leakage policies, and increased water efficiency in new homes. This additional efficiency results from legislation such as the introduction of water efficiency standards into Part G of the Building Regulations and the Code for Sustainable Homes as well as improvements in technology.

Page 3.18

SWW Water Resources Management Plan 2015 – 2040

The forecasts show how we expect the consumption of current households to change over the period to 2039/40, given constant average household size (AHS). In the future however household sizes will change both through the expected continuation of the nationwide reduction in AHS and because smaller households are more likely to opt to switch to metered billing. As household size has an impact on PCC, with smaller households using more water per person on average than larger ones, any change in AHS must be accounted for. We have done this by applying correction factors to the forecast PCCs, with these factors calculated using data from our household consumption monitors showing how average PCC varies for different household sizes.

Micro-component analysis provides the best method of forecasting future PCC, but much of the evidence required for such an analysis, such as that provided by Defra‟s Market Transformation Programme, is only available at national level. We therefore had to ensure that the results of our analysis reflected observed consumption trends in the region we serve. Whilst the forecast unmeasured PCC agreed well with recent trends, the rate at which measured PCC has fallen was much greater than that predicted by the micro-component analysis. Whilst the normalisation analysis described in Section 3.3.3.3 identified an average annual reduction of 2.7% per year in measured PCC (as indicated by the value of the year coefficient in Table 3.5), the trend from the micro-component analysis was 0.8%.

This difference can be explained by considering how this region differs from the rest of England and Wales. Levels of metering in the South West have reached around 75%, much higher than the average for the rest of England and Wales. With so many of our customers therefore having a financial incentive to reduce water usage, some of the savings forecast to happen over the next 25 years are likely to occur more quickly in this region than in other parts of the Country. As a consequence we have assumed that the measured PCC reduction forecast by the micro-component analysis will be realised more quickly. We did this by increasing the rate of reduction in the first few years of the Plan to follow the observed trend, and reducing the water savings later in the Plan, leading to a PCC at the end of the planning period that matched that forecast by the micro-component analysis.

The effect of these adjustments is shown in Figure 3.5. No adjustment of the unmeasured PCC derived from micro-component analysis was required, as this followed the observed trend.

Page 3.19

SWW Water Resources Management Plan 2015 – 2040

Figure 3.5: The effect of adjustments to the micro-component analysis derived measured PCC to match the current trend

180 Historic Dry year initial forecast

160 Dry year final forecast

Normal year initial forecast

140 Normal year final forecast

120 Average Average PCC

100 (litres per per (litres person per day)

80

2017/18 2007/08 2012/13 2022/23 2027/28 2032/33 2037/38 2002/03

3.3.7 The effect of metering on household demand

Our forecasts have been prepared assuming that our current optional meter programme continues for the duration of the planning period. As metered customers have a financial incentive to reduce their water consumption, those who opt to have a water meter installed generally reduce their consumption. In preparing this Plan it was important to understand how much water customers save when they switch to metered billing.

The data we obtain from our unmeasured household consumption monitor allows us to compare total water use before and after a household switches to metered billing. We employed a statistical consultant, Explicata Ltd, to analyse the data we have collected from around 750 survey member properties over the past 18 years. The primary conclusion of this work was that the average reduction in water consumption as a result of metering is about 16.5%, which is sustained to the second year following the switch from unmeasured to metered status.

Explicata state that the extent to which such consumption reductions might be achievable in the remaining unmeasured part of the household population is uncertain, but that the analysis shows consistent reductions up to the most recent cohort of households switching. Given these conclusions we have assumed that opting household properties will reduce their consumption by 15% compared to their pre-opting usage.

Page 3.20

SWW Water Resources Management Plan 2015 – 2040

In addition to the benefit of reducing customer consumption, measured households on average suffer a lower level of leakage from their underground supply pipes then unmeasured ones. This is because any leaks on the section of underground supply pipe downstream of a meter are noticeable through the meter. In our forecasts we have therefore assumed that underground supply pipe leakage is reduced when customers switch to metered billing. Based on 2012/13 data from our annual performance report we assume that when a customer switches to metered billing, underground supply pipe leakage is reduced by 20% in the case of households and 32% for non-households. 3.3.8 The effect of climate change on household demand

For our draft Plan we based our assessment of the effect of climate change on household demand on Climate Change and the Demand for Water9. Since we produced this Plan UKWIR have published Impact of climate change on water demand10 containing new research. We have now updated our assessment of the effect of climate change on demand to use the information contained within this report. We have used the average of the effects predicted by the Thames and Severn Trent models and forecast that climate change will increase 2039/40 household consumption by 0.8%. This is a reduction from the impact used in our draft Plan, in which we assumed that the increase would be 1.1%. 3.3.9 Household consumption forecast

Our forecast of total household consumption is shown in Figure 3.6. We expect that initially the continued reduction in PCC will significantly exceed the additional demand from a rising population. This reduction in PCC is the result of more unmeasured customers switching to metered billing and further water savings from the existing measured customers. Later in the planning period there will be fewer remaining unmeasured households, so the rate of opting will slow significantly, while fewer opportunities for water saving will be available to measured customers. Because of this we predict that the increasing population will then drive an increase in demand, in spite of a continued reduction in average PCC.

Compared to our draft Plan, these latest forecasts show higher household consumption, with a 4.2% increase in the 2039/40 figures. The stronger population growth forecast, as described in section 3.2.3, explains the majority of this increase, with the other changes described above having a small effect.

9 Downing, T.E et al, Stockholm Environment Institute, “Climate Change and the Demand For Water”, February 2003. 10 UKWIR, “Impact of climate change on water demand”, Ref 13/CL/04/12, 2013.

Page 3.21

SWW Water Resources Management Plan 2015 – 2040

Figure 3.6: Forecast household consumption

280

Final plan - dry year Final plan - normal year 260 Draft plan - dry year Draft plan - normal year 240 Historic

220

200

Total household Totalconsumption (Ml/d) 180

160

2000/01 2005/06 2010/11 2015/16 2020/21 2025/26 2030/31 2035/36

3.3.10 Forecasting PCC in the future

An understanding of the way in which household customers use water within the home is an important part of forecasting their future consumption. To assist in the production of this Plan we have used data obtained from our household consumption monitors which we operate for both measured and unmeasured properties. We send members of these surveys questionnaires every couple of years asking about how they use water in their homes. The responses to these questionnaires provide information which is very useful in estimating micro- component consumption.

We must ensure that we are able to forecast future household demand for future plans, so we will continue to operate our household consumption monitors in the future. Data from these surveys will be combined with robust national data, such as UKWIR research and information published as part of Defra‟s Market Transformation Programme.

Page 3.22

SWW Water Resources Management Plan 2015 – 2040

3.4 Non-household consumption 3.4.1 Our approach to forecasting non-household consumption

The level of metering in our non-household customers has been high for many years and currently stands at over 90%. Because of this we have a good set of data from which we can gain an understanding of non-household consumption. Non- household consumption is heavily influenced by economic factors making econometric data useful in both explaining historic data and forecasting future consumption. In producing these forecasts we have used the report Forecasting water demand components11. Details of the way we did this are given below. 3.4.2 Historic consumption

In order to understand our non-household demand and the factors that affect it, we have undertaken log-linear regression based on 1991/92 to 2012/13 annual data. We analysed both industrial sector and total measured non-household consumptions, and found that the most statistically robust modelling was at the total level. The data we used for this analysis did not include allowances for meter under- registration or plumbing losses that were not recorded by a water meter. We found that the following were the best explanatory factors for measured non-household demand:

. The level of economic output in the area we serve measured in terms of Gross Value Added (GVA)

. Rainfall between May and September expressed as a percentage of LTA

. A trend term with 1991/92 taking value 1 and increasing by one for each subsequent year

This model explains 96% of the variance in consumption over the period of the analysis. A summary of the model is shown in Table 3.8.

Table 3.8: Summary of the non-household log-linear regression model

Explanatory Estimate of Standard T-statistic Significance variable coefficient error Constant 1.913 1.004 1.906 - Natural log of 0.300 0.103 2.902 0.010 economic output Rainfall (May- -0.078 0.018 -4.420 <0.0005 Sep) Trend -0.022 0.002 -7.892 <0.0005

11 Ibid. 2

Page 3.23

SWW Water Resources Management Plan 2015 – 2040

3.4.3 Base year demand in normal and dry year scenarios

By using LTA rainfall in the regression model derived above, an estimate of what base year demand would have been had 2012/13 been a normal year can be obtained. Similarly, by using 1995/96 rainfall data (which was the last significant drought in the South West) in the model, the consumption had it been a dry year can be estimated.

The base to normal year and base to dry year factors are shown in Table 3.9, along with the estimated measured non-household consumption under these scenarios.

Table 3.9: Summary of base year measured non-household adjustment factors. The consumptions shown do not include the effect of meter under- registration or plumbing losses that were not recorded by a water meter.

Adjustment Consumption from base (Ml/d) year Base year - 77.3 Normal year +4.0% 80.5 Dry year +6.5% 82.4

3.4.4 Forecasts of measured non-household demand

To enable us to use the model we derived to produce forecasts of measured non- household demand, we obtained a macroeconomic forecast from Cambridge Econometrics (CE) of employment and economic output within the Region. This forecast incorporates longer term structural trends in the economy and employment by industry.

We used the CE forecasts in the model to obtain a projection of total measured non- household demand, applying the percentage growth obtained from the regression to the base year demand. After 2023 we restricted the trend term within the regression to deliver a sustainable longer term 1.0% annual reduction in demand based on continuing water efficiency and industry composition changes.

Where we knew about changes to large user demands, we have applied these to the results of the regression model.

We have obtained from our billing system the number of void non-households in the base year of 2012/13; 2.4% of measured non-households were void, and we have assumed the same proportion of void properties throughout the planning horizon. 3.4.5 Forecasts of unmeasured non-household demand

Currently less than 9% of non-household consumption is by unmeasured customers. To forecast future unmeasured non-household consumption we assumed that usage by current unmeasured customers would change at the same rate as that of the measured ones. Over recent years an average of 1.3% per annum of remaining unmeasured non-household customers have switched to metered billing. We have assumed that this rate of opting will continue throughout the planning horizon.

Page 3.24

SWW Water Resources Management Plan 2015 – 2040

The proportion of unmeasured non-households that were void was 12.0% during 2012/13, much higher than the 2.4% void rate in measured properties. We have assumed these void rates for the duration of our forecasts, but due to the continuing migration of unmeasured properties to the measured category, the overall void rate drops slightly from 3.3% in the base year to 3.0% in 2039/40. 3.4.6 The effect of climate change on non-household demand

In our draft Plan we based our assessment of the effect of climate change on non- household demand on results from Climate Change and the Demand for Water12. Since our draft Plan was published the report Impact of climate change on water demand13 has been released. While the data available to the authors of this report did not contain clear evidence of the impact of climate change on non-household demand, it did indicate that the impacts on household demand will be only 49% of that predicted in the earlier study. We have therefore assumed that the increase in non-household demand attributable to climate change will be 49% of that used in our draft Plan. This equates to an increase in 2039/40 dry year non-household consumption of 1.4%. 3.4.7 Overall non-household consumption forecast

Our total non-household consumption forecast is shown in Figure 3.7. After the steep decline in consumption resulting from the continued economic weakness, we predict that consumption will continue the long term downward trend but at a slowing rate.

Our latest forecasts are 2% lower in 2039/40 lower than those used for our draft Plan as a result of updated econometric forecasts from CE predicting a slightly reduced long-term economic growth rate.

12 Ibid. 9 13 Ibid. 10

Page 3.25

SWW Water Resources Management Plan 2015 – 2040

Figure 3.7: Forecast non-household consumption

130

Final plan - dry year Final plan - normal year Draft plan - dry year 110 Draft plan - normal year Historic

90 householdconsumption (Ml/d)

- 70 Total non Total

50

2000/01 2005/06 2010/11 2015/16 2020/21 2025/26 2030/31 2035/36

3.4.8 Measured non-household consumption by industry sector

The major industries in our region are tourism and agriculture. Beyond these two, the composition of industry is diverse and approximately equally split between service and non-service industries.

The South West is known as a major tourist destination which contributes to the highly seasonal demand patterns that we experience in our region. This tourism helps to explain why the hotels and catering industry is one of our largest, accounting for around 18% of non-household consumption. Historically water consumption in this sector has remained relatively stable, without the large reductions observed in other industries. We expect this relative stability to continue in the future, with further water efficiency driving a small annual consumption reduction of about a quarter of a percent.

Agriculture in the region accounts for over 15% of current non-household consumption and is predominantly livestock based, with dairy, cattle and sheep farming forming the bulk. The water use of our agricultural sector is therefore far less reliant on the use of irrigation than is the case in eastern England. Consumption has fallen dramatically over the past couple of decades and is now around 48% lower than it was in 1990. The rate of this reduction has slowed in recent years, but from our regression analysis we predict that for the duration of the forecast a year-on-year reduction in consumption of around 2% is likely due to increasing water efficiency.

We expect water use in other commercial sectors to continue to decline in spite of forecast economic growth. This is due to the increasing use of water efficient practices and technologies as a method of reducing operating costs.

Page 3.26

SWW Water Resources Management Plan 2015 – 2040

A summary of our measured non-household consumption forecasts is shown in Table 3.10.

Table 3.10: Summary non-household consumption in the major sectors in our region

Forecast consumption (Ml/d)

2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

Hotels & catering 17.82 17.60 17.39 17.20 17.04 16.90

Agriculture 15.48 13.87 12.44 11.29 10.41 9.64 Education & health 10.08 8.94 8.16 7.54 7.17 6.89 Other non-service 21.13 18.73 17.10 15.80 15.02 14.45

Normal Normal year Other service 13.03 11.55 10.54 9.74 9.26 8.91 Total 77.55 70.68 65.62 61.59 58.90 56.79 Hotels & catering 18.00 17.78 17.57 17.38 17.22 17.07

Agriculture 16.09 14.42 12.94 11.75 10.83 10.04

Education & health 10.32 9.15 8.36 7.73 7.35 7.07

Other non-service 21.63 19.18 17.51 16.19 15.40 14.81 Dry year Dry Other service 13.34 11.82 10.80 9.98 9.49 9.13 Total 79.38 72.35 67.17 63.04 60.29 58.12

3.5 Leakage 3.5.1 Determining base year leakage

Our leakage control is based on continuous monitoring of night flow data in small areas of on average 1,000 properties known as District Metering Areas (DMAs). We calculate the level of leakage by analysing DMA night flows, from which we subtract the usage of large measured customers and assessed domestic and commercial night use of the properties in the area. We then take the 27th percentile value of all the overnight readings to calculate the leakage for a particular month. Our reported annual leakage is an average of all twelve months of the year, without the removal of summer months.

We have approximately 2,200 meters collecting continuous 15-minute data with 99% of this data being transmitted through telemetry. This allows us to quickly review data and reduce the time it takes us to become aware of network problems. The flow data is automatically imported into our Leakage Analysis Software System (LASS) which provides reports on DMA prioritisation, data collection problems and is the reporting tool for regulatory returns.

Page 3.27

SWW Water Resources Management Plan 2015 – 2040

As part our comprehensive leakage control strategy we monitor losses from service reservoirs annually by comparing inlet and outlet flows at each reservoir. This method has the benefit of recording all losses associated with the reservoirs, whether from overflows, structural seepage or leaks in the mains. It also avoids the operational disturbance and risk to security of supply involved in static drop testing (where inlet and outlet valves are closed and the reservoir level is monitored to see if it falls). We currently estimate losses from service reservoirs to be 3.9 Ml/d.

For leakage reporting DMAs are aggregated to WIS zones and then summed to a regional figure. As WRZs comprise a number of WIS zones it is also easy to report leakage within each of our three WRZs. 3.5.2 Determining the sustainable economic level of leakage

Our leakage strategy optimisation model (based on the MELT principles) has continued to evolve incorporating guidelines recently identified in the Review of the calculation of sustainable economic level of leakage and its integration with water resource management planning14. In our model the region has been considered as six leakage zones each consisting of a group of WIS zones. These leakage zones are based on treatment works and the resources that supply them. For each zone we have modelled costs for leakage detection down to WIS zone level.

We have continued to develop an improved approach to costing socio-environmental aspects and have undertaken a cost/benefit analysis for reducing (customer) supply pipe leakage.

The underlying economic principles incorporated in our Sustainable Economic Level of Leakage (SELL) model are:

. It is based upon the principal of a Natural Rate of Rise (NRR) of leakage which is an estimate of how quickly leakage would rise if no leakage control activity was undertaken. The NRR in different areas will vary, and we have calculated an estimate of NRR for each WIS zone in our supply area. As property numbers change over time, the NRR will also change, ie a rise in the number of connected properties will tend to increase leakage.

. As the level of leakage is reduced, the cost of leakage control activity increases.

. Lower leakage levels reduce demand and thus reduce marginal operating costs.

. Over time improvements in leakage detection techniques are likely. Our model assumes a 1% per annum net reduction in costs for leakage detection.

14 Strategic Management Consultants, “Review of the calculation of sustainable economic level of leakage and its integration with water resource management planning”, October 2012

Page 3.28

SWW Water Resources Management Plan 2015 – 2040

. Recent and future improvements in pressure management have potential for reducing both the frequency and the severity of leaks. Although the exact relationship in complex distribution networks between pressure and leakage is not easily predicted, it is generally accepted that the two elements can reasonably be regarded as being directly proportional to one another. The average zonal night pressure (AZNP) in 2006, the year that our comprehensive NRR study was undertaken, was 58.6m and the company‟s target for the year 2019/20 is 48m. In our model, the effects of pressure reduction have been used to rescale both the NRR and the Policy Minima by the applicable pressure reduction.

. Environmental and social costs generally (but not always) tend to promote leakage detection and repair over increasing output. Willingness to pay surveys have been used to cost supply interruptions due to leakage repairs, and we have used values from the Office of National Statistics for the costs of vehicle delays, which result from leakage control roadworks.

. Our model incorporates a Marginal Value for Water (MVW) to estimate a value to society and the environment for water abstractions that subsequently contribute to leakage (above policy minimum). We have modelled for a range of MVW values from 0 to 30 pence/m³ of leakage above policy minimum. Whilst not specific to our resources the results do encompass the full (known) range of values estimated by other parties, and so allow us to make an informed decision in setting our target leakage level.

. Costs for carbon are fully included in our company unit production costs through the European Union Emissions Trading Scheme (EUETS) applied to the cost of energy, as recommended in the UKWIR report A Framework for Accounting for Embodied Carbon in Water Industry Assets15.

. All costs/benefits have been scaled to their 2012 values using the Retail Price Index (RPI) measure of inflation.

As in our 2009 Water Resources Plan (WRP09)16 our model has enabled us to produce a range for the SELL which is shown in Figure 3.8, along with our forecast of leakage. The upper limit of the band reflects our company costs and the social costs of the supply interruptions and road traffic delays associated with leakage control. The lower limit also reflects the additional MVW of 30 pence/m³ of leakage over policy minimum.

15 UKWIR, “A Framework for Accounting for Embodied Carbon in Water Industry Assets”, Ref 12/CL/01/15, 2012 16 South West Water, “Water Resources Plan 2010-2035”, November 2009

Page 3.29

SWW Water Resources Management Plan 2015 – 2040

Figure 3.8: Results of SELL modelling, showing the effect of a range of values of MVW

130

0 120 1

110

5 100

10 15 90 20 25 Leakage Leakage (Ml/day) 30

80

SELL range 70 SELL at specified MVW (pence/m³) Forecast

60 2012 2017 2022 2027 2032 2037

We estimated the costs of a policy of performing free replacement of leaking customer supply pipes, as this would encourage customers to report suspected leaks and would speed up repairs. Our analysis estimated a potential reduction in customer supply pipe leakage of 9.7 Ml/d at a cost of just over £4m per year. For comparison, this is more than ten times the cost of treating of a similar volume of water at one of our works. It is also more than half the annual leakage detection and repair budget, which enables us to find and fix around eight times as much leakage (that results from NRR). 3.5.3 Leakage forecast

Our leakage target since 1999/00 has been 84 Ml/d and we have met this target each year. Although guidance for the calculation of SELL has evolved over the intervening years, it has always been the case that our target is below this economic level. The calculation above shows that we are currently operating significantly below the SELL and that our current target will remain below it for the duration of the planning period.

Page 3.30

SWW Water Resources Management Plan 2015 – 2040

It is a Government aspiration that total leakage will not rise17 and following our extensive programme of research and engagement activities with our customers18, we will be reducing our total leakage from 84 Ml/d to 64 Ml/d over the planning period. However, this reduction is not considered cost effective or affordable at this time, so it will take place between 2020 and 2040, with leakage staying at 84 Ml/d until 2020. Full details of this planned reduction can be found in our Business Plan which was published on 2 December 2013.

This reduction in leakage is driven by customer preference and Government aspirations. Even without this reduction we forecast a supply demand surplus for the duration of the Plan, and therefore there is no supply demand driver. Therefore we have included this reduction in our baseline forecasts. 3.5.4 Meeting our leakage target

As our leakage target is below the SELL, maintaining this level has been extremely challenging and has required us to manage leakage control operations in the most efficient way. With the continued housing growth and the resultant expansion of our existing network, maintaining leakage at 84 Ml/d has required reductions in both the average leakage per property served, and the average leakage per kilometre of main. As our 84 Ml/d target for the future remains below the economic level, this level of challenge will remain, intensifying as we begin to reduce leakage towards 64 Ml/d.

This section details the ways in which we will continue to achieve the reductions to average leakage per property and average leakage per kilometre of main that are required to maintain total leakage of 84 Ml/d until 2020, and reduce it thereafter.

3.5.4.1 Leakage control

Our current leakage control is based on detection and repair, with constant development undertaken to optimise the processes that support these activities. Recent developments have included:

. The critical appraisal of the design and performance of existing DMAs with the aim of improving both data integrity and leakage monitoring. This continues to drive new meter installations and the replacement of poorly-performing existing meters. We have also sub-divided larger DMAs into smaller areas to localise leakage and improve the efficiency of our leak detection activity.

. During 2011/12 all leak detection staff were provided with mobile devices which allows work to be assigned remotely from depots to field-based staff and detection staff to capture and raise jobs remotely. This significantly reduces the time taken from detection of a leak to the job being assigned to our repair partners.

. We have increased our use of telemetry for remote data collection to cover 99% of our metered flows, which has significantly improved our awareness of leak events. Combined with strict repair partner Key Performance Indicator (KPI) targets, this has reduced the runtime of leaks, and therefore their potential to impact customer service.

17 Environment Agency, Ofwat, Defra and the Welsh Government, “Water resources planning guideline: The guiding principles for developing a water resource management plan”, June 2012 18 South West Water, “Customer & stakeholder research & engagement”, December 2013

Page 3.31

SWW Water Resources Management Plan 2015 – 2040

. Our LASS software which was introduced in 2004/05 continues to be developed, improving the targeting of our leak detection resource and the reporting of leakage performance.

DMAs cover 100% of our network, ensuring that all parts of the distribution system are closely and continuously monitored throughout the year. Every DMA has an individual leakage target set which reflects both our leakage target and the unit cost of leak detection in each DMA. DMAs with a high level of leakage or high unit cost are reviewed and where necessary, we redesign them to improve active leakage control targeting. A continuous picture of the distribution of leakage across the region is therefore available, which we can use to direct leakage detection staff.

The effectiveness of leak detection staff is essential to successful leakage control especially given the intensive level of activity required to operate below the SELL. All leak detectors receive comprehensive training and are provided with a wide range of current equipment. We monitor detection performance with monthly reviews and corrective action plans are instigated where appropriate. Equipment is regularly reviewed and trialled to ensure that staff are provided with the most efficient and accurate means of locating leaks.

The speed and quality of leak repairs is equally important and applies to both reported and detected leaks. We set KPIs for leak repair times and these are banded according to the assessed severity of the leak and the potential of disruption to customer service. Leak repair time performance is reported monthly and we produce a daily report to ensure that the number of leak repairs waiting to be carried out is within agreed service levels.

Since 2007/08 we have operated a policy for the opportunistic renewal of leaking communication pipes where leaks are from galvanised and black plastic pipework rather than joints. This avoids repeated repairs of defective communication pipes and is undertaken wherever Street Works Notices permit the additional work to be undertaken.

3.5.4.2 Customer supply pipe leakage

Our customer supply pipe repair policy since 2005 has provided one hour‟s free detection for commercial customers and contributions towards the cost of the repair or replacement of supply pipes for all privately owned domestic properties. The contributions are weighted in favour of replacement to encourage customers to replace pipework in which deterioration has occurred and which is likely to suffer repeat failure. This has been very successful with the majority of contribution claims now being for supply pipe replacements and it has markedly increased the rate of service pipe asset renewal. The contributions are conditional on the work being completed within the period of a waste notice which is issued for all supply pipe leaks; this encourages timely repair. The increase in household metering is assisting the identification of supply pipe leaks, as 85% of meters are installed in the highway.

We estimate that 80% of leakage is from our distribution network and 20% from customer supply pipe leakage.

Page 3.32

SWW Water Resources Management Plan 2015 – 2040

3.5.4.3 Pressure management

We undertake a continuous programme of pressure logging resulting in a 5-year recording cycle. This gives us a continuous update of the average zone pressures within each Pressure Managed Area (PMA) in the region. Results are aggregated to give a property weighted average zone pressure for each DMA, which in turn are aggregated to depot and a regional average zone pressure. The information we collect also enables us to update DMA Hour to Day factors, which we use in our leakage analysis.

Since 2010 we have been undertaking a specifically funded pressure management initiative which aims to reduce our Average AZNP by 10m by 2020. Our current assessment of the regional AZNP is 59m, which according to Water UK Leakage Performance Indicators is among the highest in the industry. Efforts to reduce pressures are being targeted to areas by combining a number of factors such as excess operating pressures, high burst frequency indexes, other supply disruptions and elevated leakage rates. The projected long-term benefits from the initiative will be a reduction in burst rates on both mains and communication pipes, a reduction in leak detection and repair costs, a reduction in DI and improved security of supply for customers.

3.6 Other components of demand 3.6.1 Water taken unbilled

Water taken unbilled can be taken both legally and illegally. Over 90% of the water taken legally unbilled is used in the operation of our waste water treatment works, the small remainder includes water used for fire fighting and highway washing. Examples of illegal use are connections that have been made to our distribution system without permission and consumption at void properties which have been occupied without us having been informed. Where we have evidence of water being taken illegally, we investigate and bring prosecutions where necessary.

We have assumed that there will be a slight drop in the amount of water taken unbilled as consumption at void properties will fall. This is the result of more properties becoming measured, allowing us to easily identify and bill for water that has been used. 3.6.2 Distribution system operational use

This component of demand covers the water that we use in the operation and maintenance of our distribution system for purposes such as mains flushing and service reservoir cleaning. We have assumed that the volume of water we use for these purposes will remain at the current level for the duration of the planning period.

Page 3.33

SWW Water Resources Management Plan 2015 – 2040

3.6.3 Overall forecast of other components

The forecast of the total of these other components of demand are shown in Figure 3.9. The chart does not show both dry and normal year forecasts as the only difference between them is the additional consumption of occupied void properties during a dry summer, which is very small. Our forecast of other component consumption has risen slightly from our draft Plan. This is because further analysis of the amount of water used in our waste water treatment works has indicated that this is higher than previously estimated.

Figure 3.9: Forecast of other components of demand

26

23

20

Historic

17 Final plan forecast Other consumptionOther (Ml/d) Draft plan forecast

14

2001/02 2006/07 2011/12 2016/17 2021/22 2026/27 2031/32 2036/37

3.7 Total demand 3.7.1 Summary of forecast demand

Our demand forecast has been prepared assuming that we continue our current optional metering programme, maintain our current leakage target of 84 Ml/d until 2019/20 then reduce leakage to 64 Ml/d over the next 20 years and continue to meet our current water efficiency target of 0.75 Ml/d per annum. We do not envisage that the total demand will be affected due to any changes brought about by retail licensees under the Water Supply Licensing regime.

We predict that total demand will continue to fall, initially driven by water savings in both the household and non-household sectors, and after 2019/20 by the planned reduction in leakage from 84 to 64 Ml/d.

Page 3.34

SWW Water Resources Management Plan 2015 – 2040

Our forecast is close to that presented in our draft Plan until 2019/20, after which the planned reductions in leakage lead to demand falling below the level previously presented.

With demand in a normal year falling from 418 to 363 Ml/d and average PCC from 136 to 112 l/hd/d, this Plan meets Defra‟s water efficiency and demand management aspirations set out as part of the planning guideline19.

Our total forecast is shown in Figure 3.10.

Figure 3.10: Total demand forecast 490 Historic Final plan - normal year 460

Final plan - dry year

Draft plan - normal year 430 Draft plan - dry year

400 Totaldemand (Ml/d)

370

340

2000/01 2005/06 2010/11 2015/16 2020/21 2025/26 2030/31 2035/36

3.7.2 Water efficiency activity

The forecasts we have produced assume a continuation of our base water efficiency activity, resulting in a saving of 0.75 Ml/d for each year of the Plan. These savings will be achieved by continuing our current activities with both household and non- household customers. These activities include:

. Guidance via our water conservation helpline, our website, face to face water audits, talks to special interest groups and events such as country and county shows.

. Supporting schools with educational water efficiency tools available via our website and with talks on request.

. Targeted promotions to our region‟s gardeners of discounted water butts advertised through our web site and through bill message promotions.

19 Ibid. 17

Page 3.35

SWW Water Resources Management Plan 2015 – 2040

. The promotion of free water saving devices for household customers to self select via our water conservation website.

. The provision of a series of online tools to help our business customers, including a water efficiency calculator and embedded carbon calculator, together with an option for customers to log and monitor their daily consumption.

. Undertaking water use reviews with business customers, including support in funding proposals that pay back in less than 12 months.

. Undertaking water use reviews and free device installation for customers who struggle with water affordability.

. Undertaking water efficiency reviews targeted at customers within the hospitality sector during key periods of the year. These businesses will also be subject to the general water efficiency activity that we undertake in both the household (many small B&Bs are billed as households) and non-household sectors.

. While the economic incentive to save water is greater for metered customers, the services listed above are also available to unmeasured customers to help reduce their consumption. We also provide a calculator tool to help unmeasured customers evaluate their water use. This is particularly helpful for those considering switching to a meter.

In addition we will work to enhance the national evidence base for water efficiency by our involvement in water efficiency research and trials and engagement with appropriate industry bodies. 3.7.3 Profile of annual demand

We have used weekly demand profiles in the supply demand modelling to allow us to account for the high seasonal variation; these profiles are shown in Appendix B. Different profiles have been used for the various parts of our region to reflect the difference in peak demands observed.

The profiles we have used are the same as those we used in our WRP0920. Whenever we experience high levels of demand during hot, dry periods such as in July 2013, we compare our profiles (both at WIS zone and WRZ level) with actual distribution input data to ensure that they adequately reflect the demand experienced. All the comparisons that we have performed since our WRP09 was published have shown the profiles to be valid, both in terms of overall level of demand and the seasonal pattern of consumption. We are therefore confident that it is still appropriate to use these demand profiles in our WRMP14 supply demand modelling.

We can occasionally experience higher levels of demand for short periods during the winter as a result of freezing and subsequent leakage. However, in all of our resources zones, this level of demand has historically always been lower than the summer peak we plan for in a dry year. Given the nature of our water resources, these short periods of high winter demand have no impact on our estimation of the Deployable Output (DO). Therefore for the purposes of this Plan we represent winter demand by more typical values.

20 Ibid. 16

Page 3.36

SWW Water Resources Management Plan 2015 – 2040

3.8 Derivation of weighted average demand forecasts

Ofwat require the estimation of weighted average consumption forecasts to be used by them in setting price limits. In the past price limits have been set according to normal year demand, but in moving to weighted average demand Ofwat have recognised that the long-term average demand may not be equal to the normal year demand. Two possible reasons for this are:

. Demand doesn‟t drop as much in a wet year as it rises in a dry one. If this is the case the long-term average demand will be higher than the normal year demand.

. More years are wet than are dry, or vice-versa.

We have considered historic weather patterns and demands and have found no evidence to suggest that the long-term average demand would differ from that expected in a normal year. This may partly be due to our highly seasonal demand which does drop significantly in a wet year compared to a dry one, meaning that wet year reductions in demand are comparable to increases in dry years.

We have therefore used normal year demands to represent the weighted average.

Page 3.37

SWW Water Resources Management Plan 2015 – 2040

4 Climate change

4.1 General

Our Water Resources Management Plan (WRMP) includes an assessment of climate change and its effects on the supply demand balance. We include a direct assessment of climate change on both our resources and demand. The uncertainties associated with estimating the impacts of climate change on both supply and demand are included in the assessment of headroom uncertainty.

4.2 Supply

We have assessed and reported in our Plan the likely implications of climate change on the Deployable Output (DO) of our resources.

4.2.1 Climate change vulnerability

As advised by the Water resources planning guideline1, before undertaking the above work, we agreed the vulnerability of our sources to climate change with the Environment Agency and shared with them our assessment of the vulnerability of our system to climate change. We concluded that our system falls into the “low vulnerability” category as set out in the above guideline. A copy of our report is given in Appendix F. The Environment Agency confirmed that our approach is reasonable.

4.2.2 Assessment of the impacts of climate change on river flows

In summer 2009, Defra published revised climate change projections known as the UKCP09 scenarios2. The Water Companies worked with the Environment Agency to research and develop suitable methods to use these new projections for water resources management plan assessments3. Guidance is given within the Water resources planning guideline4. The appropriate level of climate change assessment is influenced by the output of the vulnerability assessment.

Using the above information and our knowledge of our water resources, we carried out an analysis of the two approaches which we could use within this Plan: Approach 1.2 (UKCP09 flow factors) and Approach 1.4 (Future Flows).

We shared a copy of our report containing the results of our analysis with the Environment Agency and have included a copy in Appendix F.

1 Environment Agency, “Water resources planning guideline – interim update”, October 2012 2 Defra, “Adapting to climate change - UK climate projections 2009”, June 2009 3 Environment Agency, “Climate change approaches in water resources planning – Overview of new methods”, Ref SC090017/SR3, 2012 4 Ibid. 1

Page 4.1

SWW Water Resources Management Plan 2015 – 2040

4.2.3 Assessment of the impacts of climate change on groundwater resources

We discussed with the Environment Agency the impacts of climate change on our groundwater resources and concluded that this is only of significance in Wimbleball Water Resource Zone (WRZ), given the very few groundwater sources in the other WRZs. We also confirmed that within Wimbleball WRZ, groundwater sources are not dominant compared to surface water contribution.

On this basis, an approach was adopted which reflected the level of risk posed from climate change on our groundwater sources. The majority of groundwater abstractions occur from the Otter Sandstone aquifer. The risk from climate change impacts on Otter Sandstone sources is taken to be limited due to the high groundwater storage capacity strata. This assumption was tested through consideration of the impact on the Dotton boreholes, which are located in the Otter Valley. Whilst most of the sources in the Otter Valley are not constrained by groundwater level, the DO of the Dotton boreholes could be affected should groundwater levels be reduced to a significant degree. The Dotton sources were analysed through the use of a lumped parameter spreadsheet tool in conjunction with output from the newly developed Otter Valley Groundwater Model.

In addition, particular attention has been given to the impacts of climate change on the Otterton 1A borehole where the abstraction is controlled by groundwater levels in a local monitoring borehole. Relatively small changes in groundwater level can have a disproportionate influence on this borehole’s ability to abstract. Otterton 1A is further at risk from climate change as the source has a coastal location and is vulnerable to sea level change.

For those sources located outside the Otter Valley and not covered by the model, the spreadsheet tool alone was used. These were sources abstracting from the Upper Greensand aquifer where the storage capacity is lower compared to the Otter Sandstone aquifer.

A detailed description of the assessment of climate change impacts on our groundwater sources is set out in Appendix F.

4.2.4 Assessment of the impacts of climate change on DO

We have carried out our assessment of the impacts of climate change on DO in accordance with the Water resources planning guideline5. The guideline identifies four key stages:

(a) Stage 1 - Calculate river flows and/or groundwater recharge levels

Using the information from above, we perturbed historic time series sequences of river flows for each WRZ to cover the five scenarios identified within the UKCP09 flow factors – the 5th, 25th, 50th, 75th and 95th percentiles.

Using the information from our assessment of the impacts of climate change on groundwater resources as above, we estimated the impacts of climate

5 Ibid. 1

Page 4.2

SWW Water Resources Management Plan 2015 – 2040

change on our groundwater sources when they are used conjunctively with our surface water sources.

(b) Stage 2 – Calculate DO

As identified in the Water resources planning guideline6, we needed to select one of the UKCP09 flow factors scenarios to represent the best estimate of the impacts of climate change on baseline DO. In our case we believe that the 50th percentile flow factors are the most appropriate and representative of the possible range of scenarios presented in the UKCP09 flow factors and we discussed this with the Environment Agency.

Flows covering the other scenarios from the UKCP09 flow factor work were used to develop the climate change uncertainty distribution used in the headroom uncertainty assessment.

(c) Stage 3 – Scale the impacts of climate change from the base year to 2040

We needed to scale the change in DO calculated in Stage 2 above, for each year of the planning period. However, given the factors are derived for the 2020s, we extrapolated the impact on climate change with care and hence appropriately adapted the methodology set out in the guideline to fit our circumstances.

The results of the impact on DO are input into the water resources planning tables (WRP1 BL supply).

It should be noted that although climate change causes an adverse effect on DO throughout the planning period, infrastructure constraints may influence this impact.

(d) Stage 4 – Determine the uncertainty associated with climate change for inclusion in target headroom

In the consideration of climate change there is inevitably a degree of uncertainty. This is accounted for within the target headroom calculations as set out in the Water resources planning guideline7. Further details are presented in Appendix A which covers headroom.

4.3 Demand

We have presented details of how we have included the impacts of climate change on demand in Section 3.

4.4 Impact on supply demand balance

We have shown the impacts of climate change on our DO / WAFU in the WRP1a tables.

6 Ibid. 1 7 Ibid. 1

Page 4.3

SWW Water Resources Management Plan 2015 – 2040

5 Target headroom

5.1 Method

5.1.1 Target headroom

Target headroom has been defined as:

“the minimum buffer that a prudent water company should allow between supply (including raw-water imports and excluding raw-water exports) and demand to cater for specified uncertainties (except those due to outages) in the overall supply demand resource balance”.

We endorse this definition of target headroom which appeared in a UKWIR/ Environment Agency report1. Since the publication of this report further work has been carried out on headroom, leading to the publication of another report2. This second report presents a methodology focused on headroom uncertainty which can be used as a basis for target headroom. It is this second report that we have used for the calculation of target headroom in this Plan.

5.1.2 Calculation of target headroom

We have calculated headroom uncertainty in each of the three Water Resource Zones (WRZs); these calculations are set out in detail in Appendix A. The following components of headroom uncertainty are included in the methodology:

Supply related . Vulnerable surface water licences S1 . Vulnerable groundwater licences S2 . Time limited licences S3 . Bulk imports S4 . Gradual pollution causing a reduction in abstraction S5 . Accuracy of supply-side data S6 . Uncertainty of impact of climate change on source yield S8 . Uncertain output from new resource developments S9

Demand related . Accuracy of sub-component data D1 . Demand forecast variation D2 . Uncertainty of impact of climate change on demand D3 . Uncertain outcome from demand management measures D4

1 UKWIR/Environment Agency, “A Practical Method for Converting Uncertainty into Headroom”, Ref WR-13, January 1998 2 UKWIR, “An Improved Method for Assessing Headroom”, Ref 02/WR/13/2, 2002

Page 5.1

SWW Water Resources Management Plan 2015 – 2040

The methodology states that the uncertainty associated with time limited licences (S3), “will quite legitimately be reflected in Headroom Uncertainty” (p21). However the Environment Agency’s Water resources planning guideline3 states “companies should not make allowances in their headroom calculations for the risk of time- limited licences not being renewed”. Therefore we have excluded this risk from headroom uncertainty.

In Appendix A we describe how we have applied the Environment Agency guideline to the headroom uncertainty values in order to derive target headroom.

5.1.3 Available headroom

The available headroom in a WRZ is defined as the difference between the Water Available for Use (WAFU, which is Deployable Output (DO) including raw water imports less raw water exports, less outage) and the Dry Year Annual Average Unrestricted Daily Demand. If the available headroom is predicted to be less than the target headroom, then we should take action to avoid the risk of failing to meet our chosen level of service. However it should be noted that target headroom should be regarded as a guide only.

5.1.4 Security of Supply Index

It is our policy to maintain a surplus of WAFU over projected demand plus target headroom, thereby achieving the maximum Ofwat Security of Supply Index. This Plan sets out how we will achieve this.

5.2 Target headroom

5.2.1 Approach to target headroom determination

We have applied the UKWIR Methodology4 identically to all three WRZs to calculate headroom uncertainty. We have used water resources modelling to determine WAFU up to 2039/40 and then compared this to the target headroom plus Dry Year Annual Average Unrestricted Daily Demand to show the supply demand balance over the planning horizon.

The starting value for target headroom was that determined by the headroom analysis carried out for the 2009 Water Resources Plan (WRP09)5. We reviewed this previous analysis to ensure that:

. there were no new areas of uncertainty . uncertainties identified previously were still relevant . uncertainties identified previously were modified if appropriate.

Our conclusion was that there were no new areas of uncertainty, some areas of uncertainty identified previously were no longer relevant and some required modification to the probability distributions.

3 Environment Agency, “Water resources planning guideline – interim update”, October 2012, page 95 4 Ibid. 2 5 South West Water, “Water Resources Plan 2010-2035”, November 2009

Page 5.2

SWW Water Resources Management Plan 2015 – 2040

5.2.2 Reduction in uncertainty since WRP09

At WRP09 the following two options were included in the Plan, which are equivalent to introducing new sources:

. water efficiency options . new tariff structure

Uncertainties in the estimates of the impact of these two options on Distribution Input were included in the target headroom component S9 at WRP09.

For this Plan there are no new sources proposed and hence no S9 component of headroom uncertainty is included.

5.2.3 Headroom components relevant to the current Plan

With the removal of a number of potential uncertainties (both S9 components), we consider that only the following components are relevant to the current Plan:

Supply related . Accuracy of supply-side data S6 . Uncertainty of impact of climate change on source yield S8

Demand related . Accuracy of sub-component data D1 . Demand forecast variation D2 . Uncertainty of impact of climate change on demand D3 . Uncertain outcome from demand management measures D4

We have analysed these components using the same approach as at WRP09. The detail of each method is described in Appendix A.

5.2.4 Target headroom and sustainability reductions

As recommended in the guideline6, we make no allowance for sustainability reductions in target headroom.

5.2.5 Target headroom and the level of uncertainty

The use of a probabilistic approach to headroom requires that a balance is made between the costs and risks to customers and the environment afforded by a low headroom allowance against those of a high headroom allowance. This involves judgement of an appropriate level of confidence to attach to the target headroom uncertainty. We have determined this level by reference to the Environment Agency’s guideline7, which makes it clear that:

. the Environment Agency does not expect companies to plan for 100% certainty . the Environment Agency does not expect companies to apply too low a target headroom

6 Ibid. 3, page 43 7 Ibid. 3, page 94

Page 5.3

SWW Water Resources Management Plan 2015 – 2040

. water companies should accept a higher level of risk in future than at present.

For this Plan we have adopted an 85% confidence level at the beginning of the planning period, reducing to 75% by 2025 and down to 70% by 2040.

5.2.6 Target headroom and the impact of climate change

We have used a Monte Carlo approach to the assessment of target headroom in accordance with the guideline8. This produces a joint probability distribution by combining individual probability distributions in a stochastic manner. Therefore the isolation of an element of target headroom associated with an individual risk can be misleading. The sum of headroom values calculated from individual Monte Carlo simulations of sub-groups of headroom components is unlikely to be equal to one headroom calculation containing all the components.

In order to estimate the impact of climate change uncertainty on target headroom, we have calculated target headroom for all components and also separately for groups of components (ie with / without climate change) to determine the proportion of target headroom that the climate change components contribute. The results are shown in Table 5.1 below. A chart showing the headroom calculated for each group of components and for all components is shown in the individual section for that WRZ in Appendix A.

Table 5.1: Estimated effect of climate change on target headroom

Increase in target headroom due to WRZ climate change 2015/16 2025/26 2039/40 Ml/d 1.7 2.8 3.0 Colliford % 16.2 31.6 33.1 Ml/d 2.0 3.3 3.4 Roadford % 14.3 27.6 28.7 Ml/d 0.6 1.2 1.4 Wimbleball % 14.7 29.5 31.5

8 Ibid. 3, page 59

Page 5.4

SWW Water Resources Management Plan 2015 – 2040

6 Water resources strategy 6.1 Baseline supply demand balance 6.1.1 Colliford WRZ

Figure 6.1 below shows how forecast demand plus target headroom in Colliford Water Resource Zone (WRZ) compares to the Water Available for Use (WAFU).

Whilst WAFU does fall slowly due to climate change, falling demand leads to the WAFU remaining above demand plus target headroom throughout the planning period. This surplus means that no options are required in Colliford WRZ over the planning period.

Figure 6.1: The baseline supply demand position in the Colliford WRZ 200 WAFU Forecast demand + target headroom

180 Forecast dry year demand

160

140 Megalitres per Megalitresper day

120

100

2039/40 2012/13 2014/15 2019/20 2024/25 2029/30 2034/35

6.1.2 Roadford WRZ

In addition to WAFU falling as a result of climate change, we have also made an assessment of the effect of sustainability reductions proposed by the Environment Agency to take effect in the years 2015/16 and 2016/17. In spite of the reductions in WAFU resulting from these proposals, the WAFU in this WRZ remains comfortably above demand plus target headroom throughout the planning period (Figure 6.2). Therefore no options are required in Roadford WRZ over the planning period.

Page 6.1

SWW Water Resources Management Plan 2015 – 2040

Figure 6.2: The baseline supply demand position in the Roadford WRZ 320 4.8 Ml/d Sustainability Reduction WAFU identified in our WRP09 Forecast demand + target headroom Forecast dry year demand 280 2.16 Ml/d Sustainability

Reduction

240

Megalitres per Megalitresper day 200

160

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

6.1.3 Wimbleball WRZ

As in Colliford WRZ, no sustainability reductions have been proposed by the Environment Agency for Wimbleball WRZ. WAFU remains above demand plus target headroom throughout the planning period (Figure 6.3). Therefore no options are required in Wimbleball WRZ over the planning period.

Page 6.2

SWW Water Resources Management Plan 2015 – 2040

Figure 6.3: The baseline supply demand position in the Wimbleball WRZ 140 WAFU Forecast demand + target headroom

120 Forecast dry year demand

100

80 Megalitres per Megalitresper day

60

40

2039/40 2012/13 2014/15 2019/20 2024/25 2029/30 2034/35

6.2 Options appraisal

Our Plan shows that we are not expecting to see a deficit in our supply demand position for the duration of the planning period, and therefore do not need to implement any options for water resource purposes. Other drivers, such as capital maintenance and quality, are not driving any investment that will detrimentally affect WAFU.

While the water resources supply demand position is not driving an investment need, we will continue our existing activity to ensure an efficient and sustainable position in the future. This activity will include:

. continuing our optional metering programme . continuing to meet our target for water efficiency . continuing to meet our leakage target . building on our current catchment management activity . where other drivers necessitate investment, options that also improve the robustness of our supply systems will be considered where it is economical to do so . making surplus water available to other companies should they require it.

Page 6.3

SWW Water Resources Management Plan 2015 – 2040

6.3 Final supply demand balance

As we forecast a surplus in all three of our WRZs for the duration of the Plan, no options are required and the final supply demand position is the same as the baseline shown in Section 6.1.

Whilst we are confident in the conclusions of our analysis, we will monitor our position over the coming years and should either forecast demand or WAFU differ significantly from our forecasts, we will take action should it be required. 6.3.1 Actual level of service

Figures 6.1 to 6.3 show a supply demand surplus in all of our WRZs. Therefore we currently both meet and exceed our planned level of service in all of our WRZs. Indeed, our Drought Plan advises that the use of Drought Orders or Drought Permits is not currently envisaged to be required within the lifetime of our Drought Plan. 6.3.2 Impact of level of service on deployable output

The relationship between deployable output and level of service is covered in Section 2. 6.3.3 Greenhouse Gas Emissions

Although we have no options within our Plan, there will be a change from the current way we operate our system as a result of introduction of the compensation flow at Kennick, Trenchford and Tottiford as described in section 2.4.3. The impact of this measure will slightly increase the greenhouse emissions from that given in section 2.9 to 48,227 tCO2e – an increase of approximately 0.28%.

6.4 Scenario testing

In developing a water resources strategy up to the year 2040 it is inevitable that numerous assumptions have to be made. It is important to examine the sensitivity of the strategy to these assumptions to ensure that it is both resilient and flexible.

With the predicted surplus throughout the planning period and therefore no required options, this represents a low-risk plan. The Water resources planning guideline1 states that for companies with a simple plan, showing the extent to which varying headroom risk percentile affects the supply demand balance, may be sufficient. Our headroom analysis encompasses all of the major risks inherent within the Plan and therefore we consider that this level of analysis will indeed suffice.

As explained in Section 2.4.3.3 of this Plan, our local Environment Agency team have confirmed that no scenario analysis on sustainability changes is required in this Plan.

1 Environment Agency, “Water resources planning guideline – interim update”, October 2012

Page 6.4

SWW Water Resources Management Plan 2015 – 2040

6.4.1 Sensitivity of final Plan to headroom risk percentile

As detailed in section 5.2.5 of this Plan we used the 85% risk percentile for headroom uncertainty at the beginning of the planning period, reducing to 75% by 2025 and down to 70% by 2040. A useful way of assessing the sensitivity of the Plan is to consider how it would be affected by using a higher risk percentile for headroom. The charts below show varying levels of headroom uncertainty risk percentile and how they affect the supply demand situation for each of our WRZs. The 95th percentile represents the likely worst-case scenario, while the 75th percentile is shown for comparison. Figures 6.4 to 6.6 show that even in this worst- case scenario, supply exceeds demand in all three WRZs, and therefore that the Plan is not sensitive to the choice of headroom risk percentile.

Figure 6.4: Effect of varying headroom risk percentile on the supply demand position in the Colliford WRZ 200 WAFU Forecast demand + target headroom 180 Forecast demand + headroom (95th percentile)

Forecast demand + headroom (75th percentile)

160

140 Megalitres per Megalitresper day

120

100

2039/40 2012/13 2014/15 2019/20 2024/25 2029/30 2034/35

Page 6.5

SWW Water Resources Management Plan 2015 – 2040

Figure 6.5: Effect of varying headroom risk percentile on the supply demand position in the Roadford WRZ 320 WAFU Forecast demand + target headroom Forecast demand + headroom (95th percentile) 280

Forecast demand + headroom (75th percentile)

240

Megalitres per Megalitresper day 200

160

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

Figure 6.6: Effect of varying headroom risk percentile on the supply demand position in the Wimbleball WRZ 140 WAFU Forecast demand + target headroom 120 Forecast demand + headroom (95th percentile)

Forecast demand + headroom (75th percentile)

100

80 Megalitres per Megalitresper day

60

40

2039/40 2012/13 2014/15 2019/20 2024/25 2029/30 2034/35

Page 6.6

SWW Water Resources Management Plan 2015 – 2040

6.4.2 Sensitivity of final Plan to changes in demand

While an analysis of headroom risk satisfies the requirements for a simple plan, we also performed testing against uncertainty in demand forecasts. Demand is one of the most uncertain aspects of forecasting, and while it is already reflected in our headroom uncertainty analysis, this analysis provides a quantitative understanding of the additional changes required to affect the Plan.

Of our three WRZs, Colliford has the greatest vulnerability to changes in demand. On top of the uncertainty already contained in the headroom, it would require 2039/40 household consumption to be 31% higher than forecast, or non-household 113% higher, for Colliford WRZ demand to exceed supply. This is equivalent to an overall increase in forecast consumption of 24% in this WRZ. The increase in forecast total consumption required for Roadford and Wimbleball WRZ demands to exceed supply are even higher at 33% and 28% respectively, allowing the possibility of supporting Colliford from our other WRZs even in this extreme scenario.

This analysis shows that our Plan is not sensitive to changes in demand.

Page 6.7

SWW Water Resources Management Plan 2015 – 2040

7 Choices

7.1 Introduction

Having developed baseline forecasts for water supply and demand (plus target headroom), water companies are required to check whether demand (plus target headroom) is greater than supply at any point during the 25 year planning period. If it is, then companies must investigate options to secure supplies. But if there isn‟t a deficit, as is the case in our region, then there are two possible courses of action described in the Water resources planning guideline1:

1 Do nothing: The water company does not need to take any further action apart from continuing its current policies.

2 Do the right thing: The company wishes to implement a series of measures to become more efficient, better for the environment, maintain its positive supply demand balance beyond the 25 years and achieve Government aspirations.

In either case, it is good practice for a water company to investigate whether there are any ways to lower the overall costs (financial, environmental, social and carbon) of its existing operations. If the second course of action is taken, a water company should decide on the best option for its customers (on the basis of cost and what customers would like) and for the environment (both local and global).

We are very clear in our aim to “do the right thing” and therefore this section of the Water Resources Management Plan (WRMP) will describe some of the actions that we have taken and options that we are examining in order to achieve this.

7.2 Government policy priorities and expectations

The Water White Paper2 made clear the importance that the Government attaches to secure sustainable and affordable supplies of water to customers. In order to achieve these aims there are a number of key policy priorities that the Government expects water companies to address in their plans. These include:

1 The long-term perspective

Taking a long-term perspective, beyond the 25-year planning horizon, to make companies‟ systems more resilient to future uncertainties, such as the impacts of climate change, and to allow efficient, sustainable water resources planning to meet the needs of customers and the environment.

2 Water scarcity and environmental damage

Taking better account of the value of water by reflecting its scarcity and the environmental and social costs of abstraction in order to make the water sector‟s activities more sustainable.

1 Environment Agency, “Water resources planning guideline – interim update”, October 2012, page p99 2 Defra, “Water White Paper - Water for life”, December 2011

Page 7.1

SWW Water Resources Plan 2015 – 2040

3 Further interconnection and water trading

Considering all options to balance supply with demand, including water trading, cross boundary solutions and third party supplier solutions, and providing up-to- date information about the availability of water to third parties (including any future entrants to the market), in order to reduce costs, ensure efficient allocation of available resource and improve innovation within the sector.

4 Reducing the demand for water

Reducing the demand for water by managing leakage and providing services to help customers use water efficiently where there is a reasonable prospect that the benefits of doing so will outweigh the costs.

Our responses to these policy priorities are described in the following sections.

7.3 The long-term perspective

7.3.1 An extended planning horizon

The planning horizon for water resources management plans is 25 years and our final supply demand balance projections for this period are shown in Section 6 of this Plan. However, we have also looked forward a further ten years, ie to the year 2050 and although further projection has not been carried out in the detailed manner of the first 25 years, it should nevertheless give an indication of future prospects assuming the continuation of climate change and the absence of any step change in other circumstances.

These extended projections for the three Water Resource Zones (WRZs) are shown in Figures 7.1 to 7.3 below. It can be seen that no further interventions are necessary in order to ensure a supply demand surplus up until 2050.

Page 7.2

SWW Water Resources Plan 2015 – 2040

Figure 7.1: Colliford WRZ supply demand balance with extended planning horizon

200 WAFU

180 Forecast Demand + Target Headroom

Forecast 'Dry Year' Demand

160

140 Megalitres per day

120

100

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40 2044/45 2049/50

Figure 7.2: Roadford WRZ supply demand balance with extended planning horizon

320

WAFU Forecast Demand + Target Headroom 280

Forecast 'Dry Year' Demand

240 Megalitres per day 200

160

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40 2044/45 2049/50

Page 7.3

SWW Water Resources Plan 2015 – 2040

Figure 7.3: Wimbleball WRZ supply demand balance with extended planning horizon

140

WAFU

120 Forecast Demand + Target Headroom

Forecast 'Dry Year' Demand

100

80 Megalitres per day

60

40

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40 2044/45 2049/50

7.3.2 Long-term resilience and flexibility

As advised by the Water Resources Planning Guideline3, this Plan is not required to address short-term supply losses which are caused by unforeseen events which do not impact on Deployable Output (DO). Consequently, short-term emergencies of this nature (eg temporary transfers of water to meet short-term supply deficits, risks of asset failure, low pressure and supply interruptions) are addressed by a separate resilience planning process. However, for completeness, we have included an outline summary of our work in this area for information.

Resilience planning has been an important “business as usual” activity for South West Water for many years. An example of a typical resilience investment was the construction of a flood protection scheme for Pynes Water Treatment Works (WTW) to a 1 in 200 year return period standard in 2011.

Nationally, the Mythe flooding event of 2007, the severe winter of 2010 and the eruption of Eyjafjallajokull in Iceland have all highlighted the vulnerability of the UK‟s national infrastructure and essential services to disruption caused by external natural hazards. These events led the Cabinet Office to publish guidelines to public bodies and national service providers4. Ofwat also subsequently published guidance5.

The water industry response was to commission UKWIR to develop a methodology that draws together all available data and resilience planning and good practice into

3 Ibid. 1 4 Cabinet Office, “Keeping the Country Running”, October 2011 5 Ofwat, “Resilience – outcomes focused regulation”, May 2012

Page 7.4

SWW Water Resources Plan 2015 – 2040

one document which was published in 2012. South West Water took an active role in the research and sat on the project steering group.

Using this methodology we have undertaken a considerable body of work on resilience which has been included in our Business Plan. This work included:

. Using updated Environment Agency data to re-screen all non-infrastructure assets, including WTWs, pumping stations and service reservoirs, with respect to flooding vulnerability.

. Assessing external hazards and impacts on our top 15 water quality zones where we have large single source supply areas.

. Screening water non-infrastructure assets for vulnerability to coastal erosion.

. Designing interventions to mitigate risks identified via improvements in resilience using the Cabinet Office guidewords.

- Resistance - Reliability - Redundancy - Response and recovery

. Assessing customers‟ willingness to pay for risk mitigation.

. Using cost benefit analysis on interventions to guide business planning.

As our future Security Of Supply Index does not depend upon major capital investments, which could prove either uncertain or unnecessary, we believe that we are able to respond flexibly (between price limits) to an uncertain future in order to achieve the best results for both customers and the environment.

7.4 Water scarcity and environmental damage

7.4.1 Environmental liaison

South West Water is keenly aware of its environmental responsibilities and has for many years held regular meetings with organisations interested in the aquatic environment. We have worked with the Environment Agency and such groups as the Fowey Fisheries & Resources Group and the Exe Mitigation Group to minimise the impact of our operations on water bodies and to ensure that our abstractions are sustainable. Many projects such as gravel restoration, habitat recreation, a fish hatchery and fish passes have been directly funded by South West Water. We are committed to maintaining our close links with these organisations in the future.

During the past five years we have developed two major new water resources in Cornwall: Park and Stannon Lakes. These lakes were created from redundant China Clay pits and added significantly to our DO and the resilience of supplies without the need to build large dams. Extensive landholdings were associated with the pits and South West Water willingly embraced an obligation to maximise the

Page 7.5

SWW Water Resources Plan 2015 – 2040

environmental potential of this land. We have established an Environmental Management Group comprising environmental regulators and other interested parties and have provided funds for restoration. The success of this work can be gauged from the fact that the once derelict Park Lake area has recently been designated as a County Wildlife Site. We believe that this innovative resource development demonstrates environmental sensitivity and our commitment to sustainable abstraction.

7.4.2 Catchment management

Climate change and pressures for intensive agriculture are affecting the quality and quantity of the raw water we collect in our reservoirs and abstract from rivers and underground. South West Water promoted two series of actions from 2006 to 2009 to develop responses to this based on moorland restoration on Exmoor and reducing soil loss and pollution from farms. From this, we are implementing a five-year £8.8m programme to address water quality and quantity problems at source. This new and holistic approach to water supply is termed catchment management.

Exmoor and Dartmoor have been changed significantly in the last hundred years as a result of ditch construction and various drainage schemes. The purpose of these historic and often grant-aided projects was to improve the land for agricultural purposes, but loss of natural water storage has led to significant erosion, carbon dioxide release from drying peat, biodiversity loss and increased downstream flood risks.

A restoration project on Exmoor was started in 2003 by Exmoor National Park Authority with South West Water providing additional funding from 2006 onwards. Any changes in grazing yields are offset by support payments from Natural England to the landowners involved. Restoration of a more natural drainage pattern is anticipated to have a number of beneficial effects on downstream hydrology such as increased attenuation of flood peaks and an increase in baseflow. Such impacts will ultimately increase the water resources available for public supply. The University of Exeter is undertaking an extensive hydrological investigation to measure changes.

A group of 12 farms were improved above one of our most damaged resources, Upper Tamar Lake, near Bude. Similar techniques were undertaken to those of the England Catchment Sensitive Farming Delivery Initiative to reduce soil erosion, loss of manures and damage caused by the access of stock to watercourses. Within one month, water below the farms was of better quality than that above. This showed that natural water quality improvement was occurring as farms were not losing their soil and fertilizer to streams and rivers in or beside their land. Farm productivity and incomes have been raised; both are essential for food production and ensuring a viable future for farming.

A more extensive programme of changes to uplands and farmed land was promoted in the last Periodic Review (PR09) with support from the Environment Agency, Ofwat and the Consumer Council for Water. We are now implementing a 2,000 hectare SSSI restoration on Exmoor, an extensive trial on Dartmoor and seven catchment scale farmland improvements above key intakes and reservoirs. We are also investing £0.7m in 17 two-year catchment investigations to develop further schemes for inclusion in our Periodic Review (PR14) Business Plan. This programme of water quality and quantity protection is branded as Upstream Thinking.

Page 7.6

SWW Water Resources Plan 2015 – 2040

We have very limited land holdings above our reservoirs, so this work is carried out on third party property by negotiation with owners and tenants. Clean raw surface water costs approximately 20% less to treat than water with heavy sediment loads and if deteriorating water quality can be stopped it will delay or avoid costly longer- term water treatment upgrading. When assessed over 30 years in line with Ofwat‟s instructions for PR09, with inclusion of the carbon storage benefits, the programme offers a benefit to cost ratio of 65:1 or better. Our downstream customers are offered the prospect of lower bill increases and „soft engineering‟ for flood protection, which complements essential conventional defences for urban areas and our WTWs adjacent to rivers. Slower release of water in droughts helps protect biodiversity and will improve residual river flows for the dilutions of sewage effluents downstream.

The current improvement programme for moorland and catchments raises £8.1m for local projects to restore the natural water storage ability of uplands and limit land- based damage to rivers from 2010 to 2015. These projects are included in the South West River Basin Management Plan as they will contribute to „Good Status‟ delivery for the Water Framework Directive. Moorland restoration offers carbon capture at about 1.2 tonnes CO2/hectare/year while farmers are being encouraged to create woodlands with the Woodland Trust‟s support and wetlands. The cost to our water customers is an increase of 65p on bills by 2015, compared to a customers‟ willingness-to-pay identified in PR09 of £1.80 for additional environmental projects. Our delivery partners are Westcountry Rivers Trust, Devon Wildlife Trust, Cornwall Wildlife Trust and others with a network of local experts and volunteers keen to identify risks and work with landowners to promote remedial action. This private company/third sector link is an alternative approach to centrally directed regulation and is, we consider, more efficient and cost effective.

In preparation for the reform of the Common Agriculture Policy and the next water industry Periodic Review (PR14) we are working with Defra, the Environment Agency, Natural England and our partners to design a practical method of making ecosystem service payments to reward best land management practices above our principal water sources. We believe that farmers should be able to develop new income streams from water and carbon management.

Our wider catchment management approach matches many of the aims of the Natural Environment White Paper. Increasing food security while protecting water quality, improving biodiversity and dealing with climate change threats are achievable, through the actions we have designed and are delivering with our partners through Upstream Thinking.

Our main restoration priorities are:

. Exmoor Mires restoration for the River Exe and supplies to Allers and Pynes WTWs. . Exmoor farmed land to protect inflows to Wimbleball Reservoir. . Otter Valley farmed land to protect Otter Valley groundwater resources. . Tamar catchment farmed land to protect Upper Tamar Lake and Gunnislake abstraction. . West Penwith farmed land improvements to protect raw water used at Drift Reservoir.

Page 7.7

SWW Water Resources Plan 2015 – 2040

Innovative projects like catchment management demonstrate our appreciation of the scarcity and value of water and our willingness to develop new approaches to achieve the balance between secure sustainable supplies and value for our customers.

7.5 Further interconnection and water trading

7.5.1 Conjunctive use and interconnection

As explained earlier in this Plan, our water resources system is based upon three WRZs each comprising a number of different sources such as reservoirs, river intakes and groundwater abstractions. The sources within the WRZ are operated conjunctively which means that a WTW may be supplied from a number of different sources and a demand centre may be supplied from a number of different WTWs. The conjunctive use of resources not only adds to the resilience of public water supply but also enables the total DO of the resource zone to be maximised.

To operate sources conjunctively requires interconnecting pipework, so in effect our WRZs form a grid. Over recent years we have made significant investments to increase interconnections between resource zones. For example, we are now able to use Allers WTW, which is in the Wimbleball WRZ, to supply South Molton which is in the Roadford WRZ. We can also use Pynes WTW, which is in the Wimbleball WRZ, to supply both the Crediton area and parts of South Devon both of which are in the Roadford WRZ.

These transfers enable us to even out the supply demand balance between WRZs. So for example, instead of developing a new source of raw water to remove a deficit in one WRZ, we transfer water from an adjacent zone which has a supply demand surplus. This increasing interconnectivity between WRZs allows us to make the best possible use of our water resources and avoid promoting unnecessary developments. It also means that we are gradually developing a regional water grid. This increasing interconnectivity is good for our customers and good for the environment as it means that we are making the best use of the resources we have rather than developing costly new sources with all the attendant environmental consequences.

In the future we intend to further develop the links between zones but these will be justified on the need for further resilience rather than the supply demand balance. The further development of our regional grid will, of course, facilitate new entrants to public supply by enabling them to input to the grid at a location which may be remote from their intended customers.

In order to make the best use of our conjunctive use system we use a computer model called MISER6 (Water Resources Module) to aid operational decision making. MISER is described as “a highly configurable decision-support software tool that facilitates optimal water management, asset and resource planning”. The model enables us to provide secure supplies to customers by using sources within environmental constraints in such a way as to minimise costs (eg water treatment costs, pumping costs). This mode of operation is therefore good for both customers and the environment.

6 MISER is produced by Tynemarch

Page 7.8

SWW Water Resources Plan 2015 – 2040

7.5.2 Water trading

In autumn 2012, in accordance with the Water Resources Planning Guideline7 we published our Contact Plan8. Our Contact Plan9 advised that on the basis of the work that we had completed at that time and the information that had been provided to us by the Environment Agency and other organisations, we believed it was unlikely that there would any supply demand deficit in any of our water resources zones before the end of the planning horizon of 2039/40.

Since production of our Contact Plan10 we carried out further work to quantify any surplus in our supply demand balance in each of our WRZs which we present in this Plan.

As can be seen in Section 6, we have a supply demand surplus in all three WRZs throughout the planning period to 2040. We have made this position known to other water companies and have discussed it in detail with neighbouring companies.

The amount of water that is available for transfer is based upon an assessment of our supply demand balance taking into account known and anticipated sustainability reductions. However, it is possible that numbers could change in the future as a result of changes in policy by regulators or other, as yet unknown, factors and we have therefore initially taken a conservative view of the water available for transfer. Due to various uncertainties and its greater geographical distance from likely users we have not currently produced figures showing the availability of surplus supplies in the Colliford WRZ.

This information is provided in good faith and should be sufficient for another water user to estimate whether a transfer of water from South West Water looks economically attractive in comparison with the alternatives. Should this be the case then further information and studies will clearly be required.

7.5.2.1 Location where the water is available

Water would be available at two locations:

1 Northbridge SX 93 97

This is the location of one of our main intakes on the River Exe. When the river is above the prescribed flow measured at Thorverton Gauging Station we abstract from the natural flow of the river, when it falls below this level we make releases from Wimbleball Reservoir for subsequent abstraction at Northbridge Intake.

7 Ibid. 1 8 South West Water, “Contact Plan”, September 2012 9 Ibid. 8 10 Ibid. 8

Page 7.9

SWW Water Resources Plan 2015 – 2040

2 Gunnislake SX 43 72

This is the location of our main intake on the River Tamar. When the river is above the prescribed flow measured at Gunnislake Gauging Station, we abstract from the natural flow of the river. When it falls below this level, we make releases from Roadford Reservoir for subsequent abstraction at Gunnislake Intake.

It should be noted that the water is available in the river at both locations. Although we have intakes at these locations, it cannot be assumed that our existing pumping capacity and pipework are sufficient to abstract the extra water and therefore this must be taken into account when assessing the viability of a transfer.

7.5.2.2 The average daily and peak daily volume available

The currently available water is based on the use of existing abstraction licences.

. A constant 5 Ml/d is available at Northbridge . A constant 15 Ml/d is available at Gunnislake

We plan to reduce leakage to 64 Ml/day between 2020 and 2040, which will make the following additional water available:

. A constant 3 Ml/d at Northbridge . A constant 9 Ml/d at Gunnislake

7.5.2.3 Treated or untreated?

The available water is untreated river water.

7.5.2.4 The earliest date the water will become available

. 5 Ml/d is available at Northbridge from 2015 . 15 Ml/d is available at Gunnislake from 2015

The date at which the additional water from leakage reduction is available will depend upon the phasing of the leakage reduction project.

7.5.2.5 Period over which the water is expected to be available

. 5 Ml/d is available at Northbridge until at least 2040 . 15 Ml/d is available at Gunnislake until at least 2040

The additional water that will be available as a result of further leakage reduction will be available until at least 2040.

7.5.2.6 The indicative contract price per Ml

The figures below are in 2012/13 prices, with an assumption of a link to RPI.

. For Northbridge:

Page 7.10

SWW Water Resources Plan 2015 – 2040

Fixed annual charge for constant 5Ml/d £410,625 Variable charge for volume up to 5Ml/d 11.25p/m3 Variable charge for volume over 5Ml/d 24p/m3

Price p/m3 for constant 5Ml/d – 22.5p/m3

. For Gunnislake:

Fixed annual charge for constant 15Ml/d £1,231,875 Variable charge for volume up to 5Ml/d 11.25p/m3 Variable charge for volume over 5Ml/d 24p/m3

The additional spare water made available from leakage currently costs £1.64/m3. On the basis that the additional leakage control is usually being carried out for the purposes of water supply, the cost could be structured as follows:

. Northbridge:

Fixed annual charge for provision of additional 3Ml/d £1,757,475 Variable charge for volume up to 3Ml/d 3.5p/m3

. Gunnislake:

Fixed annual charge for provision of additional 9Ml/d £5,272,425 Variable charge for volume up to 9Ml/d 3.5p/m3

If the spare water delivered through leakage had other drivers (eg the Government wanting leakage reduced to have spare water for trading) and this did not have to be reflected in the contract price, then the charge could amount to 29p/m3 (with possible 25.5p/m3 fixed for the contracted volume and 3.5p/m3 variable).

There are a number of assumptions built into this price:

. No specific additional winter pumping into Wimbleball (or into Roadford for the Gunnislake intake) is required. Even if this is required, it does not affect the Long Run Marginal Cost (LRMC) as long as demand is taken evenly throughout the year.

. The price is based on a quantity actually available. Based on the pump storage cost assumption for Wimbleball, there could be a 50% availability charge (whatever the volume was taken), with the remainder of the total price based on the actual volume up to the permitted level. If the water was available, daily amounts above the constant values quoted could be charged at a higher rate.

Figures 7.4 and 7.5 show the supply demand balance for the Roadford and Wimbleball WRZs should a water user require the maximum available transfers from South West Water (without the additional water from leakage reductions identified in Section 7.5.2.2). As can be seen from these graphs, the WRZs are still in surplus despite these transfers. As described in Section 7.5.2, we have initially taken a conservative view of the amount of water available for transfer. However, should there be further interest in transfers from these WRZs, then increased quantities may be available.

Page 7.11

SWW Water Resources Plan 2015 – 2040

Figure 7.4: Roadford WRZ supply demand balance with transfer of 15 Ml/d to another water user 320

WAFU 4.8 Ml/d Sustainability Reduction identified in our WRP09 SWW dry year demand + target 280 headroom 2.16 Ml/d Sustainability SWW dry year demand + target Reduction headroom + 15 Ml/d transfer

240

Megalitres per day 200

160

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

Page 7.12

SWW Water Resources Plan 2015 – 2040

Figure 7.5: Wimbleball WRZ supply demand balance with transfer of 5 Ml/d to another water user 140 WAFU

SWW dry year demand + target 120 headroom SWW dry year demand + target headroom + 5 Ml/d transfer

100

80 Megalitres per day

60

40

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

7.6 Reducing the demand for water

7.6.1 Water metering

The South West Water Region is not considered by the Environment Agency to be an area of “serious stress”11. There is therefore no imperative for us to promote a case for universal metering. Currently approximately 75% of our domestic customers have a water meter and we anticipate that in 2020 this will reach 85%. Ultimately we will require SMART meters to allow for new tariffs and to provide better information on water use to customers. We are currently trialling SMART meters and would anticipate their use after 2015.

7.6.2 Per capita consumption

The per capita consumption of our domestic customers is well below the national average of 147 litres per person per day (l/p/d). At just 138 l/p/d, it is already close to the Government‟s aspirational level of 130 l/p/d and is expected to reach this figure in around 3 years. Furthermore, as described in Section 3 of this Plan, we anticipate per capita consumption and demand levels to continue their downward trend throughout the period covered by this Plan without any specific additional interventions.

11 Environment Agency, “Improving the classification of water stressed areas”, November 2012

Page 7.13

SWW Water Resources Plan 2015 – 2040

7.6.3 Leakage

We have run a number of focus groups across the region to understand what water and sewerage services mean to our customers and what their preferences are for our activities in the future. From the information gathered in these focus groups, we have carried out further survey work to understand exactly what our customers‟ priorities are for compulsory activities, such as those required by legislation, and also for services where there are options over what to deliver.

The survey results to date have revealed that our customers‟ top priorities are very similar to those identified in 2007 (when our 25-year Strategic Direction Statement was first published). Unsurprisingly, the provision of safe and clean water remains the number one priority. The consistent number two priority for customers was leakage control.

We are very proud of our industry-leading leakage management record and have met or exceeded the agreed target for leakage every year that it has been set. It should also be borne in mind that this target of 84 Ml/d is significantly below the Sustainable Economic Level of Leakage (SELL) as described in Section 3.5 of this Plan. Nevertheless, we are keen to do more in response to our customers‟ clearly stated priority for leakage control.

We have therefore been consulting customers on an ambitious proposal to reduce leakage by a further 24% from a regional total of 84 Ml/d to just 64 Ml/d. The effect of this reduction on the supply demand in the three WRZs is set out below in Figures 7.6, 7.7 and 7.8.

However, a reduction in leakage to 64 Ml/d is not considered cost effective or affordable before 2020 so we are planning further reductions for future investment periods (consistent with our customers‟ views).

In the meantime, we will continue to develop effective and efficient ways to control leakage, including:

. targeting more areas for pressure management to reduce the number of leaks and bursts and our overall average pressure by 9%

. continuing to develop the efficiency and effectiveness of our teams who carry out leak detection and repair activities

. aiming to repair all significant leaks within two working days

. using computer modelling and network automation techniques to identify emerging problems and leakage issues

. proactively replacing deteriorating pipes in our network.

Our aim to repair all significant leaks within two working days represents a 30% improvement on our current target. We plan to deliver this through technology improvements to our leak detection activities, work planning and repair processes.

Page 7.14

SWW Water Resources Plan 2015 – 2040

Figure 7.6: Colliford WRZ supply demand balance showing the impact of reducing regional leakage to 64 Ml/d 200

WAFU

Dry year demand + target headroom without leakage reduction 180 Dry year demand + target headroom

160

140 Megalitres per day

120

100

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

Figure 7.7: Roadford WRZ supply demand balance showing the impact of reducing regional leakage to 64 Ml/d 320 WAFU

Dry year demand + target headroom without leakage reduction

280 Dry year demand + target headroom

240

Megalitres per day 200

160

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

Page 7.15

SWW Water Resources Plan 2015 – 2040

Figure 7.8: Wimbleball WRZ supply demand balance showing the impact of reducing regional leakage to 64 Ml/d 140 WAFU

Dry year demand + target headroom without leakage reduction 120 Dry year demand + target headroom

100

80 Megalitres per day

60

40

2012/13 2014/15 2019/20 2024/25 2029/30 2034/35 2039/40

7.7 The future

As can be seen from previous sections in the Plan, we currently enjoy secure and sustainable water resources in our region. Furthermore, we believe that this position will be maintained throughout the period covered by this Plan without significant further intervention. This situation is a direct result of our investment in water resources over many years and by company policies that are already in line with Government objectives and aspirations. Nevertheless, we are committed to work with Government, regulators and other interested parties to continue to “do the right thing” for our customers and the environment.

Page 7.16

SWW Water Resources Plan 2015 – 2040

8 Glossary of terms used in the WRMP1

Above ground supply Losses from the pressurised system after the point of pipe losses consumption.

Abstraction The removal of water from any source, either permanently or temporarily.

Abstraction licence The authorisation granted by the Environment Agency under the terms of an abstraction licence.

AISC See Average Incremental Social Costs

AMP5 Asset Management Plan 5 – the period 2010/11-2014/15 (also referred to as K5)

AMP6 Asset Management Plan 6 – the period 2015/16-2019/20 (also referred to as K6)

Annual average daily The cumulative demand in a year, divided by the number of demand days in the year.

Aquifer A geological formation, group of formations or part of a formation that can store and transmit water in significant quantities.

Atrazine A herbicide which is widely used globally, but no longer used in the UK.

Average Incremental The net present value (NPV) of the option costs, including Social Costs (AISC) social and environmental costs, divided by the net present value of the option capacity or output.

Capex Capital expenditure.

Catchment area The area of land whose rainfall feeds a particular river, lake or reservoir.

Communication pipe That part of the service pipe between the distribution main and the highway boundary.

Consumption Water delivered billed less underground supply pipe losses. Consumption can be split into customer use plus total plumbing losses.

Customer use Consumption less total plumbing losses.

Customer-side The implementation of policies or measures which serve to management control or influence the consumption or waste of water by the end user.

1 The majority of these definitions have been taken from the UKWIR/Environment Agency publication “Definitions of Key Terms for Water Resources Practitioners”, Ref No 97/WR/14/1, 1997

Page 8.1

SWW Water Resources Management Plan 2015 – 2040

Demand The implementation of policies or measures which serve to management control or influence the consumption or waste of water.

Demand A single measure, or a combination of measures (eg a public management option awareness campaign using both leafleting and radio advertising), taken to influence the demand for water.

Deployable Output The output of a commission source or group of sources or of bulk supply as constrained by:

. environment . licence, if applicable . pumping plant and/or well/aquifer properties . raw water mains and/or aqueducts . transfer and/or output main . treatment . water quality

for specified conditions and demands

Distribution input The amount of water entering the distribution system at the point of production. This is the quantity usually measured as demand by customers.

Distribution losses Made up of losses on trunk mains, service reservoirs, distribution mains and communication pipes. Distribution losses are distribution input less water taken.

Distribution Management of the transmission, storage, distribution and management mains supply pipe of potable water.

Distribution System Water knowingly used by a company to meet its statutory Operational Use obligations, particularly those relating to water quality. (DSOU) Examples include mains flushing and air scouring.

District Metering An area that is permanently defined by closed valves or Area (DMA) other physical constraints in which distribution losses are measured and managed.

DMA See District Metering Area

Drawdown period The length of time during which the contents of a reservoir are always less than a target refill storage volume.

GAC See Granular Activated Carbon

Granular Activated An adsorbent filtration media used to remove trace organic Carbon (GAC) compounds from water

Greywater Water that can be considered for non-potable re-use.

Groundwater Water within the saturated zone of an aquifer.

Page 8.2

SWW Water Resources Management Plan 2015 – 2040

Households Properties (normally occupied) receiving water for domestic purposes which are not factories, offices or commercial premises.

Hydrological yield The unconstrained output of a source that can be sustained by the catchment or aquifer feeding the source.

Internal metering Meters fitted within the household boundary which measure consumption but do not record underground supply pipe losses.

Internal plumbing Losses from the non-pressurised system after the point of losses consumption.

K5 The period 2010/11-2014/15. Also referred to as AMP5.

K6 The period 2015/16-2019/20. Also referred to as AMP6.

Leakage The sum of distribution losses and underground supply pipe losses.

Level of service The design standard used by a company for the security of supply to customers. It is expressed in terms of the average frequency with which:

. a customer might experience demand restrictions such as hosepipe bans . the Company might apply for drought orders or permits.

Licence variation The authorisation granted by the Environment Agency to change the terms of an abstraction licence.

Local reservoir Small reservoir supplying a local area. Usually supported by a strategic reservoir.

Maximum Likelihood A statistical technique where a reconciliation item is Estimation (MLE) distributed to the largest and least certain components of an estimate of the magnitude of a variable. The technique can be applied to the reconciliation of a water balance, for example.

Megalitre (Ml) Measure of volume; one million litres

Meter optants Properties in which a meter is installed at the request of its occupants.

Micro-component The process of deriving estimates of present or future analysis consumption based on expected changes in the individual components of customer use.

MLE See Maximum Likelihood Estimation

Net present value The NPV of an investment is the discounted value of (NPV) expected income less cost.

Page 8.3

SWW Water Resources Management Plan 2015 – 2040

Non-households Properties receiving water for domestic purposes but which are not occupied as domestic premises eg factories, offices and commercial premises, cattle troughs. They also include properties containing multiple households which receive a single bill (eg a block of flats).

NPV See Net Present Value

Opex Operating expenditure.

Outage A temporary loss of Deployable Output. (Note that outage is temporary in the sense that it is retrievable, and therefore Deployable Output can be recovered. The period of time for recovery is subject to audit and agreement. If an outage lasts longer than 3 months, analysis of the cause of the problem would be required to satisfy the regulating authority of the legitimacy of the outage.)

Outage allowance The value of allowable outage expressed in Ml/d.

PCV See Prescribed Concentrations or Values

Peak demand The highest demand that occurs, measured either hourly, daily, weekly, monthly, yearly, over a specified period of observation.

Planned outage A foreseen or pre-planned outage resulting from a requirement to maintain sourceworks asset serviceability.

PMA See Pressure managed area

Point of consumption The point where the supply pipe rises above ground level within the property – usually the inside stopcock or an internal meter.

Point of delivery The point at which water is transferred from mains or pipes which are vested in the water supplier into pipes which are the responsibility of the customer. In practice this is usually the outside stopcock, boundary box or external meter.

Point of production The point where treated water enters the distribution system.

Prescribed The numerical value assigned in the "Water Supply (Water Concentrations or Quality) Regulations 2000 (England)" defining the maximal or Values (PCV) minimal legal concentration or value of a parameter

Pressure Managed An area, defined by closed valves or other physical means, Area (PMA) within which hydraulic pressure is monitored, controlled and managed.

PR14 Periodic Review 2014

Page 8.4

SWW Water Resources Management Plan 2015 – 2040

Production Management of the storage, transmission and treatment of management raw water.

Pumped storage A means of increasing the natural refill of a reservoir by pumping water to the reservoir from another catchment.

Q95 The river flow which is equalled or exceeded for 95% of the time. Also referred to as the “95 percentile”.

Raw water exported Raw water exported from a specified geographical area.

Raw water imported Raw water imported from a specified geographical area.

Raw water losses The net loss of water to the resource system(s) being considered, comprised of mains/aqueduct (pressure system) losses, open channel/very low pressure system losses, and losses from break-pressure tanks and small reservoirs.

Raw water Regular washing-out of mains due to sediment build-up and operational use poor quality of source water.

Reconciliation item The difference between the estimates of the magnitude of a variable and the sum of the estimates of the individual components of that variable.

Saturated zone The zone in which the voids in a rock or soil are filled with water at a pressure greater than atmospheric.

SEA See Strategic Environmental Assessment

Selective metering Metered charging of a defined subset of households, such as a town, or a region or particular types of customers eg sprinkler users.

SELL See Sustainable Economic Level of Leakage

Service pipe The sum of the communication pipe and the supply pipe.

Source A named input to a resource zone. A multiple well/spring source is a named place where water is abstracted from more than one operational well/spring.

Sourceworks All assets used between and including the point of abstraction and the point at which water is first fit for purpose. These include:

. abstraction works . reservoir and river intakes . boreholes . raw water storage . pumping plant and mains . water treatment plant . treated water storage . treated water pumping plant

Page 8.5

SWW Water Resources Management Plan 2015 – 2040

Strategic A study of the effects of certain plans, policies and Environmental programmes on the environment. Assessment (SEA)

Strategic Reservoir A large or dominant reservoir (cf local reservoir) supplying water directly or indirectly over a wide area. South West Water has three strategic reservoirs: Wimbleball, Colliford and Roadford.

Supply pipe That part of the service pipe not within the boundary of the highway.

Supply pipe losses The sum of the underground supply pipe losses and above ground supply pipe losses.

Sustainable The Sustainable Economic Level of Leakage (ELL) is the Economic Level of point at which the cost of further leakage reduction is just Leakage (SELL) equal to the additional benefit gained. The calculation of SELL includes the social and environmental costs and benefits associated with leakage. It relies on two key relationships:

. The costs of the various activities for controlling leakage e.g. finding and repairing leaks, and how they vary with the level of leakage

. The impact that different leakage levels have on the costs of delivering water to customers (treatment and pumping costs) and the timing of planned new supply, treatment and demand management (including water efficiency) schemes

Target headroom The minimum buffer that a prudent water company should allow between supply and demand to cater for specified certainties (except those due to outages) in the overall supply demand balance.

Total leakage The sum of distribution losses and underground supply pipe losses.

Total plumbing The sum of above ground supply pipe losses and internal losses plumbing losses.

Total treated water The sum of distribution losses, underground supply pipe losses losses and total plumbing losses.

Underground supply Losses between the point of delivery and the point of pipe losses consumption.

Page 8.6

SWW Water Resources Management Plan 2015 – 2040

Unplanned outage An outage caused by an unforeseen or unavoidable legitimate outage event affecting any part of the sourceworks and which occurs with sufficient regularity that the probability of occurrence and severity of effect may be predicted from previous events or perceived risk. Note that the definitive list of legitimate unplanned outage events is:

. pollution of source . turbidity . nitrate . algae . power failure . system failure

Other events should be classified elsewhere, for instance as planning allowances.

Voids Empty properties not currently containing a household or non-household.

WAFU See Water Available For Use

Water Available For The value in Ml/d calculated by the deduction from Use (WAFU) Deployable Output of allowable outages in a resource zone.

Water balance The allocation of total distribution input across its constituent components (eg in the current year or base year of a demand forecast).

Water delivered Water delivered to the point of delivery.

Water delivered billed Water delivered less water taken unbilled. It can be split into unmeasured household, measured household, unmeasured non-household and measured non-household water delivered billed.

Water Resource Zone The largest possible zone in which all resources, including (WRZ) external transfers, can be shared and hence the zone in which all customers experience the same risk of supply failure from a resource shortfall.

Water taken Distribution input minus distribution losses.

Water taken legally Water taken legally but not charged for, such as water taken unbilled from hydrants for fire fighting.

Water taken unbilled Water taken illegally unbilled plus water taken legally unbilled.

WRZ See Water Resource Zone

WTW Water Treatment Works

Page 8.7

SWW Water Resources Management Plan 2015 – 2040

WWTW Waste Water Treatment Works

Page 8.8

SWW Water Resources Management Plan 2015 – 2040

Appendix A Headroom uncertainty calculations

SWW Water Resources Management Plan 2015 – 2040

1 Introduction

This Appendix is divided into the following three sections:

. A1: Headroom Uncertainty – Colliford Water Resource Zone (WRZ) . A2: Headroom Uncertainty – Roadford WRZ . A3: Headroom Uncertainty – Wimbleball WRZ

Each section contains details of the headroom uncertainty calculations we have used in the supply demand balance submission based upon the 2002 UKWIR headroom report1. The following components of headroom uncertainty are included in the methodology described by the Report:

Supply related . Vulnerable surface water licences S1 . Vulnerable groundwater licences S2 . Time limited licences S3 . Bulk imports S4 . Gradual pollution causing a reduction in abstraction S5 . Accuracy of supply-side data S6 . Uncertainty of impact of climate change on source yield S8 . Uncertain output from new resource developments S9

Demand related . Accuracy of sub-component data D1 . Demand forecast variation D2 . Uncertainty of impact of climate change on demand D3 . Uncertain outcome from demand management measures D4

With regard to time limited licences (S3), the report states (p 21) that “this uncertainty will quite legitimately be reflected in Headroom Uncertainty”. However the Environment Agency’s Water resources planning guideline2 states “companies should not make allowances in their headroom calculations for the risk of time- limited licences not being renewed”. Therefore we have excluded this risk from headroom uncertainty.

Our calculation of the supply-side components utilises the Water Available For Use (WAFU) figures we have derived from water resources modelling. As the modelling uses target headroom data there is an iterative process to determine the most accurate WAFU data. As a consequence the WAFU figures we have used in the current headroom assessment do not necessarily correspond exactly to the WAFU data presented in the Water Resources Management Plan (WRMP) tables of our Plan. However, the differences in the data are very small and the impact on headroom is insignificant.

1 UKWIR, “An Improved Methodology for Assessing Headroom”, Ref 02/WR/13/2, 2002 2 Environment Agency, “Water resources planning guideline – interim update”, October 2012, page 95

Page A1

SWW Water Resources Management Plan 2015 – 2040

2 Section contents

Each of the three sections follows the same format and includes the following:

. A summary of discussions on the above components in relation to individual sources within the WRZ. In many cases the choice and parameters of uncertainty distribution have been determined by expert judgement.

. The Headroom Spreadsheet containing the parameters for the uncertainty distributions at a series of time horizons. The spreadsheet utilises @RISK software for calculating probability distributions.

. The @RISK Output Details Report and chart.

. Tornado graphs for selected time horizons showing the relative significance of the individual uncertainty components.

. A chart of headroom uncertainty, showing the simulation results for the following groups of headroom components:

• Supply uncertainty excluding climate change (S1 – S7, S9) • Demand uncertainty excluding climate change (D1, D2, D4) • Impacts of climate change on supply (S8) • Impacts of climate change on demand (D3) • All components (S1 – S9 inclusive, D1 – D4 inclusive)

3 Level of headroom uncertainty

The @RISK Output Details Report provides a summary of statistics for each time horizon including percentiles of headroom uncertainty. It is from these percentiles that the appropriate level of headroom uncertainty must be obtained. We have determined this level by reference to the Environment Agency’s Water resources planning guideline3, which makes it clear that:

. the Environment Agency does not expect companies to plan for 100% certainty . the Environment Agency does not expect companies to apply too low a target headroom . water companies should accept a higher level of risk in future than at present.

We consider that the most appropriate level of headroom uncertainty is the 85th percentile for the beginning of the planning period declining to the 70th percentile by 2039/40. These target values of headroom uncertainty are shaded blue on the @RISK Output Details Report.

3 Ibid. 2, page 94

Page A2

SWW Water Resources Management Plan 2015 – 2040

4 Target headroom and the impact of climate change

We have used a Monte Carlo approach to the assessment of target headroom in accordance with the guidelines4. This produces a joint probability distribution by combining individual probability distributions in a stochastic manner. Therefore the isolation of an element of target headroom associated with an individual risk can be misleading. Nevertheless, we have calculated target headroom with and without the climate change elements. The results for each WRZ are shown in the relevant sections.

4 Ibid. 2

Page A3

SWW Water Resources Management Plan 2015 – 2040

Section A1 Colliford WRZ headroom uncertainty

S1 Vulnerable surface water licences

No vulnerable surface water licences have been identified.

S2 Vulnerable groundwater licences

No vulnerable groundwater licences have been identified.

S3 Time-limited licences

Environment Agency guidelines preclude these from the headroom analysis.

S4 Bulk imports

There are no bulk imports into the Colliford WRZ.

S5 Gradual pollution of sources causing a reduction in abstraction

No sources in the Colliford WRZ have been identified as being at risk.

S6 Accuracy of supply side data

S6/1 Uncertainty for yields constrained by pump capacity

There are no operational groundwater sources in the Colliford WRZ.

S6/2 Meter uncertainty for licence critical sources

We have assumed that all sources are subject to meter uncertainty and:

. meter error is normally distributed around a mean of 0 Ml/d . 95% probability that the meter is recording within ±5%

Table A.1: S6/2 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 WAFU 158.835 158.109 156.898 156.656 156.414 156.172 155.978 5% of WAFU 7.942 7.905 7.845 7.833 7.821 7.809 7.799 SD 4.05 4.03 4.00 4.00 3.99 3.98 3.98 RiskNormal (0,4.05) (0,4.03) (0,4.00) (0,4.00) (0,3.99) (0,3.98) (0,3.98)

Page A4

SWW Water Resources Management Plan 2015 – 2040

S6/3 Uncertainty for aquifer constrained groundwater sources

There are no operational groundwater sources in the Colliford WRZ.

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

The accuracy of river flow measurements, even from well-maintained weirs, cannot be expected to be better than ±10%. However, this does not necessarily mean that DO can be estimated to within ±10%. In order to assess the effect of errors in flow measurement on DO it would be necessary to do many water resources model runs with perturbed river flow sequences.

However, it would be incorrect to adjust all sequences by +10% and then by -10% as all sequences would not be out by the same amount at the same time. It would also be incorrect to randomly adjust figures on a daily basis as, for example, Restormel would not be reading +9% of the correct value one day and -7% the next day.

The effect on DO could be estimated by ascribing a different adjustment factor to each flow sequence within the bounds of ±10% and then running the resources model. When this process has been repeated over and over again with different factors a distribution of DO would be found. However, this process is unworkable in practice.

We have therefore concluded that the uncertainly in DO of surface water sources should be estimated pragmatically by assuming that there is a 95% probability that the value is within ±10% of the estimated value.

. error is normally distributed around a mean of 0 Ml/d . 95% probability that the value is within ±10%

Table A.2: S6/4 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 WAFU (Ml/d) 158.835 158.109 156.898 156.656 156.414 156.172 155.978 10% of WAFU 15.883 15.811 15.690 15.666 15.641 15.617 15.598 SD 8.10 8.07 8.01 7.99 7.98 7.97 7.96 RiskNormal (0,8.10) (0,8.07) (0,8.01) (0,7.99) (0,7.98) (0,7.97) (0,7.96)

S8 Uncertainty of impact of climate change on source yields

We have followed the Environment Agency guideline5 for this component applying the methodology to WAFU data for each WRZ. We have assumed a triangular distribution with the “best estimate” of loss equal to zero because the “mid” (50th percentile) scenario value is used in the resource calculation for the WRZ. The upper and lower limits are defined by the “wet” (95th percentile) and “dry” (5th percentile) climate change scenarios. The maximum loss (in 2025) is equal to the “mid” scenario value minus the “dry” scenario value whilst the maximum gain is

5 Ibid. 2

Page A5

SWW Water Resources Management Plan 2015 – 2040

taken to be the “mid” scenario value minus the “wet” scenario value. We have applied scaling factors from the guidelines for years either side of 2025.

Table A.3: S8 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 Best estimate 0.00 0.00 0.00 0.00 0.00 0.00 0.00 (Ml/d) Max loss 1.05 4.18 9.41 10.46 11.51 12.55 13.39 (Ml/d) Max gain -0.52 -2.09 -4.71 -5.23 -5.75 -6.28 -6.70 (Ml/d)

S9 Uncertainty of new sources

In PR09 two demand reduction options were included, which are equivalent to introducing new sources:

. water efficiency options . new tariff structure

Uncertainties in the estimates of the impact of these two options on DI were included in target headroom in PR09.

In PR14 there are no new sources proposed and hence no S9 component of target headroom is included.

Page A6

SWW Water Resources Management Plan 2015 – 2040

D1 Accuracy of sub-component data

The meters used to measure DI are subject to uncertainty, however they are checked and the readings are corrected. We have based projections of DI on these corrected values. Therefore the remaining uncertainty is that due to the calibration errors which is likely to be small. There is no evidence to suggest bias in one way or another. We have assumed the following:

. meter error is normally distributed around a mean of 0 Ml/d . 95% probability that the meter is recording within ±2.5%

Table A.4: D1 headroom uncertainty probability distribution summary data

Year DI 2.5% DI SD (Ml/d) 2012/13 141.495 3.54 1.80 2013/14 138.702 3.47 1.77 2014/15 136.974 3.42 1.75 2015/16 135.936 3.40 1.73 2016/17 134.942 3.37 1.72 2017/18 134.046 3.35 1.71 2018/19 133.280 3.33 1.70 2019/20 132.527 3.31 1.69 2020/21 131.764 3.29 1.68 2021/22 131.132 3.28 1.67 2022/23 130.563 3.26 1.67 2023/24 129.861 3.25 1.66 2024/25 129.234 3.23 1.65 2025/26 128.621 3.22 1.64 2026/27 128.085 3.20 1.63 2027/28 127.422 3.19 1.63 2028/29 126.786 3.17 1.62 2029/30 126.207 3.16 1.61 2030/31 125.656 3.14 1.60 2031/32 125.251 3.13 1.60 2032/33 125.036 3.13 1.59 2033/34 124.857 3.12 1.59 2034/35 124.668 3.12 1.59 2035/36 124.490 3.11 1.59 2036/37 124.325 3.11 1.59 2037/38 124.196 3.10 1.58 2038/39 124.224 3.11 1.58 2039/40 124.264 3.11 1.58

Page A7

SWW Water Resources Management Plan 2015 – 2040

D2 Demand forecast variation

A simple triangular distribution can be used to express the probability distribution, starting with zero forecast variation in 2012/13 and leading linearly to an assumed error of ±15% at the end of the planning period:

Table A.5: D2 headroom uncertainty probability distribution summary data

Year DI Min loss Most likely Max loss (Ml/d) (Ml/d) (Ml/d) (Ml/d) 2012/13 141.50 0.00 0 0.00 2013/14 138.70 -0.69 0 0.69 2014/15 136.97 -1.38 0 1.38 2015/16 135.94 -2.07 0 2.07 2016/17 134.94 -2.76 0 2.76 2017/18 134.05 -3.45 0 3.45 2018/19 133.28 -4.14 0 4.14 2019/20 132.53 -4.83 0 4.83 2020/21 131.76 -5.52 0 5.52 2021/22 131.13 -6.21 0 6.21 2022/23 130.56 -6.90 0 6.90 2023/24 129.86 -7.59 0 7.59 2024/25 129.23 -8.28 0 8.28 2025/26 128.62 -8.97 0 8.97 2026/27 128.08 -9.66 0 9.66 2027/28 127.42 -10.36 0 10.36 2028/29 126.79 -11.05 0 11.05 2029/30 126.21 -11.74 0 11.74 2030/31 125.66 -12.43 0 12.43 2031/32 125.25 -13.12 0 13.12 2032/33 125.04 -13.81 0 13.81 2033/34 124.86 -14.50 0 14.50 2034/35 124.67 -15.19 0 15.19 2035/36 124.49 -15.88 0 15.88 2036/37 124.32 -16.57 0 16.57 2037/38 124.20 -17.26 0 17.26 2038/39 124.22 -17.95 0 17.95 2039/40 124.26 -18.64 0 18.64

Page A8

SWW Water Resources Management Plan 2015 – 2040

D3 Uncertainty of impact of climate change on demand

We have used the results contained in the UKWIR report Impact of climate change on water demand6 which suggests a 1.0% increase in consumption attributable to climate change in 2039/40. We have assumed a potential variation of 20% either way for headroom purposes and have assumed a triangular distribution.

Table A.6: D3 headroom uncertainty probability distribution summary data

Year Increase in Min loss Most likely Max loss consumption (Ml/d) (Ml/d) (Ml/d) attributable to climate change (Ml/d) 2012/13 0.00 0.00 0 0.00 2013/14 0.03 -0.01 0 0.01 2014/15 0.05 -0.01 0 0.01 2015/16 0.10 -0.02 0 0.02 2016/17 0.15 -0.03 0 0.03 2017/18 0.20 -0.04 0 0.04 2018/19 0.24 -0.05 0 0.05 2019/20 0.28 -0.06 0 0.06 2020/21 0.31 -0.06 0 0.06 2021/22 0.35 -0.07 0 0.07 2022/23 0.38 -0.08 0 0.08 2023/24 0.42 -0.08 0 0.08 2024/25 0.45 -0.09 0 0.09 2025/26 0.48 -0.10 0 0.10 2026/27 0.51 -0.10 0 0.10 2027/28 0.54 -0.11 0 0.11 2028/29 0.56 -0.11 0 0.11 2029/30 0.59 -0.12 0 0.12 2030/31 0.62 -0.12 0 0.12 2031/32 0.66 -0.13 0 0.13 2032/33 0.69 -0.14 0 0.14 2033/34 0.72 -0.14 0 0.14 2034/35 0.76 -0.15 0 0.15 2035/36 0.79 -0.16 0 0.16 2036/37 0.83 -0.17 0 0.17 2037/38 0.86 -0.17 0 0.17 2038/39 0.90 -0.18 0 0.18 2039/40 0.94 -0.19 0 0.19

6 UKWIR, “Impact of climate change on water demand”, Ref 13/CL/04/12, 2013

Page A9

SWW Water Resources Management Plan 2015 – 2040

D4 Uncertainty of impact of demand management

In this Plan we assume that demand management measures save 0.75 Ml/day each year throughout the planning period. As these are savings, they are the equivalent of a supply scheme. In order to estimate savings due to demand management measures for each WRZ, we have calculated these values pro rata on the basis of forecast DI between the three WRZs. In 2012/13 the ratio of Roadford WRZ: Wimbleball WRZ : Colliford WRZ DI was 214.020 : 78.376 : 141.495, giving savings in the Colliford WRZ of 0.245 Ml/day each year.

We have assumed a triangular distribution and the most likely result is that the savings will be made and therefore this has a value of zero. We have used a potential variation of 10% either way for headroom purposes.

Table A.7: D4 headroom uncertainty probability distribution summary data

Year DI Min loss Most likely Max loss (Ml/d) (Ml/d) (Ml/d) (Ml/d) 2012/13 141.50 0.00 0 0.00 2013/14 138.70 -0.02 0 0.02 2014/15 136.97 -0.05 0 0.05 2015/16 135.94 -0.07 0 0.07 2016/17 134.94 -0.10 0 0.10 2017/18 134.05 -0.12 0 0.12 2018/19 133.28 -0.15 0 0.15 2019/20 132.53 -0.17 0 0.17 2020/21 131.76 -0.20 0 0.20 2021/22 131.13 -0.22 0 0.22 2022/23 130.56 -0.24 0 0.24 2023/24 129.86 -0.27 0 0.27 2024/25 129.23 -0.29 0 0.29 2025/26 128.62 -0.32 0 0.32 2026/27 128.08 -0.34 0 0.34 2027/28 127.42 -0.37 0 0.37 2028/29 126.79 -0.39 0 0.39 2029/30 126.21 -0.42 0 0.42 2030/31 125.66 -0.44 0 0.44 2031/32 125.25 -0.46 0 0.46 2032/33 125.04 -0.49 0 0.49 2033/34 124.86 -0.51 0 0.51 2034/35 124.67 -0.54 0 0.54 2035/36 124.49 -0.56 0 0.56 2036/37 124.32 -0.59 0 0.59 2037/38 124.20 -0.61 0 0.61 2038/39 124.22 -0.64 0 0.64 2039/40 124.26 -0.66 0 0.66

Page A10

SWW Water Resources Management Plan 2015 – 2040

Figure A.1 Colliford WRZ headroom uncertainty spreadsheet

Headroom Spreadsheet

Company Name South West Water Scenario Ref FINAL WRMP14 Version 2 Resource Zone Colliford WRZ Date 29/05/2014

Component Correlated Headroom Component (Ml/d)

With By Params 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40

Dependent Dependent

intermittent Component

Component Overlapping Continuous / Continuous S1 Vulnerable surfacewater licences Type Param. A Param. B Param. C

S1/1 S2 Vulnerable groundwater licences Type Param. A x1 Param. B x2 Param. C p1 Param. D p2 S2/1 S3 Time Limited Licences - These risks are now excluded by the EA Guidelines Type Param. A Param. B Param. C Param. D S3/1 S4 Bulk transfers Type Param. A Param. B Param. C Param. D S4/1 S5 Gradual pollution of sources causing a reduction in abstraction Type Param. A Min Param. B Best Param. C Max S5/1 S6 Accuracy of supply side data Type Param. A Param. B Param. C S6/1 Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 4.05 SD 4.03 SD 4.00 SD 4.00 SD 3.99 SD 3.98 SD 3.98 S6/2 Meter uncertainty for licence critical sources 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Type Param. A Param. B Param. C S6/3 Uncertainty for aquifer constrained groundwater sources Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 8.10 SD 8.07 SD 8.01 SD 7.99 SD 7.98 SD 7.97 SD 7.96 S6/4 Uncertainty for climate and catchment characteristics affecting surface waters 0.00 0.00 0.00 0.00 0.00 0.00 0.00 S8 Uncertainty of impact of climate change on Deployable Output Type Triangular Triangular Triangular Triangular Triangular Triangular Triangular Param. A -0.52 Min -2.09 Min -4.71 Min -5.23 Min -5.75 Min -6.28 Min -6.70 Param. B 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C 1.05 Max 4.18 Max 9.41 Max 10.46 Max 11.51 Max 12.55 Max 13.39 S8/1 Uncertainty of impact of climate change on source yields 0.17 0.70 1.57 1.74 1.92 2.09 2.23 S9 Uncertainty over New Sources Type Param. A Min Min Min Min Min Min Min Param. B Best Best Best Best Best Best Best Param. C Max Max Max Max Max Max Max S9/1 Type Param. A Min Min Min Min Min Min Min Param. B Best Best Best Best Best Best Best Param. C Max Max Max Max Max Max Max S9/1 D1 Accuracy of sub-component data Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 1.80 SD 1.73 SD 1.68 SD 1.64 SD 1.60 SD 1.59 SD 1.58 D1/1 Uncertainty of distribution input arising from meter inaccuracy 0.00 0.00 0.00 0.00 0.00 0.00 0.00 D2 Demand forecast variation Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -2.07 Min -5.52 Min -8.97 Min -12.43 Min -15.88 Min -18.64 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 2.07 Max 5.52 Max 8.97 Max 12.43 Max 15.88 Max 18.64 D2/1 Demand forecast variation. 0.00 0.00 0.00 0.00 0.00 0.00 Type Param. A Param. B Param. C

D2/2 D3 Uncertainty of impact of climate change on demand Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -0.02 Min -0.06 Min -0.10 Min -0.12 Min -0.16 Min -0.19 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 0.02 Max 0.06 Max 0.10 Max 0.12 Max 0.16 Max 0.19 D3/1 Effect of climate change on demand 0.00 0.00 0.00 0.00 0.00 0.00 D4 Uncertainty of impact of demand management Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -0.07 Min -0.20 Min -0.32 Min -0.44 Min -0.56 Min -0.66 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 0.07 Max 0.20 Max 0.32 Max 0.44 Max 0.56 Max 0.66 D4/1 Uncertainty of impact of demand management 0.00 0.00 0.00 0.00 0.00 0.00 Overlapping components 14 16 16 16 16 20 22 Group 1 Group 2 Group 3

2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 All uncertainties: 0.17 0.70 1.57 1.74 1.92 2.09 2.23

Page A11

SWW Water Resources Management Plan 2015 – 2040

Table A.8 Colliford WRZ headroom uncertainty @RISK output report

Outputs 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 Minimum -36.19 -46.06 -40.55 -44.96 -38.15 -40.06 -41.53 Maximum 34.58 36.05 39.22 39.60 44.95 45.57 46.18 Mean 0.17 0.70 1.57 1.74 1.92 2.09 2.23 Std Deviation 9.26 9.25 9.88 10.41 11.04 11.76 12.56 Variance 85.73 85.61 97.60 108.42 121.97 138.21 157.85 Skewness -0.010 -0.002 -0.035 -0.002 0.000 0.023 0.000 Kurtosis 3.048 3.053 2.971 2.928 2.899 2.935 2.895 Number of Errors 0 0 0 0 0 0 0 Mode -0.71 1.68 0.07 1.33 2.09 1.80 8.44 5.0% -15.08 -14.50 -14.81 -15.40 -16.40 -17.10 -18.07 10.0% -11.49 -11.11 -11.38 -11.58 -12.44 -12.80 -13.76 15.0% -9.31 -8.79 -8.80 -9.28 -9.58 -10.20 -10.80 20.0% -7.58 -7.07 -6.84 -7.22 -7.38 -7.91 -8.49 25.0% -5.99 -5.55 -5.06 -5.35 -5.63 -5.95 -6.51 30.0% -4.58 -4.09 -3.57 -3.80 -3.94 -4.21 -4.63 35.0% -3.37 -2.91 -2.16 -2.35 -2.44 -2.64 -2.75 40.0% -2.23 -1.71 -0.84 -0.92 -0.91 -1.06 -1.09 45.0% -1.05 -0.52 0.40 0.42 0.54 0.54 0.51 50.0% 0.14 0.64 1.60 1.74 1.97 1.92 2.14 55.0% 1.33 1.78 2.88 3.21 3.29 3.45 3.84 60.0% 2.55 3.04 4.21 4.54 4.80 5.02 5.62 65.0% 3.68 4.18 5.47 5.88 6.26 6.53 7.29 70.0% 4.94 5.45 6.84 7.35 7.75 8.30 8.92 75.0% 6.34 6.92 8.19 8.78 9.45 10.13 10.75 80.0% 7.87 8.46 9.78 10.54 11.28 12.00 12.91 85.0% 9.79 10.33 11.67 12.63 13.46 14.38 15.50 90.0% 12.07 12.58 14.19 15.05 16.11 17.45 18.46 95.0% 15.53 15.96 17.74 18.74 20.09 21.77 22.67

Blue cells are 85th, reducing to 70th, percentile figures used in the supply demand balance assessment.

Page A12

SWW Water Resources Management Plan 2015 – 2040

Figure A.2 Colliford WRZ headroom uncertainty @RISK output chart

Headroom Uncertainty - Colliford WRZ (Ml/day) HEADROOM 5% Perc 10% Perc 15% Perc 20% Perc 25% Perc 30% Perc 35% Perc 40% Perc 45% Perc 50% Perc 55% Perc 60% Perc 65% Perc 70% Perc 75% Perc 80% Perc 85% Perc 90% Perc 95% Perc 25.0

20.0

15.0

10.0

5.0

0.0

2012/13 2020/21 2025/26 2030/31 3035/36 3039/40 -5.0 2015/16

-10.0

-15.0

-20.0

Page A13

SWW Water Resources Management Plan 2015 – 2040

Colliford WRZ headroom uncertainty @RISK sensitivity analysis - tornado graphs

The tornado graphs below show how sensitive the headroom is to each of the headroom components for the 2012/13, 2015/16, 2025/26 and 2039/40 time horizons.

Figure A.3 Colliford WRZ sensitivity tornado graph 2012/13

Sensitivity Tornado S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D1/1 Uncertainty of distribution input arising from meter inaccuracy S8/1 Uncertainty of impact of climate change on

source yields

5

5 0

-

20 15 10

10 15 20

- - - Mean of 2012/13

Figure A.4 Colliford WRZ sensitivity tornado graph 2015/16

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D1/1 Uncertainty of distribution input arising from meter inaccuracy S8/1 Uncertainty of impact of climate change on source yields

D2/1 Demand forecast variation.

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

5

0 5

-

15 10 20

15 20 10 25

- - - Mean of 2015/16

Page A14

SWW Water Resources Management Plan 2015 – 2040

Figure A.5 Colliford WRZ sensitivity tornado graph 2025/26

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D2/1 Demand forecast variation.

S8/1 Uncertainty of impact of climate change on source yields D1/1 Uncertainty of distribution input arising from meter inaccuracy

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

5

0 5

-

20 15 10

10 15 20 25

- - -

Mean of 2025/26

Figure A.6 Colliford WRZ sensitivity tornado graph 2039/40

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

D2/1 Demand forecast variation.

S6/2 Meter uncertainty for licence critical sources

S8/1 Uncertainty of impact of climate change on source yields D1/1 Uncertainty of distribution input arising from meter inaccuracy

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

5

0 5

-

15 10 20

10 15 20 25

- - -

Mean of 2039/40

Page A15

SWW Water Resources Management Plan 2015 – 2040

Colliford WRZ headroom uncertainty @RISK sensitivity analysis – climate change components

In accordance with Environment Agency proposals7, headroom was calculated for the five scenarios:

. supply uncertainty excluding climate change (S1 – S7, S9) . demand uncertainty excluding climate change (D1, D2, D4) . impacts of climate change on supply (S8) . impacts of climate change on demand (D3) . all components (S1 – S9 inclusive, D1 – D4 inclusive)

The results are shown in the chart below for the 75th percentile. As can be seen from the chart, the total headroom is not simply a sum of the headroom calculated from groups of components. However, the chart does give an indication of the relative contribution each group of components makes to the overall headroom calculation.

As would be expected from inspection of the values of the D3 component of headroom (effect of climate change on demand), this makes a very small contribution to target headroom. The S8 component (the impacts of climate change on supply) makes an increasing contribution to the target headroom as we move through the planning period. However, by the end of the planning period the contribution to headroom from this component is still no more than the non-climate change impacts on supply.

Figure A.7 Colliford WRZ – graph showing climate change contribution to headroom uncertainty

Headroom Uncertainty - Colliford WRZ (Ml/day) Climate change contribution to headroom (75th percentile) 20

Impacts of climate change 18 on demand (D3):

16

Demand uncertainty 14 excluding climate change (D1, D2, D4): 12

Impacts of climate change 10 on supply (S8):

8

6 Supply uncertainty excluding climate change

Headroom Headroom uncertainty (Ml/day) (S1 - S7, S9): 4

2 All components

0

2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 2012/13

7 Environment Agency, “Climate change approaches in water resources planning – Overview of new methods”, Ref SC090017/SR3, 2012, page 63

Page A16

SWW Water Resources Management Plan 2015 – 2040

Section A2 Roadford WRZ headroom uncertainty

S1 Vulnerable surface water licences

No vulnerable surface water licences have been identified.

S2 Vulnerable groundwater licences

No vulnerable groundwater water licences have been identified.

S3 Time-limited licences

Environment Agency guidelines preclude these from the headroom analysis.

S4 Bulk imports

There are no bulk imports into the Roadford WRZ.

S5 Gradual pollution of sources causing a reduction in abstraction

No sources in the Roadford WRZ have been identified as being at risk.

S6 Accuracy of supply side data

S6/1 Uncertainty for yields constrained by pump capacity

In the groundwater DO assessments we have used actual pumping rates rather than nominal pumping capacities, hence this component does not apply.

S6/2 Meter uncertainty for licence critical sources

We have assumed that all sources are subject to meter uncertainty and that:

. meter error is normally distributed around a mean of 0 Ml/d . 95% probability that the meter is recording within ±5%

Table A.9: S6/2 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 WAFU (Ml/d) 258.194 251.460 246.076 245.431 244.786 244.141 243.626 5% of WAFU 12.910 12.573 12.304 12.272 12.239 12.207 12.181 SD 6.59 6.41 6.28 6.26 6.24 6.23 6.21 RiskNormal (0,6.59) (0,6.41) (0,6.28) (0,6.26) (0,6.24) (0,6.23) (0,6.21)

Page A17

SWW Water Resources Management Plan 2015 – 2040

S6/3 Uncertainty for aquifer constrained groundwater sources

In the Roadford WRZ, there are no groundwater sources currently constrained by the aquifer.

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

The accuracy of river flow measurements, even from well-maintained weirs, cannot be expected to be better than ±10%. However, this does not necessarily mean that DO can be estimated to within ±10%. In order to assess the effect of errors in flow measurement on DO it would be necessary to do many water resources model runs with perturbed river flow sequences.

However, it would be incorrect to adjust all sequences by +10% and then by -10% as all sequences would not be out by the same amount at the same time. It would also be incorrect to randomly adjust figures on a daily basis as, for example, Gunnislake would not be reading +9% of the correct value one day and -7% the next day.

The effect on DO could be estimated by ascribing a different adjustment factor to each flow sequence within the bounds of ±10% and then running the resources model. When this process has been repeated over and over again with different factors a distribution of DO would be found. However, this process is unworkable in practice.

We have therefore concluded that the uncertainly in DO of surface water sources should be estimated pragmatically by assuming that there is a 95% probability that the value is within ±10% of the estimated value.

. error is normally distributed around a mean of 0 Ml/d . 95% probability that the value is within ±10%

Table A.10: S6/4 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 WAFU (Ml/d) 241.987 235.261 229.892 229.261 228.630 228.000 227.496 10% of WAFU 24.199 23.526 22.989 22.926 22.863 22.800 22.750 SD 12.35 12.00 11.73 11.70 11.66 11.63 11.61 RiskNormal (0,12.35) (0,12.00) (0,11.73) (0,11.70) (0,11.66) (0,11.63) (0,11.61)

S8 Uncertainty of impact of climate change on source yields

The Environment Agency guideline8 has been followed for this component applying the methodology to WAFU data for each WRZ. We have assumed a triangular distribution with the “best estimate” of loss equal to zero because the “mid” (50th percentile) scenario value is used in the resource calculation for the WRZ. The upper and lower limits are defined by the “wet” (95th percentile) and “dry” (5th percentile) climate change scenarios. The maximum loss (in 2025) is equal to the

8 Ibid. 2

Page A18

SWW Water Resources Management Plan 2015 – 2040

“mid” scenario value minus the “dry” scenario value whilst the maximum gain is taken to be the “mid” scenario value minus the “wet” scenario value. We have applied scaling factors from the guidelines for years either side of 2025.

Table A.11: S8 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 Best estimate 0.00 0.00 0.00 0.00 0.00 0.00 0.00 (Ml/d) Max loss 1.41 5.63 12.66 14.07 15.47 16.88 18.01 (Ml/d) Max gain -0.90 -3.62 -8.15 -9.05 -9.95 -10.86 -11.58 (Ml/d)

S9 Uncertainty of new sources

In PR09 two demand reduction options were included, which are equivalent to introducing new sources:

. water efficiency options . new tariff structure

Uncertainties in the estimates of the impact of these two options on DI were included in target headroom in PR09.

In PR14 there are no new sources proposed and hence no S9 component of target headroom is included.

Page A19

SWW Water Resources Management Plan 2015 – 2040

D1 Accuracy of sub-component data

The meters used to measure DI are subject to uncertainty, however they are checked and we correct the readings. Projections of DI are based upon these corrected values. Therefore the remaining uncertainty is that due to the calibration errors which is likely to be small. There is no evidence to suggest bias in one way or another. We have assumed the following:

. meter error is normally distributed around a mean of 0 Ml/d loss . 95% probability that the meter is recording within ±2.5%

Table A.12: D1 headroom uncertainty probability distribution summary data

Year DI 2.5% DI SD (Ml/d) 2012/13 214.020 5.35 2.73 2013/14 209.478 5.24 2.67 2014/15 206.493 5.16 2.63 2015/16 204.719 5.12 2.61 2016/17 203.111 5.08 2.59 2017/18 201.703 5.04 2.57 2018/19 200.464 5.01 2.56 2019/20 199.346 4.98 2.54 2020/21 198.216 4.96 2.53 2021/22 197.261 4.93 2.52 2022/23 196.415 4.91 2.51 2023/24 195.365 4.88 2.49 2024/25 194.401 4.86 2.48 2025/26 193.432 4.84 2.47 2026/27 192.570 4.81 2.46 2027/28 191.486 4.79 2.44 2028/29 190.432 4.76 2.43 2029/30 189.455 4.74 2.42 2030/31 188.523 4.71 2.40 2031/32 187.795 4.69 2.40 2032/33 187.327 4.68 2.39 2033/34 186.909 4.67 2.38 2034/35 186.471 4.66 2.38 2035/36 186.048 4.65 2.37 2036/37 185.642 4.64 2.37 2037/38 185.287 4.63 2.36 2038/39 185.166 4.63 2.36 2039/40 185.059 4.63 2.36

Page A20

SWW Water Resources Management Plan 2015 – 2040

D2 Demand forecast variation

A simple triangular distribution can be used to express the probability distribution, starting with zero forecast variation in 2012/13 and leading linearly to an assumed error of ±15% at the end of the planning period:

Table A.13: D2 headroom uncertainty probability distribution summary data

Year DI Min loss Most likely Max loss (Ml/d) (Ml/d) (Ml/d) (Ml/d) 2012/13 214.02 0.00 0 0.00 2013/14 209.48 -1.03 0 1.03 2014/15 206.49 -2.06 0 2.06 2015/16 204.72 -3.08 0 3.08 2016/17 203.11 -4.11 0 4.11 2017/18 201.70 -5.14 0 5.14 2018/19 200.46 -6.17 0 6.17 2019/20 199.35 -7.20 0 7.20 2020/21 198.22 -8.22 0 8.22 2021/22 197.26 -9.25 0 9.25 2022/23 196.42 -10.28 0 10.28 2023/24 195.37 -11.31 0 11.31 2024/25 194.40 -12.34 0 12.34 2025/26 193.43 -13.37 0 13.37 2026/27 192.57 -14.39 0 14.39 2027/28 191.49 -15.42 0 15.42 2028/29 190.43 -16.45 0 16.45 2029/30 189.45 -17.48 0 17.48 2030/31 188.52 -18.51 0 18.51 2031/32 187.80 -19.53 0 19.53 2032/33 187.33 -20.56 0 20.56 2033/34 186.91 -21.59 0 21.59 2034/35 186.47 -22.62 0 22.62 2035/36 186.05 -23.65 0 23.65 2036/37 185.64 -24.67 0 24.67 2037/38 185.29 -25.70 0 25.70 2038/39 185.17 -26.73 0 26.73 2039/40 185.06 -27.76 0 27.76

Page A21

SWW Water Resources Management Plan 2015 – 2040

D3 Uncertainty of impact of climate change on demand

We have used the results contained in the UKWIR report Impact of climate change on water demand9 which suggests a 1.0% increase in consumption attributable to climate change in 2039/40. We have assumed a potential variation of 20% either way for headroom purposes and have assumed a triangular distribution.

Table A.14: D3 headroom uncertainty probability distribution summary data

Year Increase in Min loss Most likely Max loss consumption (Ml/d) (Ml/d) (Ml/d) attributable to climate change (Ml/d) 2012/13 0.00 0.00 0 0.00 2013/14 0.04 -0.01 0 0.01 2014/15 0.07 -0.01 0 0.01 2015/16 0.16 -0.03 0 0.03 2016/17 0.23 -0.05 0 0.05 2017/18 0.30 -0.06 0 0.06 2018/19 0.36 -0.07 0 0.07 2019/20 0.42 -0.08 0 0.08 2020/21 0.48 -0.10 0 0.10 2021/22 0.53 -0.11 0 0.11 2022/23 0.58 -0.12 0 0.12 2023/24 0.63 -0.13 0 0.13 2024/25 0.68 -0.14 0 0.14 2025/26 0.72 -0.14 0 0.14 2026/27 0.77 -0.15 0 0.15 2027/28 0.81 -0.16 0 0.16 2028/29 0.85 -0.17 0 0.17 2029/30 0.89 -0.18 0 0.18 2030/31 0.94 -0.19 0 0.19 2031/32 0.99 -0.20 0 0.20 2032/33 1.04 -0.21 0 0.21 2033/34 1.09 -0.22 0 0.22 2034/35 1.14 -0.23 0 0.23 2035/36 1.19 -0.24 0 0.24 2036/37 1.24 -0.25 0 0.25 2037/38 1.30 -0.26 0 0.26 2038/39 1.35 -0.27 0 0.27 2039/40 1.40 -0.28 0 0.28

9 UKWIR, “Impact of climate change on water demand”, Ref 13/CL/04/12, 2013

Page A22

SWW Water Resources Management Plan 2015 – 2040

D4 Uncertainty of impact of demand management

In this Plan we assume that demand management measures save 0.75 Ml/day each year throughout the planning period, over our area. As these are savings, they are the equivalent of a supply scheme. In order to estimate savings due to demand management measures for each WRZ, we have calculated these values pro rata on the basis of forecast DI between the three WRZs. In 2012/13 the ratio of Roadford WRZ: Wimbleball WRZ : Colliford WRZ DI was 214.020 : 78.376 : 141.495, giving savings in the Roadford WRZ of 0.370 Ml/day each year.

We have assumed a triangular distribution and the most likely result is that the savings will be made and therefore this has a value of zero. We have used a potential variation of 10% either way for headroom purposes.

Table A.15: D4 headroom uncertainty probability distribution summary data

Year DI Min loss Most likely Max loss (Ml/d) (Ml/d) (Ml/d) (Ml/d) 2012/13 214.02 0.00 0 0.00 2013/14 209.48 -0.04 0 0.04 2014/15 206.49 -0.07 0 0.07 2015/16 204.72 -0.11 0 0.11 2016/17 203.11 -0.15 0 0.15 2017/18 201.70 -0.18 0 0.18 2018/19 200.46 -0.22 0 0.22 2019/20 199.35 -0.26 0 0.26 2020/21 198.22 -0.30 0 0.30 2021/22 197.26 -0.33 0 0.33 2022/23 196.42 -0.37 0 0.37 2023/24 195.37 -0.41 0 0.41 2024/25 194.40 -0.44 0 0.44 2025/26 193.43 -0.48 0 0.48 2026/27 192.57 -0.52 0 0.52 2027/28 191.49 -0.55 0 0.55 2028/29 190.43 -0.59 0 0.59 2029/30 189.45 -0.63 0 0.63 2030/31 188.52 -0.67 0 0.67 2031/32 187.80 -0.70 0 0.70 2032/33 187.33 -0.74 0 0.74 2033/34 186.91 -0.78 0 0.78 2034/35 186.47 -0.81 0 0.81 2035/36 186.05 -0.85 0 0.85 2036/37 185.64 -0.89 0 0.89 2037/38 185.29 -0.92 0 0.92 2038/39 185.17 -0.96 0 0.96 2039/40 185.06 -1.00 0 1.00

Page A23

SWW Water Resources Management Plan 2015 – 2040

Figure A.8 Roadford WRZ headroom uncertainty spreadsheet

Headroom Spreadsheet

Company Name South West Water Scenario Ref FINAL WRMP14 Version 2 Resource Zone Roadford WRZ Date 28/05/2014

Component Correlated Headroom Component (Ml/d)

With By Params 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40

Dependent Dependent

intermittent Component

Component Overlapping Continuous / Continuous S1 Vulnerable surfacewater licences Type Param. A Param. B Param. C

S1/1 S2 Vulnerable groundwater licences Type Param. A x1 Param. B x2 Param. C p1 Param. D p2 S2/1 S3 Time Limited Licences - These risks are now excluded by the EA Guidelines Type Param. A Param. B Param. C Param. D S3/1 S4 Bulk transfers Type Param. A Param. B Param. C Param. D S4/1 S5 Gradual pollution of sources causing a reduction in abstraction Type Param. A Min Param. B Best Param. C Max S5/1 S6 Accuracy of supply side data Type Param. A Param. B Param. C S6/1 Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 6.59 SD 6.41 SD 6.28 SD 6.26 SD 6.24 SD 6.23 SD 6.21 S6/2 Meter uncertainty for licence critical sources 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Type Param. A Param. B Param. C S6/3 Uncertainty for aquifer constrained groundwater sources Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 12.35 SD 12.00 SD 11.73 SD 11.70 SD 11.66 SD 11.63 SD 11.61 S6/4 Uncertainty for climate and catchment characteristics affecting surface waters 0.00 0.00 0.00 0.00 0.00 0.00 0.00 S8 Uncertainty of impact of climate change on Deployable Output Type Triangular Triangular Triangular Triangular Triangular Triangular Triangular Param. A -0.90 Min -3.62 Min -8.14 Min -9.05 Min -9.95 Min -10.86 Min -11.58 Param. B 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C 1.41 Max 5.63 Max 12.66 Max 14.07 Max 15.47 Max 16.88 Max 18.01 S8/1 Uncertainty of impact of climate change on source yields 0.17 0.67 1.51 1.67 1.84 2.01 2.14 S9 Uncertainty over New Sources Type Param. A Min Min Min Min Min Min Min Param. B Best Best Best Best Best Best Best Param. C Max Max Max Max Max Max Max S9/1 Type Param. A Min Min Min Min Min Min Min Param. B Best Best Best Best Best Best Best Param. C Max Max Max Max Max Max Max S9/1 D1 Accuracy of sub-component data Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 2.73 SD 2.61 SD 2.53 SD 2.47 SD 2.40 SD 2.37 SD 2.36 D1/1 Uncertainty of distribution input arising from meter inaccuracy 0.00 0.00 0.00 0.00 0.00 0.00 0.00 D2 Demand forecast variation Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -3.08 Min -8.22 Min -13.37 Min -18.51 Min -23.65 Min -27.76 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 3.08 Max 8.22 Max 13.37 Max 18.51 Max 23.65 Max 27.76 D2/1 Demand forecast variation. 0.00 0.00 0.00 0.00 0.00 0.00 Type Param. A Param. B Param. C

D2/2 D3 Uncertainty of impact of climate change on demand Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -0.03 Min -0.10 Min -0.14 Min -0.19 Min -0.24 Min -0.28 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 0.03 Max 0.10 Max 0.14 Max 0.19 Max 0.24 Max 0.28 D3/1 Effect of climate change on demand 0.00 0.00 0.00 0.00 0.00 0.00 D4 Uncertainty of impact of demand management Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -0.11 Min -0.30 Min -0.48 Min -0.67 Min -0.85 Min -1.00 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 0.11 Max 0.30 Max 0.48 Max 0.67 Max 0.85 Max 1.00 D4/1 Uncertainty of impact of demand management 0.00 0.00 0.00 0.00 0.00 0.00 Overlapping components 14 16 16 16 16 20 22 Group 1 Group 2 Group 3

2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 All uncertainties: 0.17 0.67 1.51 1.67 1.84 2.01 2.14

Page A24

SWW Water Resources Management Plan 2015 – 2040

Table A.16: Roadford WRZ headroom uncertainty @RISK output report

Outputs 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 Minimum -55.10 -48.59 -59.76 -58.29 -56.40 -70.71 -64.78 Maximum 50.13 56.61 56.39 62.72 62.59 75.91 76.86 Mean 0.17 0.67 1.51 1.67 1.84 2.01 2.14 Std Deviation 14.21 14.06 14.59 15.27 16.10 17.49 18.65 Variance 201.93 197.56 212.80 233.28 259.30 305.83 347.88 Skewness 0.009 -0.026 0.015 -0.040 0.013 0.030 0.004 Kurtosis 2.976 2.991 3.038 2.989 3.018 3.015 2.885 Number of Errors 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Mode 0.59 -3.15 1.93 -0.73 -2.88 -1.88 2.79 5.0% -23.16 -22.39 -22.27 -23.50 -24.76 -26.82 -28.75 10.0% -18.01 -17.44 -17.08 -18.09 -18.58 -20.61 -21.86 15.0% -14.57 -14.05 -13.66 -14.33 -14.78 -16.24 -17.28 20.0% -11.82 -11.14 -10.78 -11.31 -11.73 -12.66 -13.60 25.0% -9.50 -8.81 -8.26 -8.54 -9.07 -9.86 -10.55 30.0% -7.37 -6.80 -6.11 -6.20 -6.65 -7.27 -7.72 35.0% -5.27 -4.78 -4.09 -4.10 -4.37 -4.78 -5.05 40.0% -3.44 -3.01 -2.32 -2.06 -2.31 -2.54 -2.72 45.0% -1.67 -1.05 -0.36 -0.19 -0.38 -0.35 -0.13 50.0% 0.07 0.79 1.45 1.70 1.77 1.92 2.19 55.0% 1.89 2.61 3.21 3.75 4.00 4.20 4.55 60.0% 3.72 4.33 5.09 5.63 6.06 6.46 7.01 65.0% 5.65 6.22 7.13 7.74 8.11 8.78 9.47 70.0% 7.60 8.20 9.21 9.82 10.32 11.20 12.02 75.0% 9.64 10.15 11.38 12.10 12.69 13.79 14.66 80.0% 12.09 12.50 13.82 14.63 15.40 16.69 17.94 85.0% 14.83 15.18 16.63 17.66 18.42 20.04 21.52 90.0% 18.66 18.48 20.14 21.24 22.24 24.26 26.37 95.0% 23.72 23.66 25.53 26.38 28.18 31.04 33.13

Blue cells are 85th, reducing to 70th, percentile figures used in the supply demand balance assessment.

Page A25

SWW Water Resources Management Plan 2015 – 2040

Figure A.9 Roadford WRZ headroom uncertainty @RISK output chart

Headroom Uncertainty - Roadford WRZ (Ml/day)

Headroom 5% Perc 10% Perc 15% Perc 20% Perc 25% Perc 30% Perc 35% Perc 40% Perc 45% Perc

50% Perc 55% Perc 60% Perc 65% Perc 70% Perc 75% Perc 80% Perc 85% Perc 90% Perc 95% Perc 35.0

30.0

25.0

20.0

15.0

10.0

5.0

0.0

-5.0

2012/13 2035/36 2039/40 2015/16 2020/21 2025/26 2030/31

-10.0

-15.0

-20.0

-25.0

-30.0

Page A26

SWW Water Resources Management Plan 2015 – 2040

Roadford WRZ headroom uncertainty @RISK sensitivity analysis - tornado graphs

The tornado graphs below show how sensitive the headroom is to each of the headroom components for the 2012/13, 2015/16, 2025/26 and 2039/40 time horizons.

Figure A.10 Roadford WRZ sensitivity tornado graph 2012/13

Sensitivity Tornado S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D1/1 Uncertainty of distribution input arising from meter inaccuracy S8/1 Uncertainty of impact of climate change on source

yields

0

10 30 20

30 40 10 20

- - - Mean of 2012/13

Figure A.11 Roadford WRZ sensitivity tornado graph 2015/16

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D1/1 Uncertainty of distribution input arising from meter inaccuracy S8/1 Uncertainty of impact of climate change on source yields

D2/1 Demand forecast variation.

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

0

20 10 30

30 10 20

- - - Mean of 2015/16

Page A27

SWW Water Resources Management Plan 2015 – 2040

Figure A.12 Roadford WRZ sensitivity tornado graph 2025/26

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D2/1 Demand forecast variation.

S8/1 Uncertainty of impact of climate change on source yields D1/1 Uncertainty of distribution input arising from meter inaccuracy

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

0

30 20 10

10 30 20

- - -

Mean of 2025/26

Figure A.13 Roadford WRZ sensitivity tornado graph 2039/40

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

D2/1 Demand forecast variation.

S6/2 Meter uncertainty for licence critical sources

S8/1 Uncertainty of impact of climate change on source yields D1/1 Uncertainty of distribution input arising from meter inaccuracy

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

0

30 20 10

10 20 30 40

- - -

Mean of 2039/40

Page A28

SWW Water Resources Management Plan 2015 – 2040

Roadford WRZ headroom uncertainty @RISK sensitivity analysis – climate change components

In accordance with Environment Agency proposals10, headroom was calculated for the five scenarios:

. supply uncertainty excluding climate change (S1 – S7, S9) . demand uncertainty excluding climate change (D1, D2, D4) . impacts of climate change on supply (S8) . impacts of climate change on demand (D3) . all components (S1 – S9 inclusive, D1 – D4 inclusive)

The results are shown in the chart below for the 75th percentile. As can be seen from the chart, the total headroom is not simply a sum of the headroom calculated from groups of components. However, the chart does give an indication of the relative contribution each group of components makes to the overall headroom calculation.

As would be expected from inspection of the values of the D3 component of headroom (effect of climate change on demand), this makes a very small contribution to target headroom. The S8 component (the impacts of climate change on supply) makes an increasing contribution to the target headroom as we move through the planning period. However, by the end of the planning period the contribution to target headroom from this component is still no more than the non- climate change impacts on supply.

Figure A.14 Roadford WRZ – graph showing climate change contribution to headroom uncertainty

Headroom Uncertainty - Roadford WRZ (Ml/day) Climate change contribution to headroom (75th percentile) 30

Impacts of climate change on demand (D3):

25

Demand uncertainty excluding climate change 20 (D1, D2, D4):

Impacts of climate change on 15 supply (S8):

10 Supply uncertainty excluding

climate change (S1 - S7, S9): Headroom Headroom uncertainty (Ml/day)

5 All components

0

2012/13 2025/26 2039/40 2020/21 2030/31 2035/36 2015/16

10 Ibid. 7, page 63

Page A29

SWW Water Resources Management Plan 2015 – 2040

Section A3 Wimbleball WRZ headroom uncertainty

S1 Vulnerable surface water licences

No vulnerable surface water licences have been identified.

S2 Vulnerable groundwater licences

No vulnerable groundwater water licences have been identified.

S3 Time-limited licences

Environment Agency guidelines preclude these from the headroom analysis.

S4 Bulk imports

There are no bulk imports into the Wimbleball WRZ.

S5 Gradual pollution of sources causing a reduction in abstraction

No sources in the Wimbleball WRZ have been identified as being at risk.

S6 Accuracy of supply side data

S6/1 Uncertainty for yields constrained by pump capacity

In the groundwater Deployable Output (DO) assessments actual pumping rates have been used rather than nominal pumping capacities hence this component does not apply.

S6/2 Meter uncertainty for licence critical sources

It is assumed that all sources are subject to meter uncertainty. We have assumed the following:

. meter error is normally distributed around a mean of 0 Ml/d . 95% probability that the meter is recording within ±5%

Page A30

SWW Water Resources Management Plan 2015 – 2040

Table A.17: S6/2 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 WAFU (Ml/d) 89.325 89.211 89.021 88.983 88.945 88.907 88.877 5% of WAFU 4.466 4.461 4.451 4.449 4.447 4.445 4.444 SD 2.28 2.28 2.27 2.27 2.27 2.27 2.27 RiskNormal (0,2.28) (0,2.28) (0,2.27) (0,2.27) (0,2.27) (0,2.27) (0,2.27)

S6/3 Uncertainty for aquifer constrained groundwater sources

Sufficient operational and test data are available for those aquifer constrained sources to have a drought curve which can be accurately positioned to give a high confidence of source performance; therefore we have not considered this component further in the headroom analysis.

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

The accuracy of river flow measurements, even from well maintained weirs, cannot be expected to be better than ±10%. However, this does not necessarily mean that DO can be estimated to within ±10%. In order to assess the effect of errors in flow measurement on DO it would be necessary to do many water resources model runs with perturbed river flow sequences.

However, it would be incorrect to adjust all sequences by +10% and then by -10% as all sequences would not be out by the same amount at the same time. It would also be incorrect to randomly adjust figures on a daily basis as, for example, Thorverton would not be reading +9% of the correct value one day and -7% the next day.

The effect on DO could be estimated by ascribing a different adjustment factor to each flow sequence within the bounds of ±10% and then running the resources model. When this process has been repeated over and over again with different factors a distribution of DO would be found. However, this process is unworkable in practice.

We have therefore concluded that the uncertainly in DO of surface water sources should be estimated pragmatically by assuming that there is a 95% probability that the value is within ±10% of the estimated value.

. error is normally distributed around a mean of 0 Ml/d . 95% probability that the value is within ±10%

Table A.18: S6/4 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 WAFU (Ml/d) 61.049 61.008 60.940 61.023 61.106 61.190 61.257 10% of WAFU 6.105 6.101 6.094 6.102 6.111 6.119 6.126 SD 3.11 3.11 3.11 3.11 3.12 3.12 3.13 RiskNormal (0,3.11) (0,3.11) (0,3.11) (0,3.11) (0,3.12) (0,3.12) (0,3.13)

Page A31

SWW Water Resources Management Plan 2015 – 2040

S8 Uncertainty of impact of climate change on source yields

We have followed the Environment Agency guideline11 for this component applying the methodology to WAFU data for each WRZ. We have assumed that there is a triangular distribution with the “best estimate” of loss equal to zero because the “mid” (50th percentile) scenario value is used in the resource calculation for the WRZ. The upper and lower limits are defined by the “wet” (95th percentile) and “dry” (5th percentile) climate change scenarios. The maximum loss (in 2025) is equal to the “mid” scenario value minus the “dry” scenario value whilst the maximum gain is taken to be the “mid” scenario value minus the “wet” scenario value. Scaling factors from the guidelines are applied for years either side of 2025.

Table A.19: S8 headroom uncertainty probability distribution summary data

Year 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 Best estimate 0.00 0.00 0.00 0.00 0.00 0.00 0.00 (Ml/d) Max loss 0.37 1.48 3.34 3.71 4.08 4.45 4.75 (Ml/d) Max gain -0.05 -0.20 -0.46 -0.51 -0.56 -0.61 -0.65 (Ml/d)

S9 Uncertainty of new sources

At the 2009 Periodic Review (PR09) two demand reduction options were included, which are equivalent to introducing new sources:

. water efficiency options . new tariff structure

Uncertainties in the estimates of the impact of these two options on Distribution Input (DI) were included in target headroom in PR09.

In the 2014 Periodic Review (PR14) there are no new sources proposed and hence no S9 component of target headroom is included.

11 Ibid. 2

Page A32

SWW Water Resources Management Plan 2015 – 2040

D1 Accuracy of sub-component data

The meters used to measure DI are subject to uncertainty, however we check them and the readings are corrected. Our projections of DI are based upon these corrected values. Therefore the remaining uncertainty is that due to the calibration errors which is likely to be small. There is no evidence to suggest bias in one way or another. We have therefore assumed the following:

. meter error is normally distributed around a mean of 0 Ml/d loss . 95% probability that the meter is recording within ±2.5%

Table A.20: D1 headroom uncertainty probability distribution summary data

Year DI 2.5% DI SD (Ml/d) 2012/13 78.376 1.96 1.00 2013/14 76.630 1.92 0.98 2014/15 75.448 1.89 0.96 2015/16 74.786 1.87 0.95 2016/17 74.197 1.85 0.95 2017/18 73.690 1.84 0.94 2018/19 73.275 1.83 0.93 2019/20 72.914 1.82 0.93 2020/21 72.567 1.81 0.93 2021/22 72.301 1.81 0.92 2022/23 72.050 1.80 0.92 2023/24 71.741 1.79 0.92 2024/25 71.466 1.79 0.91 2025/26 71.192 1.78 0.91 2026/27 70.957 1.77 0.91 2027/28 70.651 1.77 0.90 2028/29 70.359 1.76 0.90 2029/30 70.097 1.75 0.89 2030/31 69.847 1.75 0.89 2031/32 69.675 1.74 0.89 2032/33 69.585 1.74 0.89 2033/34 69.510 1.74 0.89 2034/35 69.431 1.74 0.89 2035/36 69.355 1.73 0.88 2036/37 69.284 1.73 0.88 2037/38 69.233 1.73 0.88 2038/39 69.252 1.73 0.88 2039/40 69.275 1.73 0.88

Page A33

SWW Water Resources Management Plan 2015 – 2040

D2 Demand forecast variation

A simple triangular distribution can be used to express the probability distribution, starting with zero forecast variation in 2012/13 and leading linearly to an assumed error of ±15% at the end of the planning period:

Table A.21: D2 headroom uncertainty probability distribution summary data

Year DI Min loss Most likely Max loss (Ml/d) (Ml/d) (Ml/d) (Ml/d) 2012/13 78.38 0.00 0 0.00 2013/14 76.63 -0.38 0 0.38 2014/15 75.45 -0.77 0 0.77 2015/16 74.79 -1.15 0 1.15 2016/17 74.20 -1.54 0 1.54 2017/18 73.69 -1.92 0 1.92 2018/19 73.27 -2.31 0 2.31 2019/20 72.91 -2.69 0 2.69 2020/21 72.57 -3.08 0 3.08 2021/22 72.30 -3.46 0 3.46 2022/23 72.05 -3.85 0 3.85 2023/24 71.74 -4.23 0 4.23 2024/25 71.47 -4.62 0 4.62 2025/26 71.19 -5.00 0 5.00 2026/27 70.96 -5.39 0 5.39 2027/28 70.65 -5.77 0 5.77 2028/29 70.36 -6.16 0 6.16 2029/30 70.10 -6.54 0 6.54 2030/31 69.85 -6.93 0 6.93 2031/32 69.67 -7.31 0 7.31 2032/33 69.59 -7.70 0 7.70 2033/34 69.51 -8.08 0 8.08 2034/35 69.43 -8.47 0 8.47 2035/36 69.35 -8.85 0 8.85 2036/37 69.28 -9.24 0 9.24 2037/38 69.23 -9.62 0 9.62 2038/39 69.25 -10.01 0 10.01 2039/40 69.28 -10.39 0 10.39

Page A34

SWW Water Resources Management Plan 2015 – 2040

D3 Uncertainty of impact of climate change on demand

We have used the results contained in the UKWIR report Impact of climate change on water demand12 which suggests a 1.0% increase in consumption attributable to climate change in 2039/40. We have assumed a potential variation of 20% either way for headroom purposes and have assumed a triangular distribution.

Table A.22: D3 headroom uncertainty probability distribution summary data

Year Increase in Min loss Most likely Max loss consumption (Ml/d) (Ml/d) (Ml/d) attributable to climate change (Ml/d) 2012/13 0.00 0.00 0 0.00 2013/14 0.01 0.00 0 0.00 2014/15 0.03 -0.01 0 0.01 2015/16 0.06 -0.01 0 0.01 2016/17 0.09 -0.02 0 0.02 2017/18 0.11 -0.02 0 0.02 2018/19 0.14 -0.03 0 0.03 2019/20 0.16 -0.03 0 0.03 2020/21 0.18 -0.04 0 0.04 2021/22 0.20 -0.04 0 0.04 2022/23 0.22 -0.04 0 0.04 2023/24 0.24 -0.05 0 0.05 2024/25 0.26 -0.05 0 0.05 2025/26 0.28 -0.06 0 0.06 2026/27 0.30 -0.06 0 0.06 2027/28 0.31 -0.06 0 0.06 2028/29 0.33 -0.07 0 0.07 2029/30 0.34 -0.07 0 0.07 2030/31 0.36 -0.07 0 0.07 2031/32 0.38 -0.08 0 0.08 2032/33 0.40 -0.08 0 0.08 2033/34 0.42 -0.08 0 0.08 2034/35 0.44 -0.09 0 0.09 2035/36 0.46 -0.09 0 0.09 2036/37 0.48 -0.10 0 0.10 2037/38 0.50 -0.10 0 0.10 2038/39 0.52 -0.10 0 0.10 2039/40 0.54 -0.11 0 0.11

12 UKWIR, “Impact of climate change on water demand”, Ref 13/CL/04/12, 2013

Page A35

SWW Water Resources Management Plan 2015 – 2040

D4 Uncertainty of impact of demand management

In this Plan we assume that demand management measures save 0.75 Ml/day each year throughout the planning period. As these are savings, they are the equivalent of a supply scheme. In order to estimate savings due to demand management measures for each WRZ, we have calculated these values pro rata on the basis of forecast DI between the three WRZs. In 2012/13 the ratio of Roadford WRZ : Wimbleball WRZ : Colliford WRZ DI was 214.020 : 78.376 : 141.495, giving savings in the Wimbleball WRZ of 0.135 Ml/day each year.

We have assumed a triangular distribution and the most likely result is that the savings will be made and therefore this has a value of zero. We have used a potential variation of 10% either way for headroom purposes.

Table A.23: D4 headroom uncertainty probability distribution summary data

Year DI Min loss Most likely Max loss (Ml/d) (Ml/d) (Ml/d) (Ml/d) 2012/13 78.38 0.00 0 0.00 2013/14 76.63 -0.01 0 0.01 2014/15 75.45 -0.03 0 0.03 2015/16 74.79 -0.04 0 0.04 2016/17 74.20 -0.05 0 0.05 2017/18 73.69 -0.07 0 0.07 2018/19 73.27 -0.08 0 0.08 2019/20 72.91 -0.09 0 0.09 2020/21 72.57 -0.11 0 0.11 2021/22 72.30 -0.12 0 0.12 2022/23 72.05 -0.14 0 0.14 2023/24 71.74 -0.15 0 0.15 2024/25 71.47 -0.16 0 0.16 2025/26 71.19 -0.18 0 0.18 2026/27 70.96 -0.19 0 0.19 2027/28 70.65 -0.20 0 0.20 2028/29 70.36 -0.22 0 0.22 2029/30 70.10 -0.23 0 0.23 2030/31 69.85 -0.24 0 0.24 2031/32 69.67 -0.26 0 0.26 2032/33 69.59 -0.27 0 0.27 2033/34 69.51 -0.28 0 0.28 2034/35 69.43 -0.30 0 0.30 2035/36 69.35 -0.31 0 0.31 2036/37 69.28 -0.33 0 0.33 2037/38 69.23 -0.34 0 0.34 2038/39 69.25 -0.35 0 0.35 2039/40 69.28 -0.37 0 0.37

Page A36

SWW Water Resources Management Plan 2015 – 2040

Figure A.15 Wimbleball WRZ headroom uncertainty spreadsheet

Headroom Spreadsheet

Company Name South West Water Scenario Ref FINAL WRMP14 Version 2 Resource Zone Wimbleball WRZ Date 28/05/2014

Component Correlated Headroom Component (Ml/d)

With By Params 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40

Dependent Dependent

intermittent Component

Component Overlapping Continuous / Continuous S1 Vulnerable surfacewater licences Type Param. A Param. B Param. C

S1/1 S2 Vulnerable groundwater licences Type Param. A x1 Param. B x2 Param. C p1 Param. D p2 S2/1 S3 Time Limited Licences - These risks are now excluded by the EA Guidelines Type Param. A Param. B Param. C Param. D S3/1 S4 Bulk transfers Type Param. A Param. B Param. C Param. D S4/1 S5 Gradual pollution of sources causing a reduction in abstraction Type Param. A Min Param. B Best Param. C Max S5/1 S6 Accuracy of supply side data Type Param. A Param. B Param. C S6/1 Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 2.28 SD 2.28 SD 2.27 SD 2.27 SD 2.27 SD 2.27 SD 2.27 S6/2 Meter uncertainty for licence critical sources 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Type Param. A Param. B Param. C S6/3 Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 3.11 SD 3.11 SD 3.11 SD 3.11 SD 3.12 SD 3.12 SD 3.13 S6/4 Uncertainty for climate and catchment characteristics affecting surface waters 0.00 0.00 0.00 0.00 0.00 0.00 0.00 S8 Uncertainty of impact of climate change on Deployable Output Type Triangular Triangular Triangular Triangular Triangular Triangular Triangular Param. A -0.05 Min -0.20 Min -0.46 Min -0.51 Min -0.56 Min -0.61 Min -0.65 Param. B 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C 0.37 Max 1.48 Max 3.34 Max 3.71 Max 4.08 Max 4.45 Max 4.75 S8/1 Uncertainty of impact of climate change on source yields 0.11 0.43 0.96 1.07 1.18 1.28 1.37 S9 Uncertainty over New Sources Type Param. A Min Min Min Min Min Min Min Param. B Best Best Best Best Best Best Best Param. C Max Max Max Max Max Max Max S9/1 Type Param. A Min Min Min Min Min Min Min Param. B Best Best Best Best Best Best Best Param. C Max Max Max Max Max Max Max S9/1 D1 Accuracy of sub-component data Type Normal Normal Normal Normal Normal Normal Normal Param. A Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Avg 0.00 Param. B SD 1.00 SD 0.95 SD 0.93 SD 0.91 SD 0.89 SD 0.88 SD 0.88 D1/1 Uncertainty of distribution input arising from meter inaccuracy 0.00 0.00 0.00 0.00 0.00 0.00 0.00 D2 Demand forecast variation Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -1.15 Min -3.08 Min -5.00 Min -6.93 Min -8.85 Min -10.39 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 1.15 Max 3.08 Max 5.00 Max 6.93 Max 8.85 Max 10.39 D2/1 Demand forecast variation. 0.00 0.00 0.00 0.00 0.00 0.00 Type Param. A Param. B Param. C

D2/2 D3 Uncertainty of impact of climate change on demand Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -0.01 Min -0.04 Min -0.06 Min -0.07 Min -0.09 Min -0.11 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 0.01 Max 0.04 Max 0.06 Max 0.07 Max 0.09 Max 0.11 D3/1 Effect of climate change on demand 0.00 0.00 0.00 0.00 0.00 0.00 D4 Uncertainty of impact of demand management Type Triangular Triangular Triangular Triangular Triangular Triangular Param. A Min -0.04 Min -0.11 Min -0.18 Min -0.24 Min -0.31 Min -0.37 Param. B Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Best 0.00 Param. C Max 0.04 Max 0.11 Max 0.18 Max 0.24 Max 0.31 Max 0.37 D4/1 Uncertainty of impact of demand management 0.00 0.00 0.00 0.00 0.00 0.00 Overlapping components 14 16 16 16 16 20 22 Group 1 Group 2 Group 3

2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 All uncertainties: 0.11 0.43 0.96 1.07 1.18 1.28 1.37

Page A37

SWW Water Resources Management Plan 2015 – 2040

Table A.24: Wimbleball WRZ headroom uncertainty @RISK output report

Outputs 2012/13 2015/16 2020/21 2025/26 2030/31 2035/36 2039/40 Minimum -15.27 -14.49 -14.58 -15.02 -19.48 -16.35 -18.24 Maximum 15.06 15.17 16.09 18.37 22.77 20.59 23.23 Mean 0.11 0.43 0.96 1.07 1.18 1.28 1.37 Std Deviation 3.98 4.02 4.21 4.58 5.01 5.48 5.89 Variance 15.83 16.15 17.73 20.93 25.05 29.99 34.65 Skewness -0.037 -0.006 -0.017 0.002 0.014 -0.020 0.036 Kurtosis 2.961 2.992 2.923 2.900 2.885 2.812 2.884 Number of Errors 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Mode -0.39 1.32 0.74 1.25 1.62 1.48 0.10 5.0% -6.53 -6.25 -6.02 -6.48 -7.05 -7.79 -8.34 10.0% -5.05 -4.75 -4.50 -4.85 -5.34 -5.86 -6.30 15.0% -4.04 -3.73 -3.40 -3.72 -4.08 -4.42 -4.80 20.0% -3.26 -2.95 -2.62 -2.81 -3.11 -3.35 -3.64 25.0% -2.59 -2.28 -1.89 -2.00 -2.24 -2.47 -2.69 30.0% -1.98 -1.67 -1.27 -1.39 -1.51 -1.60 -1.79 35.0% -1.41 -1.13 -0.68 -0.74 -0.76 -0.87 -0.91 40.0% -0.86 -0.61 -0.12 -0.11 -0.13 -0.14 -0.12 45.0% -0.35 -0.08 0.44 0.52 0.53 0.54 0.58 50.0% 0.14 0.43 1.00 1.09 1.19 1.28 1.31 55.0% 0.66 0.96 1.52 1.65 1.81 1.96 2.05 60.0% 1.17 1.48 2.06 2.27 2.46 2.73 2.88 65.0% 1.70 2.01 2.64 2.89 3.08 3.46 3.68 70.0% 2.22 2.54 3.22 3.50 3.85 4.32 4.49 75.0% 2.80 3.14 3.83 4.17 4.60 5.07 5.40 80.0% 3.41 3.82 4.55 4.94 5.42 5.94 6.38 85.0% 4.23 4.54 5.37 5.86 6.45 7.06 7.50 90.0% 5.19 5.52 6.32 6.91 7.74 8.41 9.00 95.0% 6.66 7.04 7.83 8.64 9.38 10.30 11.03

Blue cells are 85th, reducing to 70th, percentile figures used in the supply demand balance assessment.

Page A38

SWW Water Resources Management Plan 2015 – 2040

Figure A.16 Wimbleball WRZ headroom uncertainty @RISK output chart

Headroom Uncertainty - Wimbleball WRZ (Ml/day) HEADROOM 5% Perc 10% Perc 15% Perc 20% Perc 25% Perc 30% Perc 35% Perc 40% Perc 45% Perc 50% Perc 55% Perc 60% Perc 65% Perc 70% Perc 75% Perc 80% Perc 85% Perc 90% Perc 95% Perc 15.0

10.0

5.0

0.0

2012/13 2015/16 2020/21 2025/26 2030/31 3035/36 3039/40

-5.0

-10.0

Page A39

SWW Water Resources Management Plan 2015 – 2040

Wimbleball WRZ headroom uncertainty @RISK sensitivity analysis - tornado graphs

The tornado graphs below show how sensitive the headroom is to each of the headroom components for the 2012/13, 2015/16, 2025/26 and 2039/40 time horizons.

Figure A.17 Wimbleball WRZ sensitivity tornado graph 2012/13

Sensitivity Tornado S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D1/1 Uncertainty of distribution input arising from meter inaccuracy S8/1 Uncertainty of impact of climate change on source

yields

6 4 8 2

0 2 6 8 4

- - - - Mean of 2012/13

Figure A.18 Wimbleball WRZ sensitivity tornado graph 2015/16

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D1/1 Uncertainty of distribution input arising from meter inaccuracy

D2/1 Demand forecast variation.

S8/1 Uncertainty of impact of climate change on source yields

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

8 4 6 2

0 2 6 8 4

- - - - Mean of 2015/16

Page A40

SWW Water Resources Management Plan 2015 – 2040

Figure A.19 Wimbleball WRZ sensitivity tornado graph 2025/26

Sensitivity Tornado

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

D2/1 Demand forecast variation.

D1/1 Uncertainty of distribution input arising from meter inaccuracy S8/1 Uncertainty of impact of climate change on source yields

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

8 6 4 2

0 2 6 4 8

- - - - 10 Mean of 2025/26

Figure A.20 Wimbleball WRZ sensitivity tornado graph 2039/40

Sensitivity Tornado

D2/1 Demand forecast variation.

S6/4 Uncertainty for climate and catchment characteristics affecting surface waters

S6/2 Meter uncertainty for licence critical sources

S8/1 Uncertainty of impact of climate change on source yields D1/1 Uncertainty of distribution input arising from meter inaccuracy

D4/1 Uncertainty of impact of demand management

D3/1 Effect of climate change on demand

8 6 2 4

2 6 8 0 4

- - - -

10 12 Mean of 2039/40

Page A41

SWW Water Resources Management Plan 2015 – 2040

Wimbleball WRZ headroom uncertainty @RISK sensitivity analysis – climate change components

In accordance with Environment Agency proposals13, headroom was calculated for the five scenarios:

. supply uncertainty excluding climate change (S1 – S7, S9) . demand uncertainty excluding climate change (D1, D2, D4) . impacts of climate change on supply (S8) . impacts of climate change on demand (D3) . all components (S1 – S9 inclusive, D1 – D4 inclusive)

The results are shown in the chart below for the 75th percentile. As can be seen from the chart, the total headroom is not simply a sum of the headroom calculated from groups of components. However, the chart does give an indication of the relative contribution each group of components makes to the overall headroom calculation.

As would be expected from inspection of the values of the D3 component of headroom (effect of climate change on demand), this makes a very small contribution to target headroom. The S8 component (the impacts of climate change on supply) makes an increasing contribution to the target headroom as we move through the planning period. However, by the end of the planning period the contribution to headroom from this component is still no more than the non-climate change impacts on supply.

Figure A.21 Wimbleball WRZ – graph showing climate change contribution to headroom uncertainty

Headroom Uncertainty - Wimbleball WRZ (Ml/day) Climate change contribution to headroom (75th percentile) 9

Impacts of climate change on 8 demand (D3):

7 Demand uncertainty excluding climate change 6 (D1, D2, D4):

5 Impacts of climate change on supply (S8): 4

3 Supply uncertainty excluding

climate change (S1 - S7, S9): Headroom Headroom uncertainty (Ml/day) 2

1 All components

0

2012/13 2025/26 2039/40 2020/21 2030/31 2035/36 2015/16

13 Ibid. 7, page 63

Page A42

SWW Water Resources Management Plan 2015 – 2040

Appendix B Demand profiles

SWW Water Resources Management Plan 2015 – 2040

Table B.1: Colliford Water Resources Zone (WRZ) demand profiles

Week WIS zones Number 101, 103-107, 102, 108, 202- 205, 207 201, 206, 407-409 204, 208 1 0.942 0.893 0.843 2 0.944 0.883 0.789 3 0.937 0.875 0.765 4 0.930 0.894 0.756 5 0.929 0.903 0.792 6 0.929 0.867 0.772 7 0.937 0.843 0.784 8 0.929 0.843 0.773 9 0.926 0.846 0.791 10 0.925 0.843 0.774 11 0.913 0.867 0.787 12 0.911 0.878 0.786 13 0.915 0.884 0.868 14 0.939 0.933 0.895 15 0.989 0.941 0.945 16 1.011 1.017 0.991 17 1.004 1.006 1.007 18 1.044 0.987 0.954 19 1.020 1.019 1.075 20 1.046 0.989 0.951 21 1.098 1.059 1.063 22 1.086 1.141 1.145 23 1.075 1.096 1.124 24 1.117 1.074 1.229 25 1.134 1.160 1.321 26 1.225 1.173 1.331 27 1.186 1.084 1.314 28 1.146 1.208 1.377 29 1.151 1.185 1.313 30 1.228 1.328 1.481 31 1.153 1.446 1.450 32 1.153 1.336 1.536 33 1.152 1.289 1.499 34 1.149 1.254 1.441 35 1.094 1.279 1.328 36 1.057 1.122 1.162 37 1.037 1.085 1.155 38 1.007 1.018 0.954 39 0.952 0.933 0.875 40 0.912 0.910 0.874 41 0.902 0.919 0.903 42 0.893 0.924 0.884 43 0.903 0.923 0.894 44 0.896 0.925 0.838 45 0.896 0.869 0.789 46 0.890 0.875 0.802 47 0.892 0.867 0.814 48 0.893 0.865 0.808 49 0.893 0.871 0.808 50 0.894 0.848 0.816 51 0.905 0.857 0.779 52 0.910 0.866 0.796

Page B1

SWW Water Resources Management Plan 2015 – 2040

Table B.2: Water Into Supply (WIS) Zones in Colliford WRZ

WIS Zone WIS Zone Name Ref 101 Penzance 102 Hayle 103 Lizard 104 Redruth 105 Falmouth 106 Truro 107 Camborne 108 Probus 201 St Austell 202 Fowey 203 Looe 204 Camelford 205 St Minver 206 Bodmin 207 St Columb Major 208 Newquay 407 Launceston 408 Torpoint 409 Saltash

Page B2

SWW Water Resources Management Plan 2015 – 2040

Table B.3: Roadford WRZ demand profiles

Week WIS zones Number 301-310, 312, 405- 502-511, 513, 406, 410, 501, 514 401-404 515 1 0.937 0.948 0.927 2 0.932 0.968 0.921 3 0.930 0.973 0.902 4 0.919 0.969 0.919 5 0.927 0.966 0.909 6 0.918 0.966 0.908 7 0.910 0.957 0.922 8 0.906 0.957 0.933 9 0.912 0.957 0.930 10 0.904 0.964 0.939 11 0.900 0.956 0.939 12 0.901 0.944 0.946 13 0.898 0.931 0.926 14 0.949 0.956 0.980 15 0.984 0.978 1.027 16 1.021 0.996 0.992 17 0.989 1.009 0.956 18 1.035 1.041 1.034 19 1.008 1.026 0.996 20 0.993 1.080 0.970 21 1.045 1.084 1.008 22 1.053 1.056 1.006 23 1.082 1.069 1.038 24 1.109 1.071 1.073 25 1.134 1.074 1.143 26 1.225 1.035 1.155 27 1.143 1.062 1.169 28 1.152 1.109 1.264 29 1.203 1.093 1.263 30 1.234 1.112 1.219 31 1.201 1.059 1.187 32 1.210 1.110 1.169 33 1.212 1.040 1.105 34 1.129 1.034 1.046 35 1.105 1.061 1.069 36 1.059 1.079 0.967 37 1.053 1.068 0.982 38 1.019 1.031 0.962 39 0.962 0.979 0.955 40 0.913 0.952 0.947 41 0.902 0.951 0.941 42 0.904 0.945 0.963 43 0.910 0.937 0.931 44 0.913 0.934 0.922 45 0.896 0.942 0.943 46 0.900 0.941 0.910 47 0.908 0.941 0.952 48 0.910 0.940 0.922 49 0.901 0.947 0.927 50 0.907 0.950 0.945 51 0.911 0.936 0.961 52 0.925 0.914 0.983

Page B3

SWW Water Resources Management Plan 2015 – 2040

Table B.4: WIS Zones in Roadford WRZ

WIS Zone WIS Zone Name Ref 301 Lynton 302 Parracombe 303 Combe Martin 304 Ilfracombe 305 Braunton 306 Barnstaple 307 Bideford 308 Clovelly 309 Okehampton 310 Winkleigh 312 South Molton 401 Plymouth 402 Yealmpton 403 Tavistock 404 Princetown 405 Broadwoodwidger 406 Bude 410 Brentor 501 Chagford 502 Moretonhampstead 503 Ashburton 504 Buckfastleigh 505 Kingsbridge 506 Brixham 507 Paignton 508 Torquay 509 Newton Abbot 510 Teignmouth 511 Dawlish 513 Chudleigh 514 Tedburn St Mary 515 Kingskerswell

Page B4

SWW Water Resources Management Plan 2015 – 2040

Table B.5: Wimbleball WRZ demand profile

Week WIS Zones 311, Number 512, 601-612 1 0.928 2 0.943 3 0.942 4 0.937 5 0.939 6 0.937 7 0.934 8 0.929 9 0.937 10 0.938 11 0.930 12 0.929 13 0.922 14 0.959 15 1.004 16 1.047 17 1.040 18 1.067 19 1.020 20 1.057 21 1.089 22 1.068 23 1.079 24 1.156 25 1.137 26 1.096 27 1.078 28 1.139 29 1.156 30 1.222 31 1.085 32 1.139 33 1.138 34 1.087 35 1.083 36 1.067 37 0.998 38 0.986 39 0.953 40 0.920 41 0.921 42 0.914 43 0.927 44 0.923 45 0.916 46 0.921 47 0.922 48 0.919 49 0.913 50 0.910 51 0.912 52 0.890

Page B5

SWW Water Resources Management Plan 2015 – 2040

Table B.6: WIS Zones in Wimbleball WRZ

WIS Zone WIS Zone Name Ref 311 Washford Pyne 512 Exminster 601 Crediton 602 Broadclyst 603 Exeter 604 Exmouth 605 Axminster 606 Chardstock 607 Stockland 608 Honiton 609 Ottery St Mary 610 Willand 611 Tiverton 6012 Woodbury

Page B6

SWW Water Resources Management Plan 2015 – 2040

Appendix C Outage determination

SWW Water Resources Management Plan 2015 – 2040

1 Introduction

The methodology adopted for the calculation of the outage allowance is similar to that we used in our 2009 Water Resources Plan (WRP09)1, and therefore is in line with best practice2. The key terms used in this Appendix are defined in the Glossary.

This Appendix describes how we have calculated outage allowances for each of the Water Resource Zones (WRZs), and how we defined and used Water Treatment Works (WTW) outputs to facilitate this. In addition, for our Wimbleball WRZ, we have carried out a Monte Carlo analysis for certain types of outage events specific to WTWs fed largely or solely by groundwater sources. This is explained in detail in Section 3.

2 Approach to identifying the outage allowance

In line with the UKWIR Outage Methodology3, we have followed a stepped approach to establishing the allowable outage for each WRZ:

1) Identify periods of low output at all WTWs which could be outage events.

2) For the periods identified, consider if the reduction in output was due to a planned event or unplanned event and confirm that planned events did not affect Deployable Output (DO).

3) Where unplanned, consider whether an event would have potentially affected DO and was therefore a legitimate outage.

4) Determine the outage allowance from the legitimate outage events identified.

5) Use Monte Carlo analysis to determine potential outage related to specific outage events identified at groundwater-fed WTWs in the Wimbleball WRZ.

6) Combine the results of steps 4 and 5 to determine the outage allowance for each WRZ.

2.1 Step 1: Identifying potential outages

As for our WRP094, for this Plan the initial step of outage event identification was through a screening of works outputs, in this case output data from the years 2009, 2010 and 2011 were considered. The output data were scrutinised through the analysis of the 30 day moving average value. This produced a typical output for that time of the year with which to compare actual outputs. This approach identified significant and/or sharp drops in output which had the potential to be genuine outages rather than drops in output associated with reduced demand or operational strategy.

1 South West Water, “Water Resources Plan 2010-35”, November 2009 2 UKWIR, “Outage Allowances for Water Resources Planning”, 1995 3 Ibid. 2 4 Ibid. 1

Page C1

SWW Water Resources Management Plan 2015 – 2040

Unusually low outputs were defined as outputs that were 30% below the 30 day moving average, ie equivalent to approximately a 7 hour total loss of output in a day. This approach takes into account daily variations or seasonal output fluctuations in output which are less than 30% and unlikely to be legitimate outages that reduce DO5.

A further check was carried out by comparing daily output to the annual average figure in order to highlight significant drops in output which may not have been revealed from the 30 day moving average analysis.

2.2 Step 2: Identification of planned events

In order to carry out specific capital maintenance projects or routine maintenance, we do not usually require works outputs to be significantly reduced or WTWs to be shut down. However, when this is unavoidable, short complete or partial shutdowns for maintenance are typically planned for when strategic sources are not being deployed eg during the winter period.

An example of this is would be the shutdown of a WTW whose supply zone is largely unsupported by other WTWs. This will result in the loss of treated water storage and an increased output following the shutdown. However, if as a result of the event, there is no overall impact on storage of the strategic reservoir this will have no effect on the total Water Available For Use (WAFU).

A further example is the shutdown of a river sourceworks in a zone which can also be supplied by a WTW fed by a reservoir, but at a non-critical time. Although the shutdown of the river sourceworks results in the increased use of water from a reservoir, if the reservoir is spilling, there will be no impact on the storage of either this reservoir or the strategic reservoir. Hence there is no reduction in the WAFU to the system.

Hence in our analysis, planned outage events are taken to have no significant impact on WAFU and therefore for the purposes of this assessment no allowance has been made for planned outages. Step 2 of the methodology therefore involves identification of any reduction in works output due to a planned event which would not be regarded as a legitimate outage. This is carried out through reference to Water Resources Review Group Meeting minutes and through discussions with regional Operations Abstraction Managers.

5 It is inherent in this assumption that the distribution system can tolerate a 7 hour shutdown of any WTW. This is considered to be a reasonable assumption.

Page C2

SWW Water Resources Management Plan 2015 – 2040

2.3 Step 3: Identification of legitimate outage

Once identified planned outages have been taken out of the outage assessment, Step 3 of the analysis is to consider only those unplanned outage events which are genuine outages that impact on WAFU. The outage events for each WRZ over for the years 2009, 2010 and 2011 are listed in Table C.1 to C.3 below with those which affect WAFU identified. Unplanned outages which affect WAFU are taken to be unexpected reductions in output capacity that occurred at times where strategic reservoirs were releasing to enhance river flows. As can be seen from the tables, the majority of unplanned outages occurred as a result of local operational issues which include:

. Inability of a WTW to cope with the range of raw water quality it receives.

. Insufficient provision of standby equipment to cover for breakdowns and maintenance.

. Power and system failures.

Outage of groundwater sources due specifically to borehole pump failure, flooding consequences at Dotton WTW and turbidity problems at Ottery St Mary have been included in a separate analysis using a Monte Carlo simulation.

It should be noted that some unplanned outages are short (i.e. less than 12 hours) which gives little time to increase the output of other WTWs or to re-valve the distribution system. The amount of stored treated water in the system is often, therefore, the deciding factor in whether customers run out of water. This is obviously a very important issue, but separate from the effect of outage on the WAFU of a WRZ. Therefore, in this Appendix, we do not consider treated water storage requirements in relation to outages.

Page C3

SWW Water Resources Management Plan 2015 – 2040

Table C.1: Outages impacting on WAFU in Colliford WRZ

WTW Start Duration Reason Potential Equivalent Total date (days) impact on reduction equivalent WAFU? in storage reduction (Ml) in storage (Ml) Drift 22/08/09 1 Local issues No 0.0 0.0 Stithians 16/12/09 1 “ No 0.0 0.0

2009 Wendron 19/10/09 1 “ No 0.0 0.0 Total 0.0 Drift 13/10/10 1 Local issues No 0.0 0.0 21/08/10 1 “ Yes 10.0 De Lank 20.0

2010 23/08/10 1 “ Yes 10.0 Total 20.0

16/02/11 1 Local issues Yes 10.0 De Lank 20.0

07/03/11 1 “ Yes 10.0 2011 Total 20.0

Page C4

SWW Water Resources Management Plan 2015 – 2040

Table C.2: Outages impacting on WAFU in Roadford WRZ

WTW Start Duration Reason Potential Equivalent Total date (days) impact on reduction equivalent WAFU? in storage reduction (Ml) in storage (Ml) Avon 25/03/09 5 Local issues No 0.0 0.0 Bratton 12/05/09 2 “ No 0.0 0.0 Fleming 20/11/09 4 “ No 0.0 11/03/09 1 “ No 0.0 Horedown 0.0 07/11/09 2 “ No 0.0 Littlehempston 13/05/09 1 “ No 0.0 0.0 14/07/09 1 “ No 0.0 Northcombe 17/07/09 2 “ No 0.0 0.0

20/07/09 1 “ No 0.0

2009 27/07/09 1 “ No 0.0 Prewley 0.0 04/09/09 1 “ No 0.0 Service 01/07/09 2 reservoir Yes 16.0 issues Tamar Lakes 04/07/09 1 “ Yes 8.0 32.0 06/07/09 1 “ Yes 8.0 08/09/09 1 Local issues No 0.0 Tottiford 28/05/09 1 “ Yes 29.0 29.0 Total 61.0 28/04/10 1 Local issues No 0.0 Bratton 09/06/10 2 “ Yes 20.0 20.0 Fleming 18/08/10 1 “ No 0.0 08/03/10 1 “ No 0.0 Dousland 0.0 04/10/10 1 “ No 0.0 02/02/10 3 “ No 0.0

04/03/10 2 “ No 0.0 30/03/10 2 “ No 0.0

2010 Horedown 0.0 24/07/10 5 “ No 0.0 23/10/10 2 “ No 0.0 08/12/10 2 “ No 0.0 24/11/10 1 “ No 0.0 Northcombe 0.0 27/11/10 1 “ No 0.0 Tamar Lakes 12/05/10 1 “ No 0.0 0.0 Total 20.0

Page C5

SWW Water Resources Management Plan 2015 – 2040

WTW Start Duration Reason Potential Equivalent Total date (days) impact on reduction equivalent WAFU? in storage reduction (Ml) in storage (Ml) Avon 25/01/11 1 Local issues No 0.0 0.0 11/04/11 1 “ No 0.0 Bratton 31/08/11 1 “ No 0.0 0.0 Fleming 10/11/11 3 “ No 0.0 22/12/11 1 “ No 0.0 24/05/11 1 “ No 0.0 Pumps 01/06/11 2 problems at Yes 8.4 Hore Down reservoir 8.4 15/08/11 1 Local issues No 0.0

30/08/11 1 “ No 0.0

Northcombe 21/12/11 1 “ No 0.0 0.0

2011 23/04/11 1 “ No 0.0 18/11/11 1 “ No 0.0 20/11/11 1 “ No 0.0 Prewley 08/12/11 1 “ No 0.0 0.0 10/12/11 1 “ No 0.0 13/12/11 1 “ No 0.0 28/12/11 2 “ No 0.0 Tamar Lakes 06/08/11 1 “ No 0.0 0.0 Tottiford 18/08/11 1 “ No 0.0 0.0 Venford 12/08/11 1 “ No 0.0 0.0 Total 8.4

Page C6

SWW Water Resources Management Plan 2015 – 2040

Table C.3: Outages impacting on WAFU in Wimbleball WRZ

WTW Start Duration Reason Potential Equivalent Total date (days) impact on reduction equivalent WAFU? in storage reduction (Ml) in storage (Ml) Bovey 26/05/09 2 Local issues Yes 1.0 1.0 Lane 13/10/09 1 “ No 0.0 04/09/09 1 “ No 0.0 Kersbrook 0.0 01/12/09 2 “ No 0.0 27/05/09 1 “ No 0.0 30/05/09 2 “ No 0.0 10/07/09 3 “ Yes 18.0

Ottery 24.0 08/10/09 1 “ Yes 6.0

2009 24/11/09 1 “ No 0.0 14/12/09 1 “ No 0.0 15/02/09 1 “ No 0.0 18/04/09 2 “ No 0.0 Wilmington 23/07/09 1 “ No 0.0 2.5 25/07/09 2 “ No 0.0 22/09/09 1 “ Yes 2.5 Total 27.5 Bovey 06/08/10 14 Local issues Yes 7.0 7.0 Lane Dotton 17/11/10 2 “ No 0.0 0.0 Kersbrook 02/03/10 2 “ No 0.0 0.0 Ottery 18/04/10 1 “ No 0.0 0.0

09/01/10 1 “ No 0.0

01/03/10 1 “ No 0.0 2010 29/05/10 1 “ Yes 2.5 Wilmington 07/06/10 1 “ Yes 2.5 12.5 16/06/10 1 “ Yes 2.5 17/08/10 1 “ Yes 2.5 25/09/10 1 “ Yes 2.5 Total 19.5

Page C7

SWW Water Resources Management Plan 2015 – 2040

WTW Start Duration Reason Potential Equivalent Total date (days) impact on reduction equivalent WAFU? in storage reduction (Ml) in storage (Ml) Included in borehole 07/01/11 33 Pump failure 0.0 failure analysis 15/02/11 1 Local issues No 0.0 28/03/11 1 “ No 0.0 18/04/11 1 “ No 0.0 Bovey 5.0 Lane 24/04/11 1 “ Yes 0.5 29/05/11 2 “ Yes 1.0 04/06/11 2 “ Yes 1.0 07/07/11 1 “ Yes 0.5

29/07/11 4 “ Yes 2.0

23/11/11 1 “ No 0.0

2011 08/12/11 1 “ No 0.0 24/09/11 3 “ No 0.0 Hook 0.0 28/09/11 1 “ No 0.0 26/07/11 1 “ Yes 1.4 Kersbrook 19/08/11 3 “ Yes 4.2 7.0 03/09/11 3 “ Yes 1.4 06/12/11 1 “ No 0.0 Ottery 09/12/11 2 “ No 0.0 0.0 19/12/11 2 “ No 0.0 12/12/11 1 “ No 0.0 Wilmington 0.0 31/12/11 1 “ No 0.0 Total 12.0

Page C8

SWW Water Resources Management Plan 2015 – 2040

2.3.1 Autumnal leaf fall affecting river water quality

As part of the outage assessment, consideration has been given to the issue of short term river water deterioration as a consequence of autumnal leaf falls. Such events have been noted most significantly at Restormel WTW (River Fowey) and to a lesser extent at Littlehempston WTW (River Dart). In the Littlehempston WTW case, leaf laden river water prevents use of the river intake until the problem has cleared whilst at Restormel WTW it is not possible to treat as much water as normal and abstractions are reduced.

Whilst events at Restormel can take several weeks before abstractions return to normal, at Littlehempston it is typical to experience only a few days (or even 24 hours) of non-availability of river water. In abstraction capability terms this can result in up to 27 Ml/d lost at Littlehempston WTW. Given this WTW has access to alternative resources in the form of groundwater supplies and reservoir storage a loss of a few days river abstraction does not result in an impact on DO and hence it has not been considered further in the outage determination.

At Restormel, the loss of output capacity is more significant and the ability of the treatment system to treat the water temporarily reduces the maximum output to approximately 60 Ml/d. Examples of this occurred in December 2009 and again in November 2010 when output was affected for around a month in each case. Consideration has been given to making allowance for this in the outage assessment as the potential exists for such events to occur at more critical times than experienced recently which could affect WAFU. However, on balance it was decided not to do so as an AMP5 project is in place to provide a solution to this issue (SWW Project Reference 29021815). This scheme is due to be completed before 2015 and therefore river water quality affected in this way should not pose a risk to Restormel output capacity in AMP6 and beyond.

2.4 Step 4: Outage allowance determination

The removal of planned outages and unplanned outages which do not impact on WAFU results in only legitimate outages remaining. These are then analysed to calculate an appropriate level of outage. From the data presented above it is clear that for the Colliford and Roadford WRZs the volumes of water lost due to legitimate outage events is very small. Consequently a de minimis value of 1 Ml/d has been adopted for both these supply areas. For the Wimbleball WRZ, the impact from groundwater fed outages is slightly more significant and an appropriate value generated by Monte Carlo simulation (carried out in Step 5) has been added to produce a total outage figure.

2.5 Step 5: Analysis of outage at groundwater supplied WTWs

In addition to assessing the outage events at WTWs in general, a further allowance has been made specifically for infrequent but recurring outages which have been identified as a specific risk to WTWs fed by groundwater sources in the Wimbleball WRZ. This is the only WRZ which has significant groundwater resources. These specific outage events have been assessed in combination through a Monte Carlo analysis which is explained in detail in Section 3.

Page C9

SWW Water Resources Management Plan 2015 – 2040

Unlike at WTWs served by river or reservoir water where partial loss of output on any one particular day can be offset by higher outputs on subsequent days, this is not the case at the majority of groundwater sources where borehole yield or licensed abstraction rates constrain the sources each day. Operating at higher rates to accommodate an earlier outage is generally not an option at most boreholes. This is why the Wimbleball WRZ is more sensitive to outage events which can be demonstrated through analysis.

2.6 Step 6: Total outage allowance

The legitimate outages calculated from the analysis of all works output data have then been combined with the specific potential outages identified for groundwater- fed WTWs estimated by Monte Carlo simulation. Together this value is taken to be the total outage allowance.

3.0 Monte Carlo simulation of three key outage events identified in the Wimbleball WRZ

In our WRP096 of outage allowance was made for the consequences of a temporary loss of supply in the Wimbleball WRZ as a result of borehole failure. This approach was thought appropriate based on examination of the historical incidence of borehole failure and the condition of South West Water boreholes as assessed at that time.

During AMP5 South West Water we have undertaken a pro-active stance to borehole condition monitoring and rehabilitation which has produced greater confidence in the reliability of our groundwater assets. As a consequence the risk of borehole failure is no longer considered appropriate for inclusion in this outage assessment.

However, over the last five years three types of outage, specific to borehole sources, have become apparent which it is felt should be taken into account in the outage determination.

The specific outage events identified are:

1. Outage of any borehole source due to submersible pump failure

2. Outage of Dotton boreholes No.1 and No.3 associated with river flooding

3. Outage due to unacceptable raw water turbidity of Greatwell boreholes No.1 No.2 and/or No.3 supplying Ottery St Mary (Intermediate WTW) linked to power or system failure.

6 Ibid. 1

Page C10

SWW Water Resources Management Plan 2015 – 2040

3.1 Borehole outage resulting from submersible pump failure

In 2006 a re-assessment of South West Water borehole pumps revealed that many were in need of replacement on the grounds that they were either old and consequently had excessive power consumption, were of an inappropriate specification, or sometimes both.

Since 2006, South West Water has embarked on a programme of submersible borehole pump replacement and by 2012 only seven units remained unchanged out of 34 borehole pumps in operational boreholes. However, an extended run of failures amongst these newly replaced pumps between 2009 and 2012 resulted in a potential for a loss of DO within the Wimbleball WRZ dependent upon the significance of the borehole in water supply terms and also the timing of the event.

This high failure rate of borehole pumps was unprecedented. A South West Water investigation in collaboration with its suppliers identified an underlying cause of failure and remedial action was taken to prevent further occurrences. Although some progress has been made in identifying a likely cause for one mode of pump failure, which may have been resolved, the Company wishes to maintain a conservative approach to the potential rate of borehole pump failures over the next few years. As a consequence this outage event has been included in a Monte Carlo simulation of the outage risks that can specifically affect the Wimbleball WRZ.

3.2 Outage of Dotton No.1 and No.3 associated with flooding of the River Otter

This particular outage event has been experienced and managed for many years but only recently has it become a significant risk to the output capacity of Dotton WTW. When the River Otter goes into spate following heavy rainfall events it can result in inundation of the floodplain. At Dotton WTW, the consequence of this is for boreholes No.1 and No.3 to experience increased turbidity levels which are considered a contamination risk.

A protection system has been put in place whereby on critical water levels in the river being reached Dotton boreholes No.1 and No.3 are automatically switched off as a precautionary measure. The boreholes are only returned to service once the river level has receded and the water quality of the boreholes has returned to normal. This outage event results in a loss of raw water supply of 6 Ml/d. This is significant in the context of a normal output from Dotton of around 20 Ml/d.

Analysis of the frequency of this type of outage event has shown that the number of high intensity rainfall events leading to floods has been above average in recent years highlighting the risk posed by such an event. In particular, 2012 saw an unprecedented six River Otter flooding events sufficient to cause the temporary loss of the Dotton boreholes. The time during which these boreholes remained out of service varied from event to event but in one case Dotton borehole No.1 was unable to be re-introduced to supply for 25 days.

In view of this, this particular event has been included within the Monte Carlo simulation of outage events for the Wimbleball WRZ.

Page C11

SWW Water Resources Management Plan 2015 – 2040

3.3 Outage due to unacceptable raw water turbidity of Greatwell boreholes No. 1 No. 2 and / or No. 3 supplying Ottery St Mary WTW

All WTWs have the potential to experience system failure and power outage from time to time. Unlike at other WTWs, where such events are usually limited to only a few minutes (through start-up of generator or by the fault being rectified remotely) or a few hours (re-started shortly after site attendance), at Ottery St Mary WTW the outage is potentially more significant. Water quality issues associated with Greatwell boreholes No. 1, No. 2 and No. 3 can result in an important loss of source DO whilst the water quality improves sufficiently for the boreholes to return to service.

The problem, turbidity, arises from a combination of ageing boreholes, poor borehole design and the presence of iron in the groundwater causing bio-fouling. Should these boreholes switch-off, even for a period of a few minutes, the restarting produces unacceptable levels of turbidity associated with sand particles and iron deposits. Depending on the borehole condition (time since last cleaning) and length of switch-off, the time required to bring turbidity levels down to acceptable levels can be considerable (as much as 24 hours should the event occur out of hours).

Although this outage event only occurs approximately twice a year on average, the operational problems it generates are considered sufficient to warrant its inclusion in the outage analysis. It has therefore been incorporated in the Monte Carlo simulation of outage for the Wimbleball WRZ.

3.4 Monte Carlo simulation

As described above, in order to make a thorough and conservative assessment of the outage risk in the Wimbleball WRZ, a Monte Carlo simulation approach has been adopted to cover three specific outage events. For each of these outage types, a duration and magnitude table was developed in line with the UKWIR outage methodology7 (Tables C.4 to C.6).

7 Ibid. 2

Page C12

SWW Water Resources Management Plan 2015 – 2040

Table C.4: Borehole pump failure data assessment for Monte Carlo simulation

Legitimate Month Duration (days) Magnitude (Ml/d) outage Least Most Maximum Least Most Maximum credible likely credible credible likely credible Jan 3 14 30 0.5 2 4 Feb 3 14 30 0.5 2 4 Mar 3 14 30 0.5 2 4 Apr 3 14 30 0.5 2 4 May 3 14 30 0.5 2 4 Pump Jun 3 14 30 0.5 2 4 failure Jul 3 14 30 0.5 2 4 Aug 3 14 30 0.5 2 4 Sep 3 14 30 0.5 2 4 Oct 3 14 30 0.5 2 4 Nov 3 14 30 0.5 2 4 Dec 3 14 30 0.5 2 4

A total of 21 operational boreholes in the Wimbleball WRZ have been considered in this analysis. The smallest pump is found at Bovey Lane WTW (0.5 Ml/d), whilst the large boreholes are Otterton 1A and Dotton No.3 (4.0 Ml/d). The majority of boreholes yield around 1.5 to 2.0 Ml/d. It is assessed that the most optimistic minimum replacement time is 3 days if a spare pump is available and water quality is acceptable with a worst case scenario of up to one month for a new pump to be installed. Based on pump failures experienced over the last three years a two week period is typically required to locate a pump, install it, run the borehole to waste for sufficient time, collect a sample and wait for sample clearance. Of the boreholes in the Wimbleball WRZ where replacement pumps had been installed since 2006, half have experienced a failure of the replacement pump in the last three years.

Page C13

SWW Water Resources Management Plan 2015 – 2040

Table C.5: Impact of River Otter flooding on Dotton boreholes No.1 & No.3 availability

Legitimate Month Duration (days) Magnitude (Ml/d) outage Least Most Maximum Least Most Maximum credible likely credible credible likely credible Jan 3 7 25 6 6 6 Feb 3 7 25 6 6 6 Mar 3 7 25 6 6 6 Apr 3 7 25 6 6 6 May 3 7 25 6 6 6 River Otter Jun 3 7 25 6 6 6 in spate Jul 3 7 25 6 6 6 Aug 3 7 25 6 6 6 Sep 3 7 25 6 6 6 Oct 3 7 25 6 6 6 Nov 3 7 25 6 6 6 Dec 3 7 25 6 6 6

The river Otter at Dotton WTW has breached its banks on an estimated 80 occasions in the last 49 years (1963-2012). It is assessed that the duration of the output loss due to potential contamination risk varies between three days minimum, seven days likely and 25 days maximum.

Table C.6: Impact of power and system failure on borehole availability at Ottery St Mary (Intermediate) WTW

Legitimate Month Duration (days) Magnitude (Ml/d) outage Least Most Maximum Least Most Maximum credible likely credible credible likely credible Jan 0.25 0.5 1 1.5 3 4.5 Feb 0.25 0.5 1 1.5 3 4.5 Mar 0.25 0.5 1 1.5 3 4.5 Apr 0.25 0.5 1 1.5 3 4.5 May 0.25 0.5 1 1.5 3 4.5 Jun 0.25 0.5 1 1.5 3 4.5 Turbidity Jul 0.25 0.5 1 1.5 3 4.5 Aug 0.25 0.5 1 1.5 3 4.5 Sep 0.25 0.5 1 1.5 3 4.5 Oct 0.25 0.5 1 1.5 3 4.5 Nov 0.25 0.5 1 1.5 3 4.5 Dec 0.25 0.5 1 1.5 3 4.5

Greatwell boreholes No.1, No.2 & No.3 at Ottery St Mary WTW each have a potential source DO of 1.5 Ml/d. The potential magnitude of an event is dependent upon how long it can take to get any of these boreholes back into supply following an unplanned switch-off associated with a power or system outage.

Page C14

SWW Water Resources Management Plan 2015 – 2040

The least credible impact is based on a single borehole suffering raised turbidity for a significant period; the most likely magnitude is taken to be the temporary loss of two boreholes, whilst the maximum credible is assumed to be where all three suffer prolonged turbidity.

For each event type the magnitude and duration distributions were combined to give a distribution for the event over a typical month of 30 days. These were analysed in @RISK software to generate a distribution for the combined effect of all three event types together. The results of this analysis are presented below (Figure C.1).

Taking the 50% probability, the outage from all three events is estimated to be just under 120 Ml (highlighted in yellow in Figure C.2) over the 30 day period. This equates to an average outage of 4 Ml/d.

Figure C.1: Input file for @RISK Monte Carlo simulation of three specific outage events

Headroom Spreadsheet

Company Name South West Water Scenario Ref DraftWRP Version 1 Resource Zone Wimbleball SSA Date 07/02/2013

Component Correlated OUTAGE Distributions determined from Duration and Magnitude Distributions (30 days)

With By Params Wimbleball Outage Monte Carlo Simulation

Dependent Dependent

intermittent Component

Component Overlapping Continuous / Continuous

PF Pump Failure in Ml Type Triangular Param. A Min 1.50 Min Param. B Best 28.00 Best Param. C Max 120.00 Max 49.83 DF Dotton Flooding in Ml Type Triangular Param. A Min 18.00 Min Param. B Best 42.00 Best Param. C Max 150.00 Max 70.00 OT Ottery Turbidity Ml Type Triangular Param. A 0.38 Min Param. B 1.50 Best Param. C 4.50 Max 2.13 Overlapping components 14 16 16 16 16 20 22

Mean 121.96

Page C15

SWW Water Resources Management Plan 2015 – 2040

Figure C.2: Output data from @RISK simulation of three specific outage events

Simulation Summary Information Workbook Name Monthly outage in Ml assessment Dataset 8-2-13.xls Number of Simulations 1 Number of Iterations 5000 Number of Inputs 0 Number of Outputs 1 Sampling Type Latin Hypercube Simulation Start Time 2/8/13 10:04:40 Simulation Duration 00:00:16 Random # Generator Mersenne Tw ister Random Seed 2005860588

Summary Statistics for Cell Q119 Statistics Percentile Minimum 28.99 5% 64.47 Maximum 249.26 10% 73.41 Mean 121.96 15% 80.76 Std Dev 38.30 20% 87.20 Variance 1466.8104 25% 93.37 Skewness 0.325587 30% 98.61 Kurtosis 2.67815 35% 103.59 Median 119.92 40% 109.12 Mode 98.95 45% 114.94 Left X 64.47 50% 119.92 Left P 5% 55% 125.09 Right X 190.26 60% 129.80 Right P 95% 65% 135.51 Diff X 125.79 70% 141.04 Diff P 90% 75% 147.84 #Errors 0 80% 154.53 Filter Min Off 85% 163.20 Filter Max Off 90% 173.83 #Filtered 0 95% 190.26

3.5 Total outage allowance for each WRZ

As shown in the Tables C.1 to C.3 above, for each WRZ the calculated outages are very small. Therefore, as for WRP098, de minimus values of 1 Ml/d have been assumed for general unplanned outages.

For the Wimbleball WRZ, the additional outage as estimated from the Monte Carlo simulation of three specific outage events associated with groundwater sites has been included. The cumulative impact on WAFU of the identified outages is shown below (Figure C.7).

Table C.7: Assumed outage for each WRZ

WRZ Assumed outage (Ml/d) Colliford 1.0 Roadford 1.0 Wimbleball 5.0 Total 7.0

8 Ibid. 1

Page C16

SWW Water Resources Management Plan 2015 – 2040

Appendix D Micro-component per capita consumption forecasts

SWW Water Resources Management Plan 2015 – 2040

Micro-component Per Capita Consumption Forecasting South West Water SSW0501

rpsgroup.com

Quality Management

Prepared by: Richard Critchley

Authorised by: Ben Sansom

th Date: 28 August 2012 Revision: 3.0 Final Project Number: SWW0501 Document Reference:

Prepared by: Prepared for: RPS Water and Environmental Consultancy South West Water, Unit 2, Cobbe Barns Peninsula House, Beddingham, Lewes Rydon Lane, East Sussex, BN8 6JU Exeter, Tel. 01273 858223 Devon. Fax 01273 858229 EX2 7HR

Our ref:

Date:

Disclaimer:

This document has been prepared by RPS Group Limited (RPS) with all reasonable skill, care and diligence and taking into account the information made available by the Client. No other warranty, expressed or implied, is made as to the professional advice included in this document or any other services provided by us. RPS disclaims any responsibility to the Client and others in respect of any matters outside the scope of this contract.

All work carried out in preparing this document has utilised and is based upon RPS’s current professional knowledge and understanding of current relevant UK standards and codes, technology and legislation. Changes in this legislation and guidance may occur at any time in the future and cause any conclusions to become inappropriate or incorrect. RPS does not accept responsibility for advising the client of the facts or implications of any such changes.

This document is confidential to the Client and is not to be disclosed to third parties. If disclosed to third parties, RPS accepts no responsibility of whatsoever nature to third parties to whom this document, or any part thereof, is made known. Any third party relies upon the contents of this document at their own risk and is not to be relied upon by any party, other than the Client without the prior and express written agreement of RPS.

The advice provided in this document does not constitute legal advice. As such, the services of lawyers may also be considered to be warranted.

rpsgroup.com

Contents 1 Introduction ...... 1 1.1 Project Aims ...... 1 1.2 Project phases ...... 1 2 Methodology ...... 2 2.1 National best practice guidance ...... 2 2.2 Application of UKWIR good practice ...... 3 2.3 Key assumptions ...... 6 2.4 Defining the micro-components and meter status categories ...... 7 3 Data ...... 8 3.1 Data sources ...... 8 3.2 Results from SWW customer surveys ...... 10 3.3 Introduction to micro-component assessments ...... 12 3.4 Component: Toilet Flushing ...... 12 3.5 Component: Personal Washing ...... 15 3.6 Component: Clothes Washing ...... 19 3.7 Component: Dishwashing ...... 21 3.8 Component: Miscellaneous Internal Use ...... 22 3.9 Component: External Water Use ...... 23 3.10 Data analysis ...... 25 3.10.1 Data input ...... 25 3.10.2 Choose scenario ...... 26 3.10.3 Base-case micro-component analysis ...... 26 3.10.4 Base year PCC reconciliation ...... 26 3.10.5 PCC and Ml/d forecasts ...... 28 3.10.6 Outputs ...... 29 4 Forecasting model ...... 30 4.1 Overview...... 30 4.2 Using Micro-F ...... 30 5 PCC forecasts ...... 32 5.1 Use of micro-components analysis ...... 32 5.2 Micro-component based PCC forecasts ...... 33

Figures Figure 1. Schematic of data and analysis steps ...... 25 Figure 2. Example of a base-year micro-component PCC reconciliation ...... 27 Figure 3. Comparison of MTP scenario forecasts of regional normal year average PCC (l/hd-d) (population weighted) ...... 35 Figure 4. Regional normal year average PCCs by MTP scenario and meter status type (l/hd-d) ...... 35 Figure 5. Historic actual and forecast normal year PCCs for unmeasured households (l/hd-d) ...... 36 Figure 6. Historic actual and forecast normal year PCCs for existing metered households (l/hd-d) ...... 36

i rpsgroup.com

Tables Table 1. Key recommendation principles identified in the UKWIR (2012) report for UKWIR and the approach to be used in this project ...... 4 Table 2. UKWIR (2012) recommended tiers of analysis ...... 5 Table 3. Comparison of forecasting approaches for Low or Intermediate tiers of analysis ...... 5 Table 4. Comparison of micro-component analysis methods for Low or Intermediate tier of analysis...... 6 Table 5. Summary of key data from SWW’s 2008 and 2011 surveys of customers on the household consumption monitor ...... 10 Table 6. Summary description of the model ...... 31 Table 7. Base year PCC reconciliation differences ...... 33 Table 8. Regional normal year micro-component PCCs (l/hd-d) for the MTP Policy scenario by meter status for 2011/12 after reconciliation adjustments ...... 34 Table 9. Regional normal year micro-component PCCs for the MTP Policy scenario (l/hd-d) ...... 34

Appendices Appendix A: MTP Reference, Policy and Earliest Best Practice (EBP) Scenarios ...... 37 Appendix B: Derivation of other data ...... 44

ii rpsgroup.com

1 Introduction

South West Water (SWW) appointed RPS to develop household per capita consumption (PCC) demand forecasts for the SWW region. The forecasts will be used as a key part of the development of the company’s 2014 Water Resources Management Plan (WRMP).

1.1 Project Aims The primary aim of the project is to develop a micro-component based PCC forecasting tool that produces PCC forecasts that SWW can use as part of its demand forecasts for the 2014 WRMP.

SWW requires that the model:-

 Covers the SWW region  Complies with the Environment Agency’s latest Water Resources Planning Guideline  Follows good practice recommended by the UKWIR report “A good practice manual and roadmap for household consumption forecasting (2012)”  Is fully editable, allowing SWW staff to update assumptions in the future

The required deliverables are:

 An Excel-based modelling tool that reliably forecasts PCC micro-components,  A technical report (this document), and  Training in the use of the model.

1.2 Project phases A phased approach has been undertaken as set out in the project proposal:-

 Phase 1. Project inception and review: to clarify SWW’s requirements, data availability, methods to be used (described in Section 2) and outputs required.

 Phase 2. Data collection: to collate data for the base-year and forecast trends for each micro-component (described in Section 3)

 Phase 3. Model development: to produce a fully tested and operational Excel-based forecasting tool to meet SWW requirements, as outlined in Section 4

 Phase 4. PCC forecasts: using the forecasting model to calculate scenario-based PCC forecasts. The calculations in the model involve using the micro-component values to derive “raw” PCC values, reconciling the PCC values to be consistent with the SWW Company Performance Review PCC values for 2011/12, applying water efficiency effects and applying the chosen scenario (e.g. choice of normal year or dry year, with or without climate change, with or without additional water efficiency). A summary of the PCC forecasts is presented in Section 5.

1 rpsgroup.com

2 Methodology

This Section describes the development of the methodology applied for deriving PCC forecasts, including:-

 Review of national guidance (Section 2.1)  Approach adopted to comply with guidance (Section 2.2)  Key assumptions used in the study (Section 2.3)  Identification of micro-components to be reported (Section 2.4)

2.1 National best practice guidance There are two main approaches that have been used by UK water companies to forecast PCC values:-

 Regression analysis to derive trends in PCC based on recent historic data. This may be based on a simple time trend or may include other explanatory variables such as weather parameters. A potential disadvantage is that future trends in PCC may be different from the past, due for example to changed customer behaviour, and so use of past trends may not be reliable.

 Micro-components analysis to estimate the consumption volumes associated with each component of household demand (e.g. toilet flushing, clothes washing, etc.). Usually micro-component values are calculated for each appliance type by multiplying estimates of % ownership (O), average frequency of use (F) and average volume per use (V). This method has the benefit of using explicit assumptions about how customer behaviour may change in the future, but has the disadvantage that many assumptions or judgement- calls are required.

The key documents that describe national best practice guidance for forecasting household consumption rates are:-

 Forecasting Water Demand Components, UKWIR/EA (1997). (Report by Cambridge Econometrics).

This report identifies micro-component analysis as the preferred approach for forecasting per capita consumption rates

 Water Resources Planning Guideline: The technical methods and instructions. Environment Agency (June 2012).

In this document, EA recommends that the micro-component approach to forecasting demand is the best way to understand household customer demand for water. It recognises that the UKWIR 2012 report (see below) will provide guidance on how to produce forecasts of consumption in different situations. The Guideline identifies six micro-components that water companies should report values for: toilet flushing; personal washing; clothes washing; dishwashing; miscellaneous internal use; and external use.

2 rpsgroup.com

 Customer behaviour and water use: A good practice manual and roadmap for household consumption forecasting. UKWIR (2012). (Report by Tynemarch for UKWIR)

This report provides detailed guidance on how to undertake micro-components analysis. It acknowledges the merits of comparing a trend-based PCC forecast and a micro- component PCC forecast to improve confidence in the forecasts.

Thus, all three documents recommend the use of micro-components analysis, and the UKWIR (2012) report provides the latest and most comprehensive guidance.

2.2 Application of UKWIR good practice

The development of PCC forecasts for SWW has therefore sought to apply the UKWIR (2012) good practice guidance, as outlined in Tables 1 to 4.

 Table 1 considers the key areas of improvement recommended by UKWIR (2012) for micro-component analysis compared with the approaches generally applied in 2009 WRMPs. A particular feature is the suggested introduction of three tiers of analysis: Low, Intermediate or High.

 Table 2 describes the data requirements and the demand forecasting drivers that apply for each of the three tiers of analysis. SWW does not expect significant supply-demand deficits to be identified in the 2014 WRMP, and so the Low tier of analysis could possibly be applied. However, the company wishes to apply the best approaches that it can, and so this study seeks to apply the Intermediate level of analysis where feasible.

 Table 3 summarises the key forecasting approaches that UKWIR (2012) identify as being consistent with the “Low” or “Intermediate” tier of analysis.

 Table 4 summarises the suggested method for calculating each micro-component O, F and V value. In most cases SWW have undertaken adequate studies to provide data to enable the Intermediate level of analysis. SWW participated in the WRc study “Increasing the value of micro-component data” (2005) and so micro-component monitoring data from SWW homes were included in the national survey by WRc. But SWW does not have any recent micro-component monitoring data from Identiflow or sub-metering studies, and so are unable to fully achieve Intermediate level of analysis for some items.

The assessments presented in Tables 1 to 4 identify the way in which this study can achieve nearly all the requirements of the Intermediate tier of analysis.

3 rpsgroup.com

Table 1. Key recommendation principles identified in the UKWIR (2012) report for UKWIR and the approach to be used in this project UKWIR recommendations for Principle improvement from the general PR09 Approach agreed for this project approach Intermediate tier to be used where Emphasis on the level of analysis being possible, as there is some vulnerability Level of analysis required linked to the nature and extent of any to future supply/demand imbalances. supply-demand imbalance. Water company specific customer survey data is available to support this. Standard, high level, micro-component categories are suggested to permit inter- Components used are to be those in Micro-component company comparison. Companies may the EA’s Table WRP2. This fully categories base the analysis on more detailed complies with requirements in EA’s categories that map to the standard revised WRP Guideline categories. Analysis is to be undertaken on each Assessment of the materiality of the meter status type without further changes in PCC estimates arising from segmentation (i.e. consistent with Accounting for segment properties in transition is suggested. intermediate tier). The transition of transition Explicit accounting for the effects of PCC customers from unmeasured to in the forecast is suggested where these metered status will be identified by the are shown to be significant. movements in property or population numbers for each meter status type. Analysis will not explicitly use the 125 l/hd-d standard in the new Building Assessment of the effect These have been introduced since PR09 Regulations, as SWW new connections of new Building and now require assessment of their PCC is already below this limit. Also Regulations effect. there is concern that actual PCC can vary significantly from the standard. A comparison is suggested to provide Comparison of micro- confidence in company forecasts. The This is not part of this study. SWW plan component forecasts differences, if any, should be considered to undertake an assessment of how the against regression-based by water companies and the adopted micro-component based PCC forecast approaches future consumption scenario explained in compares with past trends. the WRMP commentary. Alternative approaches for determining The study is to consider use of SWW Determining external use external use, not based upon the customer use data or consumption component standard micro-component approach, are values presented in WRc’s CP187 described. report Plumbing losses can be significant in unmeasured properties and their analysis SWW have derived specific estimates Specific identification of is not amenable to the standard micro- for plumbing losses. These have been the plumbing losses component approach. It is therefore included as part of Miscellaneous component suggested to include them in analysis as (internal) use a distinct component

4 rpsgroup.com

Table 2. UKWIR (2012) recommended tiers of analysis Tier Data requirement Drivers for demand forecast

Limited data except published sources that Low Limited issues with resources is able to support only limited analysis

Data which may be specific to the company Potentially some vulnerability to future Intermediate or geographical region that can support supply/demand imbalances that needs to further analysis be addressed

Extensive data from company surveys and Serious issues with resources requiring High other sources that can support detailed major investment to be considered analysis

Table 3. Comparison of forecasting approaches for Low or Intermediate tiers of analysis Issue Low tier Intermediate tier Tier achieved

Metered customers Segmentation Metered and unmetered according to origin of Intermediate metering Use of recent years Regression modelling of Base year PCC representative normal and PCC against summary Intermediate dry years weather and other factors Allocation of PCC to Base year micro- Adjustment of base year micro-components based component information using either simple pro-rata Intermediate upon proportions from and reconciliation or MLE published data Time series or regression Trends in PCC Micro-component forecasting Intermediate analysis of recent data Historic averages for Metering projections based changes in meter status. on other data sources. Segment transition Intermediate Simple assumptions Occupancy based on survey regarding occupancy results. Consumption of opting properties supported by Pre and post switching Impact of segment evidence from consumption consumption based on Intermediate transition on PCC monitoring, customer surveys published studies (for occupancy) and billing system (for post-opting PCC) Forecast micro- Apply PCC forecast Use micro-component PCC component information trends to base-year forecasts, taking account of Intermediate and reconciliation micro-component PCCs segment transition Use billing system Assume 125 l/head-d for New property information from recent new Normal year, with uplift for Intermediate consumption homes and occupancy from dry year. survey. Sensitivity analysis for Sensitivity of PCC to micro- Links to headroom assumed PCC forecast component assumptions and Intermediate analysis ranges and alternative population forecasts population forecasts

5 rpsgroup.com

Table 4. Comparison of micro-component analysis methods for Low or Intermediate tier of analysis. Component O, F or V Low tier Intermediate tier Tier achieved

O Assume 100% Assume 100% Intermediate Published data e.g. Own Identiflow/sub- F Low MTP metering studies WC flushing Own profile from V MTP profile survey and assumed Low replacement rates O Assume 100% Assume 100% Intermediate Published data e.g. Own Identiflow/sub- F Low Personal washing MTP metering studies Published data e.g. Own Identiflow/sub- V Low MTP metering studies Published data e.g. O Own survey Intermediate MTP Published data e.g. Model F with respect F Intermediate MTP to occupancy Clothes washing Published data e.g. MTP. Assume V Own profile from V reducing to current survey and assumed Intermediate most efficient over 25 replacement rates years Published data e.g. O Own survey Intermediate MTP Published data e.g. Model F with respect F Intermediate MTP to occupancy Dish washing Published data e.g. MTP. Assume V Own profile from V reducing to current survey and assumed Intermediate most efficient over 25 replacement rates years Published data e.g. Own Identiflow/sub- External use PCC Low MTP metering studies Published data e.g. Logged Individual Plumbing losses PCC Low UKWIR Household Monitor Miscellaneous Published data e.g. Own Identiflow/sub- PCC Low internal use MTP metering studies

Note: MTP = Market Transformation Programme, see Appendix A for more details

2.3 Key assumptions The key assumptions to be applied in this study have been agreed with SWW and are as follows:-

 SWW requires PCC forecasts that cover the SWW region. Zonal forecasts are not required by this project  The Environment Agency’s latest Water Resources Planning Guideline is to be complied with

6 rpsgroup.com

 The study should follow good practice as recommended by the UKWIR report “A good practice manual and roadmap for household consumption forecasting (CU02)”  Forecasts are required for each of the six micro-components specified by the EA Water Resources Planning Guideline  Separate forecasts are required for each household meter status category  Forecasts are required for each year from the base year of 2011/12 to 2039/40  Forecasts are required for the normal year and dry year scenarios, with and without climate change  The effects of base water efficiency activity are to be included  The PCC values should be consistent with SWW Company Performance Review reporting for the base year (2011/12), including meter under-registration but prior to maximum likelihood estimation (MLE)  SWW or MTP data should be used, wherever appropriate, for micro-component O, F or V values, in accordance with UKWIR good practice guidance  The MTP modelling of the effect of future replacement of appliances has been very detailed, has used a wider range of data sources than are available to SWW, and so should be used without the need for SWW modelling of future replacement of appliances.  Forecast PCCs are required for each of the three MTP scenarios (Reference scenario, Policy scenario and Earliest Best Practice scenario – see Appendix A) for sensitivity testing.

2.4 Defining the micro-components and meter status categories

The micro-components to be analysed are those required to be reported in the Environment Agency’s table WRP2, namely:

 Toilet flushing  Personal washing  Clothes washing  Dishwashing  Miscellaneous internal use (including plumbing losses)  External use

In some cases sub-components were identified for analysis, where it was important to examine distinctly different characteristics. In particular, baths, showers and washbasins for personal washing, and dishwashers and washing by hand for dishwashing.

The EA require water companies to report micro-components PCC values for unmeasured households and for metered households. However, as recommended by UKWIR for application of an intermediate tier of analysis, this study has calculated PCC forecasts for each of five meter status categories that are relevant, or could become relevant in the future, for SWW:-

 Unmeasured households  Existing metered households (i.e. all households metered up to 2011/12)  Future new connections (i.e. from 2012/13 onwards)  Future free meter options (i.e. from 2012/13 onwards)  Future compulsory metered (in case SWW consider this category in the future)

7 rpsgroup.com

3 Data

This Section describes the collation and analysis of data, including:-

 Identification of available data sources (Section 3.1)

 Review of SWW’s customer survey data (Section 3.2)

 Collation of data on O, F and V values for each micro-component (Sections 3.3 to 3.9)

 Data analysis steps (Section 3.10)

3.1 Data sources

The key data sources used for defining assumptions and calculation of micro-component O, F and V values are:

A. South West Water (2011). Results from 2011 survey of customers on SWW’s household consumption monitor (SODWAC) B. South West Water (2008). Results from 2008 survey of customers on SWW’s household consumption monitor (SODWAC) C. WRc (2005). Increasing the value of micro-component data. D. Market Transformation Programme (2010). BNW DW01: Dishwashers Government Standards Evidence Base 2009: Key Inputs E. Market Transformation Programme (2010). BNW DW03: Dishwashers Government Standards Evidence Base 2009: Policy Scenario F. Market Transformation Programme (2010). BNW01: Combined Laundry: Government Standards Evidence Base 2009: Key Inputs G. Market Transformation Programme (2010). BNW03: Combined Laundry Government Standards Evidence Base 2009: Policy Scenario H. Market Transformation Programme (2011). BNWAT01 WC market projections and product details I. Market Transformation Programme (2011). BNWAT02 Showers: market projections and product details J. Market Transformation Programme (2011). BNWAT03 Baths: market projections and product details K. Market Transformation Programme (2011). BNWAT04 Taps: market projections and product details L. Market Transformation Programme (2011). BNWAT06: Domestic water use in new and existing buildings Supplementary briefing note M. Market Transformation Programme (2011). BNWAT08: Modelling projections of water using products N. Market Transformation Programme (2012). BNXS25 UK household and population figures 1970-2030

8 rpsgroup.com

O. UKWIR (2012). Customer behaviour and water use: A good practice manual and roadmap for household consumption forecasting. (Report by Tynemarch for UKWIR Project CU02). P. Professor Paul Herrington (1996). Climate change and the demand for water (Report for Department of the Environment) Q. Which? website. http://www.which.co.uk/energy/creating-an-energy-saving- home/reviews-ns/water-saving-products/ R. Stamminger et al (2007). Washing up behaviour and techniques in Europe. S. Richter, C.P. and Stamminger, R. (2012). Water consumption in the kitchen – A case study in four European countries. Water Resource Management. DOI: 10.1007/s11269- 012-9976-5. T. Downing et al (2003). Climate Change and the Demand for Water. (Report for Defra)

The South West Water information provides recent data on appliance ownership and use for about 1000 metered households and over 300 unmetered households.

The WRc report records findings from monitoring of micro-component water consumption at 447 unmeasured households across England and Wales during 2000 to 2002. It therefore represents an important source of data on water use behaviour for unmeasured households, although the data is now over 10 years old. Some further studies have been undertaken for other water companies, but no subsequent study of this kind has been undertaken for which the results are available to SWW.

The Market Transformation Programme (MTP) documents examine current and predicted future ownership and use of domestic appliances. They are based on detailed studies of available data from within the water and energy industries, and appliance supplier organisations. Therefore they provide authoritative estimates of the ownership and usage of appliances. The relevant data available from MTP are presented in Appendix A.

The UKWIR (2012) report provides guidance to water companies on how to undertake micro- components analysis of household water consumptions, and quotes some data from other sources (in particular MTP or WRc). The report has been included in the review for this project, because in a few cases it presents data that is not available in the WRc or MTP documents.

Professor Herrington’s work represents one of the first published detailed micro-component studies. Although much of the assessment was based on desk studies and the work is over 15 years old, it remains a useful reference source for elements that are not covered by other sources of information.

The Which? website provides details of the water volumes of the most efficient white goods that are available on the market, which can be used as indicators of possible future trends in water use for washing machines and dishwashers.

The Defra report by Downing et al provides indicative values for the effect of climate change of micro-components of water demand (e.g. Table 3-14 for the SWW area), and so has been used

9 rpsgroup.com

in this study. A more up-to-date assessment is being prepared by UKWIR “Impact of Climate Change on Water Demand” (2012, in preparation) and so can be used instead when available.

3.2 Results from SWW customer surveys

SWW undertook customer surveys of appliance ownership and water use characteristics in 2008 and 2011. The customers involved were those who are part of the Company’s household consumption monitor: they were offered a cash reward for return of completed questionnaires to incentivise completion. The numbers of respondents in each case were:-

 Unmetered households in 2008: 370  Metered households in 2008: 1326  Unmetered households in 2011: 309  Metered households in 2011: 868

The numbers of respondents are small compared with the total of some 700,000 household customers across the region. However, confidence in their reliability can be gained by considering the consistency of the results with each other and with the Market Transformation Programme values from detailed national assessments.

The key data of most relevance to this study are summarised in Table 5, and are compared with values published by the Market Transformation Programme where there is a directly comparable number.

The 2011 survey average occupancy values are close to the regional average occupancies of 2.91 and 2.03, respectively. The regional average occupancy of metered households is significantly lower than for unmeasured households, due to the large numbers of low occupancy households that have opted to be metered.

Table 5. Summary of key data from SWW’s 2008 and 2011 surveys of customers on the household consumption monitor 2008 2008 2011 2011 MTP Item unmeasured metered unmeasured metered No. 370 1326 309 868 - respondents Average total 2.72 1.90 2.84 2.11 2.4 occupancy TOILETS Toilet 92.4 92.8 99.7 99.3 100 ownership % Cistern device 17.6 25.0 18.8 27.1 - ownership (%) Dual flush toilet 34.6 30.0 49.8 42.3 - ownership (%) PERSONAL WASHING Bath ownership 85.4 82.4 91.9 85.0 94 % Basin 90.8 91.5 98.4 97.8 100 ownership %

10 rpsgroup.com

Shower (over bath) ownership 53.2 56.0 60.2 58.2 - % Shower (separate) 31.4 38.2 38.2 40.4 - ownership % Shower (pumped) 10.3 16.5 12.3 15.7 - ownership % Shower (any) 78.9 86.0 89.0 92.3 81 ownership % Bidet ownership 3.0 6.3 2.9 5.1 - % WHITE GOODS Washing machine 89.5 88.8 97.7 95.0 95 ownership (%) Washing Lifespan machine age 4.9 6.0 5.0 6.2 12.6 years (years) Washing machine uses 0.63 0.48 0.66 0.50 0.71 per day Dishwasher 41.9 39.0 46.9 42.2 36 ownership (%) Dishwasher Lifespan 4.5 6.4 5.4 6.1 age (years) 13.0 years Dishwasher 0.63 0.54 0.70 0.57 0.67 uses per day EXTERNAL WATER USE Swimming pool 0.3 1.1 0.3 0.8 - o’ship (%) Hot tub 0.5 0.4 0 0.6 0.4 ownership (%) Outside tap 67.6 67.4 76.7 74.4 WRc = 65% ownership (%) Garden > 100m2 40.0 43.0 46.0 41.7 - ownership (%) Watering can 38.6 38.2 84.5 79.6 - ownership (%) Watering can uses per day (in 1.4 1.3 1.1 1.1 - garden) Hosepipe No data No data No data No data - ownership (%) Garden sprinkler 0.8 2.6 2.9 4.8 - ownership (%)

11 rpsgroup.com

The cells in Table 5 that are highlighted yellow indicate values that have been used in the study, for the reasons explained in Sections 3.4 to 3.9. The basis for selecting preferred data sources is:-

 In general SWW survey data are preferred to values from elsewhere because they are considered to better reflect local characteristics. Also, they are comparable with the MTP estimates from national assessments and so they are considered to be robust.

 Values from the 2011 surveys are preferred to use of values from the 2008 surveys because they are more recent. The differences suggested between 2008 and 2011 are often small and/or are plausible.

 100% toilet ownership has been assumed as it is expected that all households have access to a flushing toilet, even if in a very small number of cases it is not located internal to their home.

3.3 Introduction to micro-component assessments

The following sections (Sections 3.4 to 3.9) examine the available data from the SWW surveys or elsewhere for each micro-component, and give more detail on the choice of values or assumptions for the assessment for per capita consumption. Yellow highlights are used to indicate the values that have been used in preference to other available assessments.

Defra’s Market Transformation Programme (MTP) has published various documents (at: http://efficient-products.defra.gov.uk ) that describe current and future ownership and usage of domestic appliances. MTP provide projections for three scenarios:-

 The “Reference Scenario”: This is a projection of what is likely to happen without any new policy intervention. The scenario is based on current trends, technology developments and policies that are already in place.

 The “Policy Scenario”: This scenario estimates what could be achieved through an ambitious but feasible set of policy measures if the agreement of all stakeholders was obtained.

 The “Earliest Best Practice Scenario” (EBP): This is a projection of what could happen if the best available products and technologies were adopted, coupled with ambitious Government policies.

The forecast O, F and V values presented in the following sections relate to MTP’s Policy scenario, as an example. Details of how forecast values vary between the Reference, Policy and Earliest Best Practice (EBP) scenarios are shown in Appendix A.

Other key data that have been used are described in Appendix B.

3.4 Component: Toilet Flushing For the purpose of micro-components analysis all toilet flushing can be considered as a single category. Although there are a wide range of types of toilet their operation is similar and the frequency of use can be assumed to be the same.

12 rpsgroup.com

COMPONENT: TOILET FLUSHING CATEGORY: TOILET FLUSHING TYPES: MTP (ref M) identify a wide range of different toilet types with different flush volumes. For the purpose of this study, they have been grouped into the following main types:  Old toilets without cistern device (assumed average of 8 litre) – representing the assumed average volume of old toilets still in use  Old toilets with cistern device (assumed average of 7 litre)  6 litre – installed widely in the early 2000s but declining ownership, as low-volume dual- flush WCs are now preferred  6/4 litre dual flush (i.e. average 5 l) – main type currently being installed  6/3 litre dual flush & 4.5 litre single flush (i.e. average 4.5 l) – expected to increase in future  4/2.6 litre dual flush & <4.5 litre single flush (i.e. average 3.3 l) – expected to increase in future

SWW have been issuing free cistern devices for many years, as demonstrated by the high proportion of households in the customer surveys that reported having installed cistern devices (see Table 1). These cistern devices can be expected to reduce flush volume by about 1 litre.

SWW have suggested that, for the purpose of this study, old toilets without cistern devices should be grouped into a single “8 litre” category and old toilets with a cistern device should be grouped into a single “7 litre” category. These include the following types:

 9 litre syphon operated WCs – very widely installed in the past but illegal to install since 1990s. The actual flush volume can vary very widely. Some have a cistern device installed.  7.5 litre WCs – illegal to install since 2001. Some were dual-flush. Some have a cistern device installed.

MTP assume that the average replacement rate for toilets is 1 in 15 years, and have used this assumption in their modelling of future ownership of different toilet types.

TOILET OWNERSHIP (%) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. - - 92.4% 92.8% 99.7% 99.3% 100% 100% 100%

Values chosen: It is surprising that some households in the SWW surveys have not recorded having a toilet. We assume this is an omission. For this study it is assumed that all homes have a toilet.

The SWW survey indicates that about 20% of homes have a cistern device installed. MTP (ref M) estimate that there are old toilets at approximately a third of homes. So we have assumed that at

13 rpsgroup.com

2010 13% of homes have the “8 litre” type and 20% have the “7 litre” type.

We have used the estimated stock levels and forecasts in MTP (see Appendix A) for a wide range of toilet types to derive the following estimates and forecasts of ownership for each toilet group:-

All households except future new connections Toilet type 2010 ownership 2020 ownership 2030 ownership “8 litre” 13% 2% 0% “7 litre” 20% 2% 0% “6 litre” 23% 9% 1% “6/4 litre” 44% 58% 22% “6/3 litre + 4.5 litre” 0 23% 56% “4/2.6 litre” 0 6% 21% All types 100% 100% 100%

We have assumed the ownership profiles above for all meter status’ except future new connections, as they will have 6 litre maximum flush volumes installed. For future new connections we have assumed the same profile as above for current and future toilet types, but have re- allocated the “8 litre” and “7 litre” % values to the 6/4 litre type:-

Future new connections Toilet type 2010 ownership 2020 ownership 2030 ownership “8 litre” 0% 0% 0% “7 litre” 0% 0% 0% “6 litre” 23% 9% 1% “6/4 litre” 77% 62% 22% “6/3 litre + 4.5 litre” 0% 23% 56% “4/2.6 litre” 0% 6% 21% All types 100% 100% 100%

TOILET FREQUENCY OF USE (use/person-d) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. ------4.68 4.71 4.71

Values chosen: The MTP (ref M) value is based on data from various water companies and WRc micro- component studies. It is close to the WRc measured value. UKWIR quote the MTP value.

A value of 4.71 uses/person-d has therefore been assumed. In accordance with MTP, it is assumed to remain constant over time.

It has been assumed that the average frequency of use in metered households is 14% lower (see Appendix B).

14 rpsgroup.com

TOILET VOLUME PER USE (l/use) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. ------9.4 various -

Values chosen: This study has assumed the average flush volume for each toilet type considered: 8 litre, 7 litre, 6 litre, 5 litre, 4.5 litre and 3.3 litre, respectively.

3.5 Component: Personal Washing Personal washing can include washing by shower, bath, basin or bidet. The SWW customer surveys (Table 1) indicate that there are small numbers of households that use a bidet. The quantities of water involved are not well understood and are relatively very low because of low ownership levels. Therefore, the use of bidets has been included in Miscellaneous internal use (Section 3.7) rather than being specifically estimated as part of Personal washing. Hot tubs are considered in Section 3.8 under external use.

The categories of personal washing analysed for this study are: showers, baths and washbasins.

COMPONENT: PERSONAL WASHING CATEGORY: SHOWERS TYPES: MTP (refs I and M) identify 3 main types of shower: electric shower, gravity mixer shower and pumped shower.

More details are provided in Appendix A. SHOWER OWNERSHIP (%) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. - - 78.9% 86.0% 89.0% 92.3% 85.2% 81% -

Values chosen:

Table 5 summarises the average ownership of showers recorded in the SWW customer surveys. The ownership of individual shower types add to more than 100% because some homes own more than one type of shower. Overall ownership of one or more showers in 2011 was 89.0% for unmetered households and 92.3% for metered households. These are higher than the national average of 81% estimated by MTP. But the SWW values have been assumed in this study as being more representative of ownership in the region.

MTP (refs I and M) present current estimates and forecast future ownership of each shower type

15 rpsgroup.com

for the assumed MTP Policy scenario:

Type of shower 2010 ownership 2020 ownership 2030 ownership Electric 39% 46% 48% Gravity mixer 22% 24% 25% Pumped 20% 21% 21% Total 81% 91% 94%

The MTP forecasts suggest that the percentage of homes without a shower will reduce by approximately 70% by 2030, and this has been used as the basis for the forecasts in this study using the SWW values for 2010 levels, as shown below.

Values chosen for unmeasured households: Type of shower 2010 ownership 2020 ownership 2030 ownership Electric 43% 48% 49% Gravity mixer 24% 25% 26% Pumped 22% 22% 22% Total 89% 95% 97%

Values chosen for metered households: Type of shower 2010 ownership 2020 ownership 2030 ownership Electric 45% 49% 49% Gravity mixer 25% 25% 26% Pumped 22% 22% 22% Total 92% 96% 97%

SHOWER FREQUENCY OF USE (use/prop-d) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. ------1.46 1.04 -

Values chosen: MTP (ref M) estimate the average frequency of showering as 1.04 per household-day in 2010, rising to 1.21 per household-day in 2020 and to 1.33 per household-day in 2030 (according to MTP’s policy scenario). The MTP assessment has been used for this study as it has been based on a wider survey of reported shower usage than the WRc study.

SHOWER VOLUME PER USE (l/use) Values observed or reported: WRc UKWIR SWW (2008) SWW (2011) MTP (2010) (2002) (2012) Unmeasured Metered Unmeasured Metered Unm. - -

16 rpsgroup.com

26.5 (Elec.) - - - - 25.7 38.2 (Grav.) 62.7 (Pump.)

Values chosen: For unmeasured households the MTP projections for the Policy scenario have been used:

Type of shower 2010 volume (l) 2020 volume (l) 2030 volume (l) Electric 26.5 26.7 26.6 Gravity mixer 38.2 38.5 37.8 Pumped 62.7 63.2 62.1

It has been assumed that the average volume per use in metered households is 14% lower (see Appendix B).

COMPONENT: PERSONAL WASHING CATEGORY: BATH TYPES: Although there are a wide range of styles of bath available, with different volumes, this study has assessed them as a single type as their operation is similar and the frequency of use can be assumed to be the same. More details are presented in Appendix A.

BATH OWNERSHIP (%) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. - - 85.4% 82.4% 91.9% 85.0% 88.1% 94% -

Values chosen: Bath ownership levels in 2011 according to the SWW customer survey are slightly less than the national average of 94% estimated by MTP. (This is consistent with the higher ownership rates for showers than estimated by MTP). But the SWW values have been assumed in this study as being more representative of ownership in the region.

MTP (ref M) estimate that the ownership of baths in 2010 at 94%, and assume that it will reduce to 91% in 2020 and 83% in 2030. The reducing trend is based on the trend towards greater use of showers for washing. This study has therefore assumed that bath ownership in metered and unmetered households will reduce by 3% by 2020 and by 11% by 2030.

BATH FREQUENCY OF USE (use/prop-d) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. ------0.95 0.68 -

17 rpsgroup.com

Values chosen: The MTP assessment has been used for this study as it has been based on a wider survey of reported bath usage than the WRc study. MTP (ref M) estimate that the average frequency of use was 0.68 per household-day in 2010, and assume that it will reduce to 0.55 per household-day in 2020 and 0.46 per household-day in 2030 (according to MTP’s policy scenario). This reflects the expectation that in the future more people will use showers instead of baths for washing. BATH VOLUME PER USE (l/use) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. ------73.3 84.5 -

Values chosen: MTP (ref M) estimate that the average volume per use was 84.5 litres per use in 2010, and assume that it will reduce to 80.6 litres per use in 2030 (according to MTP’s policy scenario). This is in line with assumed increased use of showers in the future. The MTP assessment has been used for this study for unmeasured households as it has been based on a wider survey of reported bath usage than the WRc study. It has been assumed that the average volume per use at metered households is 14% lower (see Appendix B).

COMPONENT: PERSONAL WASHING CATEGORY: WASHBASIN TYPES: Although there are a wide range of styles of washbasins available, with different types of tap, this study has assessed them as a single type as their operation is similar and the frequency of use can be assumed to be the same. This is consistent with the information presented by MTP (ref K). More details are presented in Appendix A.

WASHBASIN OWNERSHIP (%)

Values observed or reported:

WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. - - 90.8% 91.5% 98.4% 97.8% 100.0% 100.0% -

Values chosen: In line with MTP (ref K) it is assumed that all homes include access to a washbasin for personal washing including shaving and teeth cleaning.

WASHBASIN FREQUENCY OF USE (use/prop-d)

Values chosen:

18 rpsgroup.com

The only authoritative information available is that presented by MTP (ref K). They estimate that the average frequency of washbasin use is 8 per person-day (i.e. 19.2 per household-day in 2010, assuming the MTP average occupancy of 2.4). MTP assume that this will remain unchanged in the future.

WASHBASIN VOLUME PER USE (l/use)

Values chosen: The only authoritative information available is that presented by MTP (ref K). They estimate that the average volume per washbasin use was 2.27 litres in 2010, and is forecast to reduce to 2.13 litres in 2020 and then stay at that volume (according to the MTP Policy scenario). The reductions arise due to the assumed installation of lower flow-rate taps in the future.

Note: the MTP documents (refs K and M) are inconsistent in the values reported for current (2010) average volume per use – this study has used the quoted value of 2.27 litres for the Policy scenario in ref K.

It has been assumed that the average volume per use at metered households is 14% lower (see Appendix B).

3.6 Component: Clothes Washing

Clothes washing can be undertaken by washing machine (including washer-driers), by hand or at laundrette. Clothes washing by hand is very infrequent and there is no specific data to quantify its use, and so it has been assumed to be part of the Miscellaneous internal use component. Clothes washing at laundrettes is not part of household water use and so is not considered by this study.

Therefore the only category of water use in clothes washing that has been assessed is for washing machines. COMPONENT: CLOTHES WASHING CATEGORY: WASHING MACHINE TYPES: This study has not sub-divided washing machines into various types, instead it uses a single average volume that changes with time. This is consistent with the data presented by MTP.

MTP (ref F) estimate that the average lifespan of a washing machine is 12.6 years, and have used this assumption in their modelling of future replacement of washing machine with lower water using versions. The average ages of washing machines in the SWW survey (see Table 1) are consistent with the MTP estimated lifespan of 12.6 years. WASHING MACHINE OWNERSHIP (%) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. - - 89.5% 88.8% 97.7% 95.0% 93.7% 95% >90%

19 rpsgroup.com

Values chosen: Washing machine ownership levels recorded in the SWW 2011 survey are close to the 95% level estimated by MTP and have been used as being representative of ownership in the region.

MTP (ref F) assume that washing machine ownership will remain at current levels. This study has therefore assumed no change in future.

Note: The value reported by MTP (ref F) includes washing machines (80%) and washer driers (15%). WASHING MACHINE FREQUENCY OF USE (use/prop-d) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. - - 0.63 0.48 0.66 0.50 0.81 0.71 - Values chosen: The average usage rates for washing machines are lower than the MTP estimated average of 0.71, or the WRc estimate of 0.81. The MTP assessment has been based on a wider survey of reported washing machine usage than the WRc study. But the SWW values have been assumed in this study as being more representative of usage in the region, and reflect the anticipated lower use by metered customers who have a financial incentive to use water wisely.

MTP (ref F) report that the average frequency of use of washing machines has reduced from 274 per year (i.e. 0.75 per day) in 2000 to 260 per year (i.e. 0.71 per day) in 2010, and is expected to remain at this level. This study has therefore assumed that frequency of use will remain at current levels.

In order to calculate PCC values it is necessary to divide the per-household frequency values by average occupancy to estimate average per-person frequencies of use. The frequency of use of washing machines reduces with reducing average occupancy, although the relationship is not well understood. SWW specific average occupancies for each meter status type have been used.

WASHING MACHINE VOLUME PER USE (l/use) Values observed or reported: WRc MTP UKWIR SWW (2008) SWW (2011) (2002) (2010) (2012) Unmeasured Metered Unmeasured Metered Unm. ------61 - ~60

Values chosen: MTP do not estimate the average water use by washing machines, but document MTP (ref G) predicts that the average energy consumption per wash will reduce to the typical value of a current “A” graded washing machine. Which? identify that the current lowest water using

20 rpsgroup.com

washing machine uses 31 litres per cycle.

For this study it is assumed that the average volume per use will reduce from 60 litres in 2010 to 31 litres in 2030. See Appendix A for more details.

3.7 Component: Dishwashing

Dishwashing includes washing of dishes using a dishwasher or by hand. Both categories have been assessed by this study.

COMPONENT: DISHWASHING CATEGORY: DISHWASHER TYPES:

MTP (ref D) estimate that the average lifespan of a dishwasher is 13.0 years, and have used this assumption in their modelling of future replacement of dishwashers with lower water using versions. The average ages of dishwashers in the SWW survey (see Table 1) are consistent with the MTP estimated lifespan of 13.0 years.

More details are presented in Appendix A.

DISHWASHER OWNERSHIP (%) Values observed or reported: WRc MTP SWW (2008) SWW (2011) UKWIR (2012) (2002) (2010) Unmeasured Metered Unmeasured Metered Unm. - - 41.9 39.0 46.9 42.2 37.0% 36% -

Values chosen: Dishwasher ownership levels according to the SWW customer survey are higher than the 36% level estimated by MTP. But the SWW values have been used as being representative of ownership in the region.

MTP (ref D and ref N) estimate that 36% of UK households owned a dishwasher in 2010, and predict that this will increase to 40% at 2020, and stay at that level. Therefore this study has assumed that dishwasher ownership by unmeasured and metered households will increase by 4% by 2020.

DISHWASHER FREQUENCY OF USE (use/prop-d) Values observed or reported: WRc MTP SWW (2008) SWW (2011) UKWIR (2012) (2002) (2010) Unmeasured Metered Unmeasured Metered Unm. - - 0.63 0.54 0.70 0.57 0.71 0.67 -

21 rpsgroup.com

Values chosen: The average usage rates for washing machines according to the SWW 2011 customer survey are similar to the MTP estimated average of 0.67. But the SWW values have been assumed in this study as being more representative of usage in the region, and reflect the anticipated lower use by metered customers who have a financial incentive to use water wisely.

MTP (ref D) estimate that average household dishwasher usage has reduced from 250 per year (i.e. 0.68 uses per day) in 2000 to 245 per year (i.e. 0.67 uses per day) in 2010, and assume that it will reduce to 236 per year (0.65 uses per day) at 2030. This decline takes into account potential for less frequent use in smaller households and the increase in capacity of future dishwashing machines. Therefore this study has assumed that dishwasher frequency of use by unmeasured and metered households will decrease each decade by 0.01 uses per day.

In order to calculate PCC values it is necessary to divide the per-household frequency values by average occupancy to estimate average per-person frequencies of use. The frequency of use of dishwashers reduces with reducing average occupancy, although the relationship is not well understood. SWW specific average occupancies for each meter status type have been used.

3.8 Component: Miscellaneous Internal Use

Internal tap use by washbasins has been included in personal washing (Section 3.5), and internal tap use for hand-washing of dishes has been included in dishwashing (Section 3.6).

Miscellaneous internal water use comprises all other water uses, in particular water drawn from taps at the kitchen sink, bidets or utility room sink. This may include: washing of clothes, floors, kitchen surfaces or other household items; plumbing losses; drawing off water from taps to obtain hot water; water softeners; waste disposal units; animal use; indoor watering of plants; DIY; hobbies; medical; Jacuzzis; or future new water uses.

COMPONENT: MISCELLANEOUS CATEGORY: MISCELLANEOUS INTERNAL INTERNAL USE USE

It is assumed that all households use water for miscellaneous internal water uses.

WRc (ref C) report an average internal tap use of 87.2 l/household-d and an average use for water softeners of 1.1 l/household-d, giving a total of 88.3 l/household-d.

Richter and Stamminger (ref S) have recently monitored water use in the kitchen at 81 urban households across four European countries including 20 homes in the UK. Average UK water use was observed to be 19.8 l/household-day, of which over 50% was for dishwashing. This is similar to the MTP assessments (see below).

MTP (ref K) estimate average water use at internal taps in 2010 as:  Washbasin: 8 uses/hd-d x 2.27 l/use = 18.2 l/hd-d  Kitchen: 9 uses/hd-d x 2.28 l/use = 20.5 l/hd-d

22 rpsgroup.com

 i.e. a total of 38.7 l/hd-d, which is equivalent to 92.9 l/household-d assuming an average occupancy of 2.4, and so is similar to the WRc assessment.

MTP have forecast that kitchen tap use will reduce from 20.5 l/hd-d (i.e. 49.2 l/prop-d) at 2010 to 18.3 l/hd-d (i.e. 43.9 l/prop-d) at 2030 (according to the MTP Policy scenario). This represents a reduction of 5.3 l/household-d between 2010 and 2030 due to efficiency measures such as wider use of low flow-rate taps. See Appendix A for more details.

Internal tap use for washing dishes by hand (estimated as 9 l/prop-d, see Section 3.6) has been included in the Dishwashing component, and so the values used for each year must be deducted.

Plumbing losses are not included in internal tap use calculations and so need to be added. SWW estimate plumbing losses at 14.852 l/prop-d, based on Managing Leakage (2011) and company studies.

Herrington (ref P) defined this component slightly differently, but forecast that miscellaneous use would grow by 0.247 l/hd-d per annum, in particular due to the growth in minor water uses and introduction of new ways customers would use water in the future. This would be equivalent to an increase of 11.9 l/household-d between 2010 and 2030.

The following table shows the effects (l/prop-day) of combing these elements: 2010 2020 2030 Kitchen tap use 49.2 45.6 43.9 Hand-washing dishes -9 -9 -9 Plumbing losses 14.9 14.9 14.9 Increase in uses 0 6 11.9 Total 55.1 57.5 61.7

See Appendix A for more details. For metered households, it is assumed that the average volume of use is 14% lower (see Appendix B).

3.9 Component: External Water Use

External use includes water used externally for garden watering, car washing and other uses.

COMPONENT: EXTERNAL WATER USE CATEGORY: EXTERNAL WATER USE MTP (ref L) identify the following outdoor uses of water but provide limited details of the average quantities of water used:-

 Garden watering (by hosepipe, sprinkler or watering can)  Filling/topping-up of ponds and water features  Pressure washers (for outdoor cleaning)  Recreational water use (by swimming pools, hot tubs or paddling pools).

23 rpsgroup.com

MTP suggest that, based on information from the British Swimming Pool Federation, average water use by swimming pools is 66 l/d and by hot tubs is 36 l/d. Ownership levels are very low, but the SWW customer survey suggests current ownership levels of approximately 0.5% for swimming pools and close to the MTP estimate of 0.4% for hot tubs. These suggest estimated average water uses (across all homes) of 0.33 l/household-day for swimming pools and 0.14 l/household-day for hot tubs. The quantities involved are not well understood and are negligible, and so can be ignored.

WRc (ref C) reported on water use from external taps, and found that 65% of homes used an external tap on an average of 0.89 times a day, with an average use of 46.7 l/household-d. This study has assumed that unmeasured households with an external tap use 46.7 l/household-d but that metered households use 14% less (see Appendix B).

The SWW 2011 customer survey found that 76.7% of unmeasured and 74.4% of metered households own an external tap. These are slightly higher than the WRc value of 65%, but the SWW values have been assumed in this study as being more representative of the SWW area.

The SWW survey data for ownership of watering cans, hosepipes and garden sprinklers is summarised in Table 1, but has not been used as there is no or insufficient information concerning possible usage across an average year to be able to use this data in this project.

UKWIR (2012) (ref O) has considered the possibility of combining data from WRc Identiflow surveys and garden size to quantify external water use. SWW do not have Identiflow data and so this is not feasible.

Based on the information above, this study has assumed that for a normal year, average external water use can be estimated as:

 76.7% ownership * 0.89 use/day * 46.7 l/household-d for unmetered households  74.4% ownership * 0.89 use/day * 40.2 l/household-d for metered households

It has been assumed that this will not change in the future.

The increase in water use in hot, dry weather is predominantly due to increased external water use. Therefore the external water consumption rates will be significantly increased for dry year or climate change scenarios. Data to quantify this for SWW is not yet available.

24 rpsgroup.com

3.10 Data analysis The key steps involved in analysing the O, F and V data to derive PCC forecasts are illustrated in Figure 1 and are summarised below.

Regional Populations, Properties & Occupancy

x 5 meter status types

Dry year, normal year and climate change factors

Ownership (O), Frequency (F), JR12 Table 10 regional PCCs Volume (V) SWW water efficiency effects x 5 meter status types (*) from SWW surveys + MTP + Base year only other published sources

Base year PCC Base-case micro- PCC & Ml/d forecasts reconciliation component analysis Choose x 5 meter status types scenario x 5 meter status types x 5 meter status types x Normal year, Dry year, Climate x 6 EA micro-components x 6 EA micro-components change scenarios Base-year weather Base year only Post-water efficiency Post-MUR, pre-MLE PCCs

KEY:

Data input steps Meter status types = EA micro-components = Unmetered WC flushing Existing metered Clothes washing Analysis steps Future FrOpt Personal washing (bath + shower + Future new connections handbasin) Future compulsory Dishwashing (dishwasher + hand) Miscellaneous internal use (including plumbing losses) External use Figure 1. Schematic of data and analysis steps

3.10.1 Data input The data input requirements comprise:-

 O, F and V values, as derived in Sections 3.4 to 3.9

 Regional PCC values for each meter status type for the base year (2011/12) consistent with the SWW Company Performance Review 2012. These are inclusive of meter under- registration and some plumbing losses, but before MLE. These have been provided by SWW for the project.

 Regional population, household and (implied) occupancy numbers for each year from 2011/12 to 2039/40, for each meter status type. These have been provided by SWW for the project.

25 rpsgroup.com

 Normal year, dry year and climate change factors.

 Water efficiency effects on micro-component PCCs. Dummy values have been used in the model at present.

Further details are presented at Appendix B.

3.10.2 Choose scenario This stage is to define the scenario choices for which the forecasts are required.

There are two main choices that need to be made:

 The choice of MTP forecast scenario (i.e. Reference, Policy or EBP) specifies which O, F and V values are used in the “Base-case micro-component analysis”, and the trends in O, F and V that are carried through to the later analysis steps.

 The choice of demand planning scenario (i.e. normal year or dry year, with or without climate change) determines the factors to be applied for weather effect in the “PCC and Ml/d forecasts” stage. Similarly, the choice of which (if any) water efficiency effects to be used determines the water efficiency adjustments to the calculations in the “PCC and Ml/d forecasts” stage.

3.10.3 Base-case micro-component analysis This stage derives the base-case O, F and V forecasts, and resulting PCC forecasts, for the chosen MTP scenario. No reconciliation with SWW values is undertaken at this stage.

The key steps are:

 Derive O, F and V values for each micro-component, for each meter status, and for each year from 2011/12 (base-year) to 2039/40, based on the chosen MTP scenario, using the information presented in the preceding sections.

Note: The O, F and V values remain fixed for the chosen MTP scenario. They are not subsequently adjusted for base-year PCC reconciliation or for calculation of zonal forecasts.  Multiply the O, F and V values to calculate initial PCC forecasts for each micro- component and each meter status.

3.10.4 Base year PCC reconciliation This stage examines the base year (2011/12) data. It reconciles the base-case PCC values (from the preceding stage) to the SWW PCC values for 2011/12 for the chosen water resource zone and each meter status type. The calculated adjustments are then applied to forecast micro- component PCCs as well as base-year.

The key steps are:

26 rpsgroup.com

 Identify the regional base-year total PCC values from the data input. These PCC values are consistent with the SWW Company Performance Review 2012 (inclusive of meter under-registration but before maximum likelihood estimation water balance reconciliation).

 Undertake the base-year PCC reconciliation for each meter status type. This is required because the base-case micro-component PCC values do not match the total PCC values based on SWW’s Company Performance Review 2012. The micro-component PCCs therefore have to be adjusted to achieve reconciliation. The adjustments have been calculated in proportion to the initial PCC values, which is a method recommended by UKWIR.

An example is shown as Figure 2, for unmeasured households. It assumes that the base year total PCC is 150 l/hd-d and that the sum of base-case micro-component PCC estimates is 138.95 l/hd-d.  Adjust forecast micro-component PCC values by adding the base-year adjustment values. (Figure 2 shows the effects of the adjustments for the example presented).

UNMEASURED HOUSEHOLD BASE YEAR PCC RECONCILIATION

June Return Table 10 PCC 150 l/head-day

Micro-component values for base Initial Initial Adjustment Final year (l/hd-d) (%) (l/hd-d) (l/hd-d) Toilet flushing 27.86 20.1 2.22 30.08 Personal washing Baths 21.41 15.4 1.70 23.12 Electric Shower 5.08 3.7 0.40 5.48 Gravity Mixer 4.06 2.9 0.32 4.38 Pumped 6.08 4.4 0.48 6.56 Clothes washing 15.72 11.3 1.25 16.97 Dish washing Dishwasher 2.70 1.9 0.22 2.92 Handwash - dishes 3.75 2.7 0.30 4.05 Miscellaneous internal use 40.56 29.2 3.23 43.79 External use 11.72 8.4 0.93 12.66 Total 138.95 100.0 11.05 150.00

Imbalance 11.05 0.00 Figure 2. Example of a base-year micro-component PCC reconciliation

It is important to note that the ownership (O), frequency of use (F) and volume per use (V) values are not recalculated in this study. The changes to PCC values resulting from the base year reconciliation (or subsequent application of normal year/dry year/climate change effects or water efficiency effects) are not applied to the O, F and V values. This is for three reasons: (a) it is difficult to know which values (O or F or V or all three) should be changed; (b) it could be difficult for the user to track the changes occurring; and (c) it is not necessary to change the O, F or V values as only PCC values are required to be reported in the EA’s WRP tables.

27 rpsgroup.com

3.10.5 PCC and Ml/d forecasts At this stage the PCC and volume (as Ml/d) forecasts are calculated for the chosen MTP scenario and weather scenario. It uses the forecasts from the preceding stage (base year reconciliation).

The key steps are:

 Use the choices made at “Choose scenario”. These defined the required forecasting scenario:

(1) Choice of MTP forecasting scenario: either the MTP Reference or Policy or Earliest Best Practice scenario can be chosen. (2) Choice of normal year or dry year PCC forecasts, with or without climate change.

(3) In addition a further choice can be considered: Choice of whether to deduct the effects of water efficiency on micro-component PCCs, if desired.

 Calculate final PCC forecasts for the required scenario, for each meter status type.

 Calculate volume forecasts (Ml/d) for the required scenario, for each meter status type. These are obtained by multiplying PCC values by the appropriate population estimates and forecasts.

The generalised equation for a micro-component PCC forecast for a particular year is:

PCC for micro-component = [ O*F*V + Reconciliation * Weather_effect]

– Water_efficiency Where:

O, F and V = micro-component % ownership (O), frequency of use (F), and volume per use (V) Reconciliation = adjustment to micro-component PCC at base-year

Weather_effect = combined factor effect of demand scenario (normal year, or dry year, with or without climate change Water_efficiency = effect of water efficiency on micro-component PCC

The total PCC for any meter status type for a particular year is calculated by summing the relevant individual micro-component PCC values. The generalised equation for total PCC forecast is:-

28 rpsgroup.com

Total PCC = Sum of micro-component PCCs for:

Toilet flushing Personal washing (baths + showers + washbasin) Clothes washing Dish washing (dishwasher + washing by hand) Miscellaneous internal use External use

3.10.6 Outputs The forecasts can be presented in a variety of tables and/or charts using “Micro-F”

29 rpsgroup.com

4 Forecasting model

This Section outlines the forecasting model, “Micro-F”, that has been developed by RPS to calculate the required micro-component PCC forecasts.

4.1 Overview A forecasting tool, named “Micro-F”, has been developed to produce the required micro- component PCC forecasts for each required scenario. It has been set up in Excel 2007 and the current version provided to SWW is named: SWW Microcomponent PCC Forecast Model V1.1.xls

Micro-F applies the data and analysis steps described in Figure 1 and Section 3.10. The sheets in the model are described in Table 6.

A demonstration of Micro-F was provided to SWW on 9th August 2012.

4.2 Using Micro-F Micro-F has been developed to be very easy to use. The steps are:-

 Step 1: Choose scenario. This is undertaken using the pick-lists in sheet Main Page

 Step 2: Check or update data values in data input cells. The data input cells are coloured maroon, in the “Data input” and “Regional micro-component analysis” sheets listed in Table 6. Also, the tick-boxes in the sheet “Water efficiency” should be viewed to ensure the correct use of ticks (or not) has been used.

 Step 3: Obtain results from the output tables or charts (for the scenario chosen in the Main Page). The tables or charts required can be selected by using pick-lists.

As agreed with SWW, a formal user guide is not required, and so has not been produced. However, the model includes a Readme sheet that provides an outline description of the model and how to use it. Also, RPS will provide a demonstration session to SWW personnel on how to use it.

As noted in Section 3.10.4, the ownership (O), frequency of use (F) and volume per use (V) values are not altered in the model. The changes to PCC values resulting from the base year reconciliation, application of normal year/dry year/climate change effects, or application of water efficiency effects are not applied to the O, F and V values. This is for three reasons: (a) it is difficult to know which values (O or F or V or all three) should be changed; (b) it could be difficult for the user to track the changes occurring; and (c) it is not necessary to change the O, F or V values as only PCC values are required to be reported in the EA’s WRP tables. As a result, care is needed in interpreting the results presented in the “PCC & Ml/d forecasts” sheets in Micro-F:

30 rpsgroup.com

 The PCC and Ml/d volume values presented are after reconciliation and include changes for weather scenario and water efficiency  The O, F and V values are before reconciliation or changes for weather scenario or water efficiency  Text box messages are provided on the results sheets to highlight these points.

Table 6. Summary description of the model Analysis Data input Name of sheet Purpose of sheet stage cells?

Disclaimer No Conditions of use General Readme No Outline user information User chooses the MTP scenario and Choose Main Page Yes demand scenario for which results are scenario required Population, household and occupancy Demographics Yes data for each meter status type Normal year, dry year and climate change Data input Weather scenario factors Yes factors to apply to base year Effects of proposed water efficiency on Water efficiency Yes micro-components Toilet flushing Yes For each micro-component: Personal washing Yes  Specifies the O, F & V values for each Micro- Clothes washing Yes MTP scenario and each meter status component Dish washing Yes type analysis Miscellaneous internal use Yes  Tabulates PCCs and other data in External use Yes standard format for later analysis Adjust micro-component PCC values for each meter status type so that the total Base year PCC Reconciliation Yes PCC accords with SWW Company reconciliation Performance Review 2012 reporting for the base year Table of O, F, V, PCC of Ml/d volume Results table No values as selected in a picklist, and for any chosen meter status type(s) Charts of O, F, V, PCC of Ml/d volume Meter status charts No values as selected in a picklist, and for any PCC & Ml/d chosen meter status type(s) forecasts Charts of O, F, V, PCC of Ml/d volume Micro-component charts No values as selected in a picklist, and for any chosen micro-component Table of micro-component PCCs in WRP2 Table WRP2 No format for the demand scenario chosen on the Main Page Notes:  There are also several hidden sheets which store the intermediate data analysis. They have been hidden as the user does not need to refer to them when using Micro-F.  The filename of the forecaster should not be changed as it may affect the Visual Basic (VBA) code in Micro-F.

31 rpsgroup.com

5 PCC forecasts

This Section comments on the use of the micro-components analysis method for SWW and summarises the PCC forecasts.

5.1 Use of micro-components analysis

Micro-components analysis provides an effective method for incorporating explicit assumptions about how customer appliance ownership and use of water may change in the future. The Environment Agency and others consider this feature to be very important and so, as reported in Section 2.1, national guidelines recommend micro-components analysis as the preferred method for forecasting household consumption. However, it has some significant deficiencies, in particular:-

 There are errors in the base year micro-component assessments. This is due to there being a large number of O, F and V estimates being required, and the variety of data sources and quality of estimates. Many of the items are poorly understood. The uncertainties in each element are demonstrated, for example, by the discussion of data items in Section 3. As recognised by the UKWIR (2012) guidance, base-year reconciliation of PCCs for each meter status category is required. The extent of reconciliation adjustments required in this study for SWW are summarised in Table 7. The % adjustments are generally large, and illustrate the difficulties of using micro- components analysis (MCA) to assess PCC levels or to reflect the differences in consumption occurring in practice by different meter status categories.

 There are significant uncertainties around the forecast trends. There are a large number of assumptions required to derive trends for each O, F and V forecast. In most cases MTP forecast trends have been used. MTP have used professional judgement in assessing the trends – they do not know what the actual trends will be. Furthermore, MTP have produced three forecast scenarios, which as demonstrated in Section 5.2, result in a wide range of possible PCC projections.

The UKWIR (2012) national methodology recognises these deficiencies and therefore identifies merits in producing both trend-based and micro-component based PCC forecasts, and comparing them. Such comparisons can be used to check key assumptions and/or to undertake reconciliation adjustments to the forecasts.

32 rpsgroup.com

Table 7. Base year PCC reconciliation differences Sum of micro- Estimated actual (pre- component PCC MLE) PCC values at Difference (l/hd-d) Meter status category values at 2011/12 2011/12 (l/hd-d) (%) (l/hd-d) Based on CPR(*) Based on MCA 18.6 Unmeasured 137.0 155.6 (12.0%) -3.0 Existing metered 121.3 118.3 (-2.5%) -13.8 Future new connections 118.0 104.2 (-10.7%) -16.6 Future free optants 121.3 104.7 (-16.0%) 19.0 Future compulsory 114.8 133.8 (14.2%) (*) The base year reconciliations are described in Section 3.10.4, and the derivation of base year PCCs consistent with SWW’s Company Performance Review (CPR) 2012 is described in Appendix B.

5.2 Micro-component based PCC forecasts

A wide variety of table or graphical outputs are available from Micro-F for whichever water resource zone or scenario choice that is of interest, and so detailed consideration of the results is not provided in this report. The regional micro-component PCC forecasts for a normal year (without any water efficiency effect) are summarised below in the following formats:-

 Table 8 demonstrates that personal washing is the largest component of water use in the home.

 Table 9 shows that the model predicts that PCC will decrease in the future, due in particular to reductions in water use for toilet flushing, personal washing and clothes washing, as more water efficient appliances are expected to be used in the coming years (such as low-volume flush toilets, low water using washing machines, or use of showers instead of baths).

 Figure 3 shows that the model predicts that average PCC (for all households) will reduce under any of the three MTP forecast scenarios. However, there are significant differences in the rate of change between the MTP scenarios.

 Figure 4 presents similar graphs for individual meter status types.

33 rpsgroup.com

Table 8. Regional normal year micro-component PCCs (l/hd-d) for the MTP Policy scenario by meter status for 2011/12 after reconciliation adjustments Existing Future Unmeasured Future new Future free metered compulsory households connections optants households metered Toilet flushing 32.1 23.8 18.9 21.1 28.4 Personal washing 63.0 45.6 41.3 40.4 54.5 Clothes washing 14.9 13.8 12.2 12.2 11.3

Dish washing 6.1 6.7 5.9 5.9 5.4 Miscellaneous internal use 26.4 19.6 17.7 17.3 23.4 External use 15.3 11.0 10.0 9.7 13.1 Total PCC 157.9 120.5 106.1 106.6 136.2 Note that values in Tables 8 and 9 may not sum exactly due to rounding.

Table 9. Regional normal year micro-component PCCs for the MTP Policy scenario (l/hd-d) 2011/12 2019/20 2029/30 2039/40

METERED HOUSEHOLDS

Toilet flushing 23.8 19.7 16.9 16.5

Personal washing 45.6 42.4 40.1 37.1

Clothes washing 13.8 10.2 6.8 6.2

Dish washing 6.7 6.1 5.7 5.3

Miscellaneous internal use 19.6 19.9 21.3 22.7

External use 11.0 10.7 10.6 10.6

Total PCC 120.5 109.0 101.4 98.4

UNMEASURED HOUSEHOLDS

Toilet flushing 32.1 28.1 25.0 24.7

Personal washing 63.0 60.4 58.2 54.9

Clothes washing 14.9 11.8 8.8 8.4

Dish washing 6.1 5.7 5.4 5.2

Miscellaneous internal use 26.4 27.2 29.0 30.7

External use 15.3 15.3 15.3 15.3

Total PCC 157.9 148.6 141.7 139.3

34 rpsgroup.com

Figure 3. Comparison of MTP scenario forecasts of regional normal year average PCC (l/hd-d) (population weighted)

a) Unmeasured households b) Existing metered households

c) Future new connection households d) Future free optant households

Figure 4. Regional normal year average PCCs by MTP scenario and meter status type (l/hd-d)

35 rpsgroup.com

Figures 5 and 6 compare forecast PCC values with recent actual outturn PCCs for unmeasured and metered households. In the case of unmeasured households, recent actual PCC values have been volatile but suggest a general downward trend, whereas actual PCCs for metered households have reduced more significantly. The forecast values suggest that PCC levels will reduce further in the future.

Figure 5. Historic actual and forecast normal year PCCs for unmeasured households (l/hd-d)

Figure 6. Historic actual and forecast normal year PCCs for existing metered households (l/hd-d)

36 rpsgroup.com

Appendix A: MTP Reference, Policy and Earliest Best Practice (EBP) Scenarios

37 rpsgroup.com

Defra’s Market Transformation Programme (MTP) has published various documents that describe current and future ownership and usage of domestic appliances. The documents are listed in the main report and can be found at: http://efficient-products.defra.gov.uk. They provide projections for three scenarios:-

 The “Reference Scenario”: This is a projection of what is likely to happen without any new policy intervention. The scenario is based on current trends, technology developments and policies that are already in place.

 The “Policy Scenario”: This scenario estimates what could be achieved through an ambitious but feasible set of policy measures if the agreement of all stakeholders was obtained.

 The “Earliest Best Practice Scenario” (EBP): This is a projection of what could happen if the best available products and technologies were adopted, coupled with ambitious Government policies.

This Appendix presents the key data that has been obtained from the MTP documents for use in this project. Values with blue highlight have been derived directly from the MTP documents, and those without blue highlight have been estimated or extrapolated using MTP or other information.

A1. WCs (MTP ref H, pages 10-13)

OWNERSHIP FOR REFERENCE SCENARIO (%) Model 2010 2020 2030 2040 9 + 7.5 litre 33% 4% 0% 0%

6 litre 23% 9% 3% 0%

6/4 litre 44% 81% 80% 80% (av. 5 litre) 6/3 + 4.5 l 0% 6% 14% 14% (av. 4.5 l) 4/2.6 + <4.5 litre 0% 0% 3% 6% (av.3.3l)

OWNERSHIP FOR POLICY SCENARIO (%) Model 2010 2020 2030 2040 9 + 7.5 litre 33% 4% 0% 0%

6 litre 23% 9% 1% 1%

6/4 litre 44% 58% 22% 22% (av. 5 litre) 6/3 + 4.5 l 0% 23% 56% 56% (av. 4.5 l) 4/2.6 + <4.5 litre 0% 6% 21% 21% (av.3.3l)

38 rpsgroup.com

OWNERSHIP FOR EBP SCENARIO (%) Model 2010 2020 2030 2040 9 + 7.5 litre 33% 2% 0% 0%

6 litre 23% 6% 0% 0%

6/4 litre 44% 43% 9% 9% (av. 5 litre) 6/3 + 4.5 l 0% 25% 27% 27% (av. 4.5 l) 4/2.6 + <4.5 litre 0% 24% 64% 64% (av.3.3l)

FREQUENCY OF USE (use/person-d)

4.71 for all scenarios and all time

VOLUME PER USE (l/use) Scenario 2010 2020 2030 2040 Reference 5.06 4.88 4.79 - Policy 5.06 4.44 4.09 - EBP 5.06 3.69 3.26 -

These average volumes were estimated by MTP based on the assumed ownership of each model (see above) and assumed average flush volume for each model. The average volumes calculated for this study are higher than presented here. MTP do not provide explicit calculations of how these average volumes were obtained – they do not seem to be consistent with the MTP estimated stock levels for each toilet type (see above).

A2. Baths (MTP ref J, page 9)

OWNERSHIP (%)

MTP assume the same values for all scenarios: Scenario 2010 2020 2030 2040 Reference 94 90 83 75 Policy 94 90 83 75 EBP 94 90 83 75

FREQUENCY (use/prop-d) Scenario 2010 2020 2030 2040 Reference 0.68 0.63 0.58 0.53 Policy 0.68 0.55 0.46 0.37 EBP 0.68 0.55 0.46 0.37

39 rpsgroup.com

VOLUME PER USE (l/use) Scenario 2010 2020 2030 2040 Reference 84.5 85.0 85.0 85.0 Policy 84.5 80.6 80.6 80.6 EBP 84.5 77.2 76.3 75.4

A3. Showers (MTP ref I, page 10, and ref M, pages 8-9)

OWNERSHIP (%)

MTP identify three main types of shower, and assume the same ownership projections for all scenarios:-

Shower type 2010 2020 2030 2040 Electric 39% 46% 48% 48%

Gravity mixer 22% 24% 25% 25% Pumped mixer 20% 22% 21% 21% Total 81% 92% 94% 94%

FREQUENCY OF USE (use/prop-d) 2010 2020 2030 2040 Reference 1.04 1.16 1.21 1.21 Policy 1.04 1.21 1.33 1.33 EBP 1.04 1.21 1.33 1.33

VOLUME PER USE (l/use)

Electric shower 2010 2020 2030 2040 Reference 26.5 27.9 27.9 27.9 Policy 26.5 26.7 26.6 26.6 EBP 26.5 23.4 22.4 22.4

Gravity mixer shower 2010 2020 2030 2040 Reference 38.2 38.8 38.4 38.4 Policy 38.2 38.5 37.8 37.8 EBP 38.2 30.1 28.9 28.9

The values for 2010 to 2030 have been estimated based on:-

 MTP reported average volume per use of 38.2 litres in 2010 (ref M)  MTP trend of future average volume for all mixer showers (ref I)  The values for 2040 have been assumed for this study to be the same as at 2030

Pumped mixer shower 2010 2020 2030 2040 Reference 62.7 63.7 63.0 63.0 Policy 62.7 63.2 62.1 62.1 EBP 62.7 49.4 47.5 47.5

The values for 2010 to 2030 have been estimated based on:-

40 rpsgroup.com

 MTP reported average volume per use of 62.7 litres in 2010 (ref M)  MTP trend of future average volume for all mixer showers (ref I)  The values for 2040 have been assumed for this study to be the same as at 2030

A4. Clothes washing (MTP ref F pages 4 & 12, and ref G page 3)

OWNERSHIP (%)

95% for all scenarios and all time

FREQUENCY OF USE

260 uses/prop-year (i.e. 0.71 uses/prop-day) for all scenarios and all time

VOLUME PER USE (l/use)

MTP do not provide estimates of water use by washing machines. The table of estimated values below has been based on:

 UKWIR (2012) (ref O) estimate that current average volume per use is 60 litres  Assumption that for the Policy scenario, the average volume will move to the current best volume (31 litres according to Which?) by 2030  The profiles in MTP Ref F of how energy consumption will vary between scenarios and through time to 2030

Scenario 2010 2020 2030 2040 Reference 60 48 33 33 Policy 60 45 31 31 EBP 60 35 24 24

A5. Dishwashers (ref D, pages 3, 10, and ref E page 3)

OWNERSHIP (%)

MTP report that the ownership of dishwashers is expected to increase between 2010 and 2030. The data presented by MTP is summarised below. They assume the same ownership forecasts for all scenarios.

2010 2020 2030 2040 Dishwasher ownership (mill.) 9.65 11.78 12.81 No. households in UK (million) 26.59 29.44 32.11 % ownership 36% 40% 40% 40%

FREQUENCY OF USE

The frequency of use has been calculated from the MTP projections of average annual usage between 2010 and 2030, as shown below. MTP assume the same frequency of use for all scenarios

2010 2020 2030 2040 Dishwasher use 245 242 236 per year Dishwasher use 0.67 0.66 0.65 0.64 per day

41 rpsgroup.com

VOLUME PER USE (l/use)

MTP do not provide estimates of water use by dishwashers. The table of estimated values below has been based on:

 UKWIR (2012) (ref O) estimate that current average volume per use is 20 litres  MTP (ref E) predictions for the Policy Scenario that average energy use per wash will reduce from 1.22 kWh/use in 2010 to 1.00 kWh/use at 2020, and to 0.87 kWh/use at 2030  The profiles in Ref E of how energy consumption will vary between scenarios and through time to 2030  The 2040 values have been estimated by extrapolation

2010 2020 2030 2040 Reference 20 17.7 17.2 16.7 Policy 20 16.4 14.3 12.2 EBP 20 14.8 11.4 8.0

A6. Internal tap use (ref MTP K, page 9)

FREQUENCY OF USE (use/person-d)

MTP (K) presents the following estimates of internal tap use, which are assumed to be the same for all scenarios.

2010 2020 2030 2040 Washbasin 8 8 8 8 Kitchen 9 9 9 9

VOLUME PER USE (l/use)

MTP (refs K and M) present estimates of the average volume per use of internal taps. The volumes vary between scenarios as they assume different rates of take-up of low-flow taps or inserts. The MTP documents are inconsistent in the values reported for 2010 average volume per use – this study has used the Policy scenario quoted values of 2.27 litres for washbasins and 2.28 litres for kitchen taps.

Reference scenario 2010 2020 2030 2040 Washbasin 2.27 2.27 2.27 2.27 Kitchen 2.28 2.28 2.28 2.28

Policy scenario 2010 2020 2030 2040 Washbasin 2.27 2.13 2.13 2.13 Kitchen 2.28 2.11 2.03 1.95

EBP scenario 2010 2020 2030 2040 Washbasin 2.27 2.00 2.00 2.00 Kitchen 2.28 2.04 1.89 1.74

42 rpsgroup.com

TAP USE (l/hd-d)

The values presented above have been used to calculate the following internal tap use volumes.

Reference scenario 2010 2020 2030 2040 Washbasin 18.2 18.2 18.2 18.2 Kitchen 20.5 20.5 20.5 20.5 Total (l/hd-d) 38.7 38.7 38.7 38.7

Policy scenario 2010 2020 2030 2040 Washbasin 18.2 17.0 17.0 17.0 Kitchen 20.5 19.0 18.3 17.6 Total (l/hd-d) 38.7 36.0 35.3 34.6

EBP scenario 2010 2020 2030 2040 Washbasin 18.2 16.0 16.0 16.0 Kitchen 20.5 18.4 17.0 15.6 Total (l/hd-d) 38.7 34.4 33.0 31.6

43 rpsgroup.com

Appendix B: Derivation of other data

44 rpsgroup.com

This Appendix describes the derivation of other data (i.e. other than the O, F and V values described in section 3) that have been input into the Micro-F micro-component PCC forecasting model.

B1. Base year household, population and occupancy data

Estimates and forecasts (2011/12 to 2039/40) of the number of households and population have been extracted from a data-file provided by SWW on 14th May 2012: PCC & property numbers.xlsx

The data file provides details for each meter status type, based on SWW’s current regional households and population forecasts. However, as currently no compulsory metering programme is planned the forecast numbers are zero for that meter status type. This study has produced forecast PCCs for compulsory metered households assuming that the average occupancy in each year is the same as for unmeasured households.

B2. Base year PCC values

Estimates of base year (2011/12) PCCs are required for each meter status category as data input to Micro-F so that the raw PCC values derived by the micro-component analysis can be reconciled to be consistent with SWW reported PCCs. The values used are presented in Table B1. They are consistent with SWW’s Company Performance Review 2012 before maximum likelihood estimation (MLE), and include estimates of plumbing losses.

Table B1. Base year (2010/11) PCC values Meter status category Base year Derivation PCC(l/hd-d) Unmeasured Households 156.609

Existing Metered Households 118.320 Spreadsheet 2012-08-07 – PCCs.xls from Paul Merchant presenting PCC Future New Connection 104.214 values for each meter status category Households inclusive of total average plumbing losses of 14.852 l/prop-d Future Free Optant Households 104.684

Future Compulsory Households 133.824 Calculated as 14% less than the unmeasured household PCC (see B5 below)

B3. Normal year and dry year factors

SWW’s Final Water Resources Management Plan (November 2009) assessed the effect of dry weather on average PCC, as reported in Section 1.3. It identified a normal to dry year PCC conversion factor of 1.072 for unmeasured households and 1.053 for metered households.

SWW have reanalysed the normal year to dry year factors in preparation for their 2014 Water Resources Management Plan, and have examined the characteristics of the base year (2011/12) weather. The factors for 2011/12 are summarised in Table B2, as provided in an email sent by Paul Merchant in dated 24th August 2012.

45 rpsgroup.com

Table B2. Normal year and dry year factors derived by SWW for use in Micro-F

Measured households Unmeasured households

Base year to normal year 1.018 1.015 factor

Base year to dry year factor 1.079 1.081

The normal year and dry year factors have been applied uniformly to all components of demand as there in the absence of clear evidence of how the demand for individual components responds to dry weather. In practice it is likely that the effect would be largest for external water use (in particular garden watering). The dry year factor for each component is entered at data input cells, and so can be easily modified if desired.

B4. Climate change effects

The effects of climate change on PCC have been estimated using the Defra (2003) report “Climate Change and the Demand for Water” (ref T). The values assumed for this study were taken from Table 3-14 of the report, which relates to the Penzance area. The % impact values for Medium-high climate change scenario and the EA’s Beta scenario for 2025 and 2055 have been averaged to provide the following estimates for 2040:-

 Garden watering: 3.0%  Baths and showers: 3.6%

It is anticipated that updated values will be available later in 2012 from the current UKWIR project “Impact of Climate Change on Demand”.

B5. Metering effects

Section 3.3 of SWW’s Final WRMP (November 2009) reported on studies to estimate the effect of optional metering on household water demand, and identified an estimated 14% impact. SWW requested that this value be used for the current study.

The following approach has currently been used in Micro-F to specify the effect of metering on individual components in the “Base-case regional micro-component analysis”:-

 The effect of metering has been set at 14% for the following components of PCC: toilet flushing, personal washing, hand-washing of dishes, miscellaneous internal use and external use. (The 14% value is a data input and so can be easily changed if desired).

 For clothes washing and dishwashers, the difference in use between unmeasured and metered households is assumed to be represented by the different frequencies of use recorded in the MORI survey (see Table 5).

As explained in Section 3.10 the micro-component PCC values are adjusted for each meter status category during the “Base-year PCC reconciliation”, and so the relative metering effects are altered by the reconciliation.

46 rpsgroup.com

B6. Water efficiency effects

No estimates of the effect of SWW’s water efficiency programme on future micro-component PCC’s are currently available. At present, dummy values have been inserted in Micro-F in the sheet “Water efficiency”. Specific values, expressed as l/hd-d effects for each PCC component, can be inserted by SWW when known.

The values in sheet “Water efficiency” are not applied to the PCC forecasts unless the tick- box(es) are ticked.

47 rpsgroup.com

Appendix E Demographic forecasts

SWW Water Resources Management Plan 2015 – 2040

Population, Household and Dwelling Forecasts for WRMP14: Phase 2 Draft Final Report South West Water June 2013

Population, Household and Dwelling Forecasts for WRMP14: Phase 2 Draft Final Report South West Water June 2013

Contents

______Introduction ...... 1 1 Approach ...... 2 1.1 Outputs ...... 2 1.2 Data sources ...... 2 1.3 Methodology ...... 2 2 Deriving a most-likely forecast ...... 8 2.1 Background ...... 8 2.2 Population trends ...... 8 2.3 Trends in household occupancy...... 10 2.4 Dwelling completions ...... 10 2.5 Most-likely forecast - approach ...... 13 3 Phase 2 Results ...... 14 4 Uncertainty analysis ...... 18 4.1 Background ...... 18 4.2 Stochastic analysis ...... 18 4.3 Scenario analysis ...... 18 Appendix A: Data Sources ...... 19 Appendix B: How property pipeline level data is built into the demographic forecasts ...... 21 Appendix C ...... 22

______

This output is based on and comprises both your input and information sourced from third parties (which may include public data sources). Whilst we will use all reasonable care and skill in the collection and collation of this output we cannot warrant or guarantee the accuracy of the output. You acknowledge that outputs which use empirical data and/or statistical data and/or data modelling techniques cannot be taken as a guarantee of any particular outcome and are not intended to be the sole basis of your business decisions. Our standard terms of business apply.

Introduction

This report updates the methodology used to produce a range of population and household projections for a group of water companies for WRMP14. This report includes the results from Phase 2 of the project which provides an update of the projections provided to companies in 2012 to include the latest available information, including the Census 2011. This report updates the relevant sections of the approach, results and the most-likely forecast to reflect the changes between Phase 1 and Phase 2.

The projections for Phase 1 and Phase 2 have been produced in accordance with Water Resource Planning Guideline (Joint Regulator, 2012) and the Method of Estimating Population and Household Projections (EA, 2012) report. The companies included in the study include:

 Southern Water  Thames Water  Wessex Water  Sembcorp Bournemouth Water  Portsmouth Water  South East Water  Sutton & East Surrey Water  Affinity Water (Central, East and South East)  Welsh Water  South West Water (joined after the other members of the group)

Three sets of forecasts have been provided each year in the period 2010/11 to 2039/40 for:

 Total Population;  Household population;  Communal population;  Households;  Household Occupancy;  Dwellings

The three sets of forecasts are:

 Plan-based (using information provided by local authorities)  Trend-based (using the latest information from official statistics)  Most Likely (Experian’s best view on likely outcomes based on information available).

The first section of the report presents the methodology used to produce the projections whilst section 2, explains the rationale and approach for producing the most-likely forecast.

Section 3 presents the results of the forecasts for Phase 2 and compares them with Phase 1 for each company. Section 4 provides an explanation of estimates of uncertainty associated with the forecasts for each company.

1

1 Approach

1.1 Outputs

Three sets of forecasts have been produced:

 Plan-based (using information provided by local authorities)  Trend-based (using the latest information from official statistics)  Most Likely (Experian’s best view on likely outcomes based on information available).

The forecasts have been produced at Census output area1 and provided to each company at this detailed level and aggregated to entire (water and sewerage) supply area and Water Resource Zone (WRZ).

1.2 Data sources

A wide array of data has been used to produce the Phase 2 forecasts. The key data inputs for Phase 1 and Phase 2 and their vintage are detailed in the table below. Further details of these sources are provided in Appendix A.

Table 1: Summary of data sources for Phase 1 and Phase 2 Source Phase 1 Phase 2 Local authority provided planned dwelling data and local authority Local authority provided planned dwelling data and local Local Plans plans authority plans ONS 2011-interim sub-national population projections - Local LA population projections ONS 2010 sub-national population projections – Local Authority Authority ONS 2010 National population projections, principal and variant National projections projections Household projections DCLG 2008 household projections – Local Authority DCLG 2011-interim household projections

ONS indicative mid-year estimates 2006-2010 – Local Authority Mid-year population estimates ONS revised mid-year estimates 2002-2010 - Local authority ONS mid-year estimates 2001-2005 – Local Authority

ONS mid-year estimates 2001-2010 – Lower Super Output Area Census 2011 Small Area estimates Census 2001 Experian Output Area level datasets, 2001-2040 Experian Output Area level datasets, 2011 to 2040 Property Pipeline information supplied by Emap Glenigan (April Property Pipeline information supplied by Emap Glenigan Property pipeline information 2012). (April 2012).

1.3 Methodology

1.3.1 Collecting information for the plan-based projections

The first task was to update the information collected from each of the local authorities from Phase 1 that are covered by the company boundaries (water and sewerage) of the companies involved in this study. This involved confirming with local authorities whether the information provided for phase 1 was still up to date and relevant. For those local authorities that did not respond to the data request in phase 1 a data collection template was sent. Emails were sent to local authorities over a two day period from 12th to 13th March with a request for response by 8th April. Follow up emails were sent in the following days and weeks depending on the responses received.

The contacts list was generated from a combination of water company contacts, Experian contacts and contacts provided by DCLG.

1 Experian have maintained the use of 2001 Census Output Areas. Further information is contained elsewhere in this document.

2

Information for London was taken from the London plan:

The London Plan Spatial Development Strategy for Greater London, July 2011 http://www.london.gov.uk/priorities/planning/londonplan

Some key points regarding the data collection exercise:

 Where authorities provided data for Phase 1 the relevant contacts were asked to update/ verify the data for Phase 2  E-mails were targeted to individuals within the local authority where we had named contacts  E-mails were tailored to each local authority (each e-mail was sent individually) so it was clear which water company we were collecting the information from (particularly important as some local authorities have 3 or more water companies operating in their area).  Collaborative approach was extremely helpful to local authorities as the potential burden was greatly reduced.  The data collection exercise will be re-run in full for Phase 2. We expect that data collection rates will be higher for phase 2, as we have developed a relationship with local authorities and more local authorities will have completed their local plans.  A full log of contact for all local authorities has been produced and provided to the companies.

The table below shows the response rate achieved for each water company for Phase 1 and Phase 2. The response rates for South West Water improved from 42% for Phase 1 to 67% for Phase 2.

Table 2: Local authority response rates from Phase 1 and Phase 2 by company % of LAUAs responded Water company Phase 1 Phase 2 Sembcorp Bournemouth Water 63% 63% Portsmouth Water 67% 67% South East Water 54% 59% Southern Water 54% 74% Sutton and East Surrey Water 36% 73% Thames Water 54% 63% Affinity Water Central 48% 70% Affinity Water East 67% 67% Affinity Water South East 40% 40% Wessex Water 47% 65% South West Water 42% 67%

Where information was not supplied by the local authority directly, it was collected from alternative sources. A hierarchical system was used, with the most recent sources given preference if contact with an authority was not established:

1. Directly from each local authority 2. Directly from County Councils 3. From Local Authority Plans, Core Strategies, Local Development Frameworks or Annual Monitoring Plans – depending on availability and date of publication.

3

1.3.2 District and household level forecasts

In accordance with EA guidance, the starting point for our output area (OA) level population and household projections is to create a set of district level population and household targets, which are used as control totals for the subsequent OA level work.

1.3.2.1 Trend-based projections

The first set of household projections are trend based: they are neither a forecast of what analysts expect to happen nor a statement of policy. The Phase 2 trend based local authority district level population and household targets are based on the ONS 2011-based interim sub-national population projections and the 2011-based interim DCLG household projections. The ONS and DCLG projections only extend to 2021 – they have been extended to 2040 using a simple extrapolation of the last 5 years of the projection.

The DCLG 2011 household projections and ONS 2011-based interim sub-national population projections include results from the Census 2011.

The Phase 1 projections were based on the 2010-based sub-national population projections and the 2008-based DCLG household projections, neither of which included results from the Census 2011.

Trend-based projections are a key input to producing plan-based projections.

Further analysis of the trend based population projections is included in section 3.2.

1.3.2.2 Local authority plan-based projections

WRPG states that water companies should take account of local authority plans in their population and household projections. To account for planned future developments, local authority plan household and population projections are constructed. These take information from the local authority data collection exercise discussed in 1.3.1. Here annual dwelling figures from each of the plans from 2011 onwards are converted to households and added on to the base year to produce a plan-based household forecast.

Estimates of district level plan based population are recalculated by applying projections of average household size from the trend-based projections to the plan based household projections, above.

The local authority plans cover different periods of time – typically they only extend as far as 2025 but most are shorter. Once the plans finish there is a decision on the likely trajectory of the plan-based projections. Given the wide range of plans (and different statuses of plans) covered by this study, we have applied trend based assumptions to extend the plan-based forecasts. Here we apply the growth in household numbers from the trend- based forecasts to the number of dwellings. This is an option presented in the Environment Agency methodology report.

The Greater London Authority (GLA) produced population and household projections for London boroughs as part of the evidence base for the London Plan. A GLA controlled set of plan-based projections has also been produced as part of this project.

1.3.2.3 Most-likely projections

As part of the project specification there is a requirement to produce a most-likely forecast of population and households – which is what Experian think will be the most likely outcome given data available and our expertise.

4

Experian have revised the methodology used for creating the most-likely projections for Phase 2 in light of recent evidence and data availability. The approach selects the most appropriate population projection for each local authority based on analysis of recent trends. The most-likely household forecast is based on first controlling the plan based household projection to Experian’s regional dwelling completions forecast and a second adjustment to account for underlying long-term growth. The approach for the most likely forecast is detailed in section 2.

1.3.3 OA population and household targets

The next task is to drill down below the district level targets to a more refined geographic area. Experian have used Census OAs (e.g. 33UGFY0003) for the analysis of small spatial areas. This is so that information from Census 2011, the key source of data for small area demographics, can be used. Moreover, it facilitates the incorporation of new property developments that are easily coded at OA level. Output areas figures can then be aggregated to the geographical levels required.

Experian have produced the OA projections using the Census 2001 output area boundaries. ONS have made small changes to the output areas for the 2011 Census but have provided a mapping between the 2001 and 2011 boundaries. Experian have used the 2001 boundaries to maintain consistency with Phase 1 and to incorporate Experian’s output area forecasts which are currently based on 2001 boundaries. However note that the data includes the Census 2011 results for households, dwellings, household and communal population.

The various stages taken to construct the OA population and household projections are set out below:

1. Age forwards Census 2001 OA residents in households using a cohort survival approach (e.g. the number of 20_24 year olds this year is based on 4/5 times the number of 20_24 year olds in the previous year (i.e. 1/5 move up to the next age group) plus 1/5 times the number of 16_19 year olds the previous year (i.e. 1/5 move up to the 20_24 year olds from the 16-19 age group).

2. Births are estimated by applying district level fertility rates to its constituent OA level population of females aged 15_44. Death and migration rates at OA level are also estimated by applying district level rates.

3. Control the aged forwards OA figures from step 1 to Census 2011 values.

4. Source OA level counts of communal population from Census 2011. The counts are controlled to district level targets post 2011.

5. Calculate household population by subtracting communal population from total population.

6. Estimates of the number of households in each OA are taken from Census 2011 and pushed forward by combining the growth in OA household population (from Stage 5) with changes to average household size in its encompassing district.

7. Calibrate the OA household estimates to align with district level household targets for the trend-based, plan and most-likely. At this stage we have an initial set of household and population projections, 2011-2040, by OA. The methodology that we use to build residential property pipeline information into our demographic forecasts utilizes site level planning application and contract progress data that is sourced from Emap Glenigan. To utilize Emap Glenigans site level planning application and contract progress data in our demographic forecasts we first need to establish the likelihood that each site in the property pipeline has of being “built-out”. To do this we use a procedure (developed in consultation) with Emap Glenigans that assigns “build out” probabilities according to the stage that each site has reached in the planning /contracting process and the insight (based on experience) that this information provides regarding the likelihood that the associated scheme will be completed (for more details, see appendix C).

8. All projects are assumed to start and be completed between 2011 and 2023. All developments are aggregated to OA level. This adjustment adds additional local flavour to the household projections by accounting for possible new developments.

9. Overall constraining procedures are applied to the OA household and population projections to ensure that they are consistent with our broader view of population and household projections at the district level for the trend-

5

based, plan-based and most-likely approach. Not every output area will have housing projects sourced from Emap Glenigans. As a consequence of the LAUAD constraint, those OAs without a new development will see a reduction in housing growth to balance developments elsewhere in the district for the LAUAD target to be achieved.

1.3.4 Bespoke spatial analysis

Experian’s household and population calculations for each of the WRZs areas were carried out using Alteryx and Micromarketer, two spatial analysis programmes. The methodology follows the Environment Agency guidance and ONS postcode best fit approach to producing small area estimates.

Three inputs are fed into the calculations:

 Client supplied WRZ GIS boundaries  Output Area (OA) boundaries  Current year population and area (in sq km) for each OA and postcode

The Alteryx programme first identifies which OAs are located entirely within each boundary of a given WRZ. The sum of the total population of all of these OAs can then be derived and will account for the majority of each WRZs total population.

This leaves only areas around the borders of the WRZs for examination, areas which will not contain any complete OAs but will be made up typically of elements of a number OAs (the remainder of the OA falling into another WRZ or falling outside each water companies total area). For each of these OAs we calculate the proportion of cut OA population that is inside each WRZ as a proportion of the full OA population using Census postcode area level data. These rates are kept fixed in the forecast.

The proportions are then applied to the population and households of these OAs to give the population falling inside the given WRZs. For each WRZ these population shares can then be aggregated, and combined with the population calculated from the ‘whole’ OAs we reach a final figure for the WRZs total population.

An example of the Alteryx output is shown below for a small WRZ area in the East of England. The total population for this area is comprised of the sum of the seven OAs that fall entirely within the area boundaries plus the shares of an additional twelve OAs where the area boundary splits the OA boundary. Note that where the Output Area splits the area that the share values can range between 0% and 100%. Where the share is 0% the OA cut population is zero however some of the OA area falls within the area boundary. Where the share is 100%, the OA cut population equals the full population but not necessarily all of the OA area falls within the DMA boundary.

Table 3: Best-fit example Output Area falls entirely within WRZ WRZ OutputArea CutArea CutPop FullArea FullPop Share Example X 00KFNA0015 0.016 276 0.016 276 100% Example X 00KFNA0021 0.084 334 0.084 334 100% Example X 00KFNA0023 0.017 298 0.017 298 100% Example X 00KFNA0028 0.028 272 0.028 272 100% Example X 00KFNA0030 0.020 305 0.020 305 100% Example X 00KFNA0031 0.043 254 0.043 254 100% Example X 00KFNA0032 0.022 186 0.022 186 100%

Output Area splits WRZ WRZ OutputArea CutArea CutPop FullArea FullPop Share Example X 00KFNA0006 0.313 266 0.316 266 100% Example X 00KFNA0007 0.033 284 0.092 318 89% Example X 00KFNA0012 0.008 0 0.031 213 0%

6

Example X 00KFNA0014 0.018 294 0.020 294 100% Example X 00KFNA0017 0.018 293 0.028 293 100% Example X 00KFNA0018 0.009 0 0.028 264 0% Example X 00KFNA0022 0.020 300 0.023 300 100% Example X 00KFNA0026 0.016 98 0.059 323 30% Example X 00KFNA0029 0.030 85 0.044 250 34% Example X 00KFNG0009 0.135 62 0.253 304 20% Example X 22ULGD0005 0.045 47 0.273 327 14% Example X 22ULGD0006 0.117 0 0.425 331 0%

7

2 Deriving a most-likely forecast

2.1 Background

The most-likely forecast is what we think is the most likely outcome for population and households based on our expertise and the latest information available. It was decided to review the methodology used to derive the most- likely forecast for Phase 1 in light of recent evidence of growth. This section presents the drivers, rationale and approach for the Phase 2 most likely forecast. The following drivers are considered:

1. Population trends 2. Trends in household occupancy 3. Dwelling completions

2.2 Population trends

For Phase 1 we found little evidence of population coming off trend and therefore decided that the ONS trend- based projections would be the most likely outcome for population growth at local authority level. The most-likely population projections were therefore the same as the trend-based projections for Phase 1. The release of the Census 2011 confirmed that ONS was underestimating population growth, with almost 500,000 additional people found in 2011 compared with previous estimates for the same year in England and Wales. However, this effect was not uniform across local authorities as shown in the table below. For example, the Census found that population in City of London population was 51% lower than previously estimated and the population of Isles of Scilly was 14.7% higher than previously estimated. From a regional perspective, the Census 2011 found that the population in London was 1.3% higher than estimated; in the South East it was 1% higher and 0.6% higher in the East of England. In Wales the population was 1.2% higher than previously estimated whilst in the South West, the Census 2011 only 0.1% higher than previously estimated.

Table 4: Top 10 and bottom 10 differences between mid-year population estimates (MYE) and the Census 2011 based MYE

Difference Difference (MYE - % difference (% (MYE - % difference (% of Census Bottom of Census MYE Census Top 10 Area Census MYE 2011) based MYE) 10 Area 2011) based MYE) 1 City of London 51% 3751 1 Isles of Scilly UA -14.7% -327 2 Westminster 12% 26877 2 Cambridge -13.2% -16237 3 Camden 8% 18626 3 Brent -11.1% -34783 4 Kingston upon Thames 8% 13476 4 Newham -9.3% -28794 5 Tendring 7% 10218 5 Waltham Forest -8.6% -22266 6 Runnymede 7% 5664 6 Watford -8.0% -7255 7 Welwyn Hatfield 7% 7582 7 Hackney -7.9% -19616 8 Oadby and Wigston 6% 3622 8 Leicester UA -7.8% -25639 9 Wokingham UA 6% 9018 9 Greenwich -7.7% -19735 10 Merton 6% 11421 10 Bournemouth UA -7.4% -13667

Since the release of the 2011 Census, ONS has published updated population projections which have been used to produce the Phase 2 trend-based projections. However ONS have created the 2011-based sub-national population projections by applying assumptions from the 2010-based projections to the 2011 Census results. The assumptions for the 2010-based projections were based on trends taken from the mid-year estimates prior to the release of the Census 2011. This approach has implications in some areas – particularly in areas where the Census 2011 results are significantly different to the previous mid-year estimates.

8

Figure 1a below demonstrates the issue for Camden in London, where the Census 2011 found that the population was 8.5% lower than previously estimated. The revised mid-year estimates show a shallower growth profile for the 5 year period that is used to inform trend-based growth projections. Figure 1b shows that when the assumptions from the 2010-based projections are applied to the 2011 Census point for Camden that growth is stronger than under the 2010-based projections and the growth profile compared with the mid-year estimates looks too strong. The opposite effect also occurs in areas where ONS underestimated the population and resulting projections will typically look too weak.

Figure 1a and 1b: Revised mid-year population estimates and the interim population projections, Camden

245000 270000

240000 260000 235000 250000 230000 225000 240000 2010-based ONS 220000 230000 projections 215000 220000 2011-based ONS 210000 projections 205000 210000 200000 200000 195000 190000 2006 2007 2008 2009 2010 2011 Pre-Census mid-year estimates Revised mid-year estimates

For the reasons outlined above Experian has selected the most likely projection for each local authority from the following set of projections:

 2010-based sub-national projections (controlled to Census 2011)  2011-based interim sub-national projection  Extrapolation of 2002-2011 revised mid-year estimates.

An example of these alternative projections is shown in figure 2 below for Kingston-upon-Thames. For this example, we have chosen the MYE extrapolated projection for the most-likely population forecast, since the 2011- based and 2010-based projections appear too strong given the trend between 2001 and 2011.

Figure 2: Alternative trend-based projections for Kingston-upon-Thames 300000

250000 2011-based 200000 projections

150000 2010-based controlled

100000 MYE Extrapolated

50000

0

Further analysis of population growth is available in the Phase 1 report.

9

2.3 Trends in household occupancy

The results from the Census 2011 showed that household occupancy did not decline as rapidly as previously projected by previous official household projections. Results from the Census 2011 showed that in some areas – most notably in London – occupancy actually increased between 2001 and 2011. Figure 3 shows that the latest DCLG household projections capture the slowed decline in occupancy rates – although as these are long-term projections they do not capture short-term deviations to trend that may occur as a result of economic and policy changes.

Figure 3: Comparison between occupancy (average household size) projections from the 2008 and 2011 DCLG household projections

For Phase 1 our analysis looked at estimated changes in occupancy and under the most-likely forecast we derived adjustments to trend-based occupancy forecasts to account for the economic slowdown and drive our household forecast. In section 2.4 we find that the plan-based forecasts are much closer to our expectations of future growth than under Phase 1 and therefore are a good basis for use for the most-likely forecast rather than modelling occupancy directly. Changes to occupancy will be derived under the most-likely forecast when household population is divided by the number of households.

2.4 Dwelling completions

As a result of the weak economic conditions, dwelling completions have slowed at national level and these trends have also been prevalent in the South West Water company area as shown in figure 4 below. The number of dwelling completions per annum has fallen during the recessionary period – with an average of around 5,000 dwellings completed per annum in 2009/10 to 2011/12 compared with an average of 7,000 between 2004/5 to 2008/9.

10

Fig 4: Net additions to dwelling stock by water company area South West Water 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000 - 2004-05 2005-06 2006-07 2007-08 2008-09 2009-10 2010-11 2011-12

The plan-based forecasts produced for this project take dwelling targets from each of the local authority local plans.

For the targets in the plans to be achieved there will need to be an improvement in the volume of house building – which we expect to begin to come on stream in 2014. Figure 5 shows the annual dwelling targets for the authorities which cover the South West water area – figure 6 represents these targets relative to the levels delivered on average between 2004 and 2008 and 2009-2012. For South West Water the plans require annual build of around 7,400 dwellings per annum which is higher (120%) than recent trends but lower than levels achieved between 2004 and 2012 (90%). The plan-based target then appears realistic but is still challenging given market conditions.

Fig 5: Annual dwelling targets 2013-2018 10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 2011 2012 2013 2014 2015 2016 2017 2018

11

Fig 6: Future annual dwelling targets % relative to previous period average annual housing delivery 140%

120%

100%

80% Plan v 2004-2008 60% Plan v 2009-2012

40%

20%

0% South West Water

Experian’s dwelling completion forecasts by region suggest that the plan based forecasts will not be achieved in the short-run as the market remains weak as shown in figure 7.

Fig 7: Future planned dwellings and Experian dwelling completions forecasts by region

30000 35000

25000 30000 25000 20000 20000 15000 15000 10000 10000

5000 5000

0 0 2013 2014 2015 2016 2017 2018 2013 2014 2015 2016 2017 2018 Experian East of England Plan East of England Experian London Plan London

40000 30000

35000 25000 30000 20000 25000

20000 15000 15000 10000 10000 5000 5000

0 0 2013 2014 2015 2016 2017 2018 2013 2014 2015 2016 2017 2018 Experian South East Plan South East Experian South West Plan South West

12

2.5 Most-likely forecast - approach

The most-likely forecast has been created using a three stage process:

1. Select the most-likely trend-based population projection 2. Control the plan-based household forecasts for each local authority to Experian’s regional household completions forecast 3. Adjust the controlled forecast to the difference between the plan and trend based projection in the medium to long term.

The first step ensures that the number of households forecast is in line with our forecasts of new dwellings produced by Experian’s construction futures team. This forecast considers economic conditions and other factors facing house builders over the short to medium term before assuming a trend.

The second step ensures that the most-likely forecast considers not only what local authorities are planning for but also underlying trends that may be above or below what is being planned for in the medium to long-run. Most local authority plans do not cover the entire WRMP period and many assume slower growth in the long-term, whilst at the same time population trends suggest many more houses will need to be built than are currently planned for. The most-likely therefore seeks to find a compromise between the two in the long-run.

13

3 Phase 2 Results

The Phase 2 results have been provided to each of the companies via Experian’s FTP site. Results have been provided at output area level and aggregated to water resource zone level for each company taking part in the study. In this section we present the results and compare the plan, trend and most likely forecasts. We also explore how the forecasts differ to those provided for Phase 1. Analysis of the projections used for PR09 and the impact of the economic downturn is available in the Phase 1 report.

3.1.1 South West Water population projections

At company level, the plan-based population projections are stronger than the trend-based and most-likely projections. This is reflected in Roadford and Wimbleball. Generally speaking, the most-likely projections tend to be weaker relative to the other projections, particularly towards the end of the forecast period.

Fig 8: WRZ population projections under trend, plan based and most-likely forecasts

700000 1100000

650000 1000000

600000 900000

Trend based - Colliford Trend based - Roadford 550000 800000 Most likely - Colliford Most likely - Roadford 500000 Plan based - Colliford 700000 Plan based - Roadford

450000 600000

400000 500000

2032 2014 2017 2020 2023 2026 2029 2035 2038 2011 440000

420000

400000 Trend based - 380000 Wimbleball Most likely - Wimbleball 360000

340000 Plan based - Wimbleball

320000

300000

2023 2013 2015 2017 2019 2021 2025 2027 2029 2031 2033 2035 2037 2039 2011

3.1.2 South West Water household projections

In line with the population projections, plan-based household projections are higher for the South West Water area than the most-likely and trend-based projections. At WRZ level the plan-based forecast tends to be higher than the other forecasts – this is evident for Roadford and Wimbleball. Additionally, for these areas the most-likely projection overtakes the trend-based in the medium term and remains above the trend for the remainder of the projection period. However, Colliford posts a stronger trend-based population and household projection relative to the other projections.

14

Fig 9: WRZ household projections under, plan based and most-likely forecasts

300000 490000 290000 470000 280000 270000 450000

260000 430000 Trend based - Colliford Trend based - Roadford 250000 Most likely - Colliford 410000 Most likely - Roadford 240000 Plan based - Colliford Plan based - Roadford 230000 390000 220000 370000 210000

200000 350000

2013 2015 2017 2019 2021 2023 2025 2027 2029 2031 2033 2035 2037 2039 2011 200000

190000

180000 Trend based - Wimbleball 170000 Most likely - Wimbleball

160000 Plan based - Wimbleball

150000

140000

3.1.3 Comparison with Phase 1 projections

Figure 10 shows the change in occupancy projections between Phase 1 and Phase 2. The chart shows that at over 2.26, occupancy was slightly higher in the Census 2011 than 2.23 people per household estimated in Phase 1. The trend-based projection is for a slower decline in household occupancy between 2011 and 2040. The trend- based occupancy projection is applied to the plan-based household projections to produce plan-based population projections.

Fig 10: Phase 1 and Phase 2 trend-based occupancy projections 2.30

2.25

2.20

2.15 South West Water - P1 2.10 South West Water - P2 2.05

2.00

1.95

2023 2013 2015 2017 2019 2021 2025 2027 2029 2031 2033 2035 2037 2039 2011

15

The population projections for Phase 2 include the results of the Census 2011. This has an impact on the level of population in 2011. At company level, the population is 3,300 lower in 2011 than estimated in Phase 1. Most of this change stems from Colliford which accounts for over 71% of the change. The differences are less marked in Roadford and Wimbleball – with a difference of less than 1,000 people in these areas.

The population forecasts generally have a stronger profile for Phase 2 compared with Phase 1 across most of South West Waters WRZ’s. At company level the trend-based projections for Phase 2 show average annual growth of 0.7% per annum between 2011 and 2040 compared with 0.5% for Phase 1. The stronger growth is due to the 2011-based ONS projections – as discussed in section 2. The plan-based projections are stronger than the Phase 1 plan-based projections. This is mainly due to the change in occupancy assumptions in the latest forecasts – with occupancy declining at a slower rate than under the Phase 1 projections.

The Phase 2 most-likely forecasts use the most appropriate population projection given the information available – including the revised mid-year estimates. The most-likely forecasts are lower than Phase 2 trend for company level – indicating that alternative projections were chosen for many local authorities across the South West Water area in place of the ONS 2011-based population projections.

Fig 11: Comparing population projections for Phase 1 and Phase 2

680000 1100000

660000 P1 TP Trend-based - P1 TP Trend-based - Colliford 1050000 Roadford 640000 P1 TP Plan-based - P1 TP Plan-based - 620000 Colliford 1000000 Roadford 600000 P1 TP Most-likely - P1 TP Most-likely - Colliford 950000 Roadford 580000 P2 TP Trend-based - P2 TP Trend-based - 560000 Colliford 900000 Roadford 540000 P2 TP Most-likely - P2 TP Most-likely - Colliford 850000 Roadford 520000 P2 TP Plan-based - P2 TP Plan-based - 800000

500000 Colliford Roadford

2015 2033 2013 2017 2019 2021 2023 2025 2027 2029 2031 2035 2037 2039 2011 440000 P1 TP Trend-based - 420000 Wimbleball 400000 P1 TP Plan-based - Wimbleball 380000 P1 TP Most-likely - Wimbleball 360000 P2 TP Trend-based - Wimbleball 340000 P2 TP Most-likely - 320000 Wimbleball P2 TP Plan-based -

300000 Wimbleball

2021 2025 2013 2015 2017 2019 2023 2027 2029 2031 2033 2035 2037 2039 2011

The Phase 2 household projections show around 10,100 fewer households in 2011 than estimated in Phase 1 in the South West Water area. The difference was not distributed equally across the WRZ’s – Colliford accounts for approximately 51% of the difference, Roadford 32% and Wimbleball 18%.

At company level, household forecasts are stronger than Phase 1 for the most-likely projections over long-term. This is most notable under trend where stronger population projections offset the slowed decline in occupancy, resulting in stronger household growth. On the other hand, both plan and trend-based projections are weaker in Phase 2 compared to Phase 1.

16

Fig 12: Comparing household projections for Phase 1 and Phase 2

300000 550000 290000 P1 HH Trend-based - P1 HH Trend-based - 280000 Colliford 500000 Roadford 270000 P1 HH Plan-based - P1 HH Plan-based - Colliford Roadford 260000 450000 P1 HH Most-likely - P1 HH Most-likely - 250000 Colliford Roadford 240000 P2 HH Trend-based - 400000 P2 HH Trend-based - 230000 Colliford Roadford 220000 P2 HH Most-likely - 350000 P2 HH Most-likely - Colliford Roadford 210000 P2 HH Plan-based - P2 HH Plan-based -

200000 Colliford 300000 Roadford

2021 2039 2021 2039 2013 2015 2017 2019 2023 2025 2027 2029 2031 2033 2035 2037 2011 2013 2015 2017 2019 2023 2025 2027 2029 2031 2033 2035 2037 2011 220000 P1 HH Trend-based - 200000 Wimbleball P1 HH Plan-based - 180000 Wimbleball P1 HH Most-likely - 160000 Wimbleball P2 HH Trend-based - 140000 Wimbleball P2 HH Most-likely - 120000 Wimbleball P2 HH Plan-based -

100000 Wimbleball

2021 2039 2013 2015 2017 2019 2023 2025 2027 2029 2031 2033 2035 2037 2011

17

4 Uncertainty analysis 4.1 Background

Projections become increasingly uncertain the further they are carried forward and long-term projections should be used with caution. The outputs from the project include estimates of uncertainty for the population and household projections. The three different forecasts are built from different assumptions and therefore recognise that the future is inherently uncertain.

There is no objective basis to put error bands around the plan based projections due to the recent changes in the planning system. Previous plans and the scale of housing development was set at the regional level, whilst the latest plans are produced at the local level. Measuring the accuracy of previous plans is also complicated by the economic downturn. Analysis for the most-likely forecast found that the plan-based forecasts are more credible for Phase 2 in many cases than they were for the Phase 1 forecasts, reflecting the fact that local authority plans are in a more advanced state than they were12 months ago. The plans also appear to factor in the current sluggishness in house building, with targets reduced for the next 5 years compared with under Phase 1.

One of the limitations of the traditional deterministic approach – used in the UK to produce the official population projections – is that no probabilities are attached to the principal projections, so users are given no information about the uncertainty associated with them or, with respect to the variants, are given no indication of how these compare to the principal projections in terms of certainty2. In theory it is possible however to produce a range of uncertainty around the trend-based population projections, based on comparisons with previous official projections against mid-year population estimates and we explore using this approach in this section.

ONS themselves do not produce measures of uncertainty around population projections. To help understand the uncertainty, a number of variant projections are produced based on alternative plausible demographic scenarios at national level. We have applied the assumptions to local authority projections to produce alternative scenarios at water resource zone level.

ONS have not produced variant projections for the 2011-based projections, however the assumptions used to create the 2011-based projections are the same as those used to create the 2010-based analysis so the scenario analysis using the old variants is still valid.

4.2 Stochastic analysis

Experian are currently working on an update the stochastic analysis based on the revised mid-year estimates and the results from the 2011 Census.

4.3 Scenario analysis

ONS do not produce measures of uncertainty around the population projections as it is not possible to do so using the deterministic approach used. Instead they provide a number of variant projections to demonstrate the uncertainty that different assumptions have on the population projections. However these variants are only produced at national level. Experian have applied the national variant assumptions from 8 scenarios to local authorities to simulate the variants at a lower geographical level that can then be applied to forecasts at water resource zone level. The variants chosen for this analysis:

2 ONS Chapter 6: Variants, 2010-based NPP Reference Volume, March 2012

18

 High fertility  Low fertility  High life expectancy  Low life expectancy  High net migration  Low net migration  High population (combining all the ‘high scenarios’ above)  Low population (combining all the ‘low scenarios’ above)

ONS have not produced variants for the 2011-based projections so we are unable to update our previous scenario analysis. However, the analysis from Phase 1 is still valid since ONS did not update the assumptions for natural change and migration for the 2011-based projections.

4.3.1 Scenario results

Results are presented in terms of percentage difference from the trend projection. In all cases the ‘high population’ and ‘low population’ scenarios represent the greatest difference from baseline – which at most is around +/- 10% compared with the baseline. However, it should be considered that the scenarios apply national assumptions and differences in migration flows particularly could have a much larger impact on population growth or decline at a local level.

The upper band and lower band uncertainty scenarios provide a maximum extent that the projections could fall within and therefore the bands are typically quite wide – especially on the downside in the case of South West Water. The upper and lower band scenarios can be interpreted as the outcome if average errors that have occurred at local authority level in the region continue and are compounded into the future. The results for South West suggest that in the past, ONS projections have tended to overestimate rather than underestimate the scale of population growth in the region.

Fig 13: Uncertainty scenarios for South West Water

15.0%

10.0% High Fertility

5.0% Low Fertility High Life Expectancy 0.0% Low Life Expectancy

-5.0% High Migration Low Migration -10.0% High Population

-15.0% Low Population

2010 2012 2014 2016 2018 2020 2024 2026 2030 2032 2036 2038 2022 2028 2034 2040

19

Appendix A: Phase 2 Data Sources

 Local authority provided planned dwelling data and local authority plans (see Contact Log)

 ONS 2011 interim sub-national population projections – Local Authority, released 28 September 2012 – projections used to produce the trend-based Phase 2 projections http://www.ons.gov.uk/ons/rel/snpp/sub-national-population-projections/Interim-2011-based/stb-2011-based- snpp.html

 ONS 2010 sub-national population projections – Local Authority, released 21st March 2012 – these were used to inform the Phase 2 most-likely projection and were the trend-based/ most-likely projection for Phase 1 http://www.ons.gov.uk/ons/rel/snpp/sub-national-population-projections/2010-based-projections/stb-2010- based-snpp.html

 ONS 2010 National population projections, principal and variant projections, released 26 October 2011 http://www.ons.gov.uk/ons/rel/npp/national-population-projections/2010-based-projections/index.html

 DCLG 2011 household projections – Local Authority, released 9th April 2013 http://www.communities.gov.uk/publications/corporate/statistics/2033household1110

 ONS revised mid-year estimates 2002-2010 – Local Authority, released 30th April 2013 http://www.ons.gov.uk/ons/guide-method/method-quality/imps/improvements-to-local-authority-immigration- estimates/index.html

 ONS mid-year estimates 2001-2010 – Lower Super Output Area, released September 2011 http://www.ons.gov.uk/ons/rel/sape/soa-mid-year-pop-est-engl-wales-exp/mid-2010-release/index.html

 Census of population, 2011, released January/ February 2013 http://www.nomisweb.co.uk/census/2011/key_statistics

 Census of population, 2001, released 30 March 2004. http://www.nomisweb.co.uk/home/census2001.asp

 London Plan The London Plan Spatial Development Strategy for Greater London, released July 2011 http://www.london.gov.uk/priorities/planning/londonplan

 GLA Population Projections 2011 Round, SHLAA, Borough SYA, released 16th December 2011 http://data.london.gov.uk/datastore/package/gla-population-projections-2011-round-shlaa-borough-sya

 2011 round SHLAA based household projections - standard fertility variant, released 16th December 2011 http://data.london.gov.uk/datastore/package/2011-round-shlaa-based-household-projections-standard-fertility- variant

 Experian Output Area level datasets, 2001-2040, derived from Census 2001, and pushed forwards using information from the electoral role, PAF files and responses to household lifestyle surveys. Controlled to 2011 Census output area results (released Jan/ Feb 2013).

 Property Pipeline information supplied by Emap Glenigan (April 2012).

20

Appendix B: How property pipeline level data is built into the demographic forecasts

The methodology that we use to build residential property pipeline information into our demographic forecasts utilizes site level planning application and contract progress data that is sourced from Emap Glenigan. The approach adopted by Emap Glenigan involves weekly visits to the local planning authorities to gather information regarding new planning applications. In addition to this Emap Glenigan’s information gathering approach features regular phone calls to “plan applicants” (undertaken by a dedicated team of around 40 people) in order to establish the planning application/contract progress stage that each site has reached. Accordingly, through Emap Glenigan we are able to access real time information regarding the country’s residential property pipeline.

To utilize Emap Glenigan’s site level planning application and contract progress data in our demographic forecasts we first need to establish the likelihood that each site in the property pipeline has of being “built-out”. To do this we use a procedure (developed in consultation) with Emap that assigns “build out” probabilities according to the stage that each site has reached in the planning /contracting process and the insight (based on experience) that this information provides regarding the likelihood that the associated scheme will be completed. In particular the “build out” probability that is assigned to each site reflects the maximum of the probabilities that are shown in Table 2 regarding site planning and contract stages.

Table 2: Emap Glenigan Probabilities

Planning Stage Probability Contract stage Probability Planning Not Required 0.98 Start on Site 1.00 Plans Appr on Appeal 0.95 Contract Awarded 0.75 Detail Plans Granted 0.90 Preferred Bidder Appt 0.50 Reserved Matters Granted 0.85 Bills Called 0.45 Detailed Plans Submitted 0.80 Tenders Returned 0.40 Detail Plans Withdrawn 0.60 Tender Currently Invited 0.30 Detail Plans Refused 0.55 Applications to Tender 0.25 Outline Plans Granted 0.54 Pre-Tender 0.20 Circular 18/84 0.53 Outline Plans Submitted 0.52 Appr Reserved Matters 0.55 Listed Building Consent 0.48 Pre-Planning 0.45 Public Enquiry 0.40 Outline Plans Refused 0.30 Outline Plans Withdrawn 0.20

To calculate the population that is associated with each site in the residential property pipeline the “build out” probability is simply multiplied by the number of units that are planned for each site and then multiplied again by our estimate of the average occupancy rate in the relevant Output Area.

The final stage in the methodology that we use to build residential property pipeline information into our demographic forecasts requires us to estimate when each “potential“ new development is likely to be completed. If start and completion dates are not available for a given site we take a conservative view that the site will be completed 4 years after the date at which we are making our forecasts (if the number of units in the project is less than one thousand). If the number of units exceeds one thousand, the project is given a completion date 12 years after the start date. Finally simple linear interpolation techniques are used to determine the speed at which each site is “built out” (and hence population accumulated) over the construction period.

21

Appendix C About us

22

Experian

Experian is a global leader in providing information, analytical and marketing services to organisations and consumers to help manage the risk and reward of commercial and financial decisions.

Combining its unique information tools and deep understanding of individuals, markets and economies, Experian partners with organisations around the world to establish and strengthen customer relationships and provide their businesses with competitive advantage.

For consumers, Experian delivers critical information that enables them to make financial and purchasing decisions with greater control and confidence.

Clients include organisations from financial services, retail and catalogue, telecommunications, utilities, media, insurance, automotive, leisure, e-commerce, manufacturing, property and government sectors.

Experian Group Limited is listed on the London Stock Exchange (EXPN) and is a constituent of the FTSE 100 index. It has corporate headquarters in Dublin, Ireland, and operational headquarters in Costa Mesa, California and Nottingham, UK. Experian employs around 15,500 people in 36 countries worldwide, supporting clients in more than 65 countries. Annual sales are in excess of $3.8 billion (£1.9 billion/€2.8 billion).

For more information, visit the Group's website on www.experiangroup.com

The word 'Experian' is a registered trademark in the EU and other countries and is owned by Experian Ltd and/or its associated companies.

23

Appendix F Climate change

Page F0

SWW Water Resources Management Plan 2015 – 2040

Appendix F: Climate Change

1 Vulnerability of our system to climate change

As part of our pre-consultation process, we shared with the Environment Agency our assessment of the vulnerability of our system to climate change. We concluded that our system falls into the “low vulnerability” category as set out in the Water resources planning guideline1.

A copy of our report is shown below.

1 Environment Agency, Ofwat, Defra and the Welsh Government, “Water resources planning guideline”, October 2012

Page F1

SWW Water Resources Management Plan 2015 – 2040

Page F2

SWW Water Resources Management Plan 2015 – 2040

Page F3

SWW Water Resources Management Plan 2015 – 2040

Page F4

SWW Water Resources Management Plan 2015 – 2040

Page F5

SWW Water Resources Management Plan 2015 – 2040

Page F6

SWW Water Resources Management Plan 2015 – 2040

Page F7

SWW Water Resources Management Plan 2015 – 2040

Page F8

SWW Water Resources Management Plan 2015 – 2040

Page F9

SWW Water Resources Management Plan 2015 – 2040

Page F10

SWW Water Resources Management Plan 2015 – 2040

Page F11

SWW Water Resources Management Plan 2015 – 2040

Page F12

SWW Water Resources Management Plan 2015 – 2040

Page F13

SWW Water Resources Management Plan 2015 – 2040

Page F14

SWW Water Resources Management Plan 2015 – 2040

Page F15

SWW Water Resources Management Plan 2015 – 2040

Page F16

SWW Water Resources Management Plan 2015 – 2040

Page F17

SWW Water Resources Management Plan 2015 – 2040

Page F18

SWW Water Resources Management Plan 2015 – 2040

Page F19

SWW Water Resources Management Plan 2015 – 2040

Page F20

SWW Water Resources Management Plan 2015 – 2040

Page F21

SWW Water Resources Management Plan 2015 – 2040

Page F22

SWW Water Resources Management Plan 2015 – 2040

Page F23

SWW Water Resources Management Plan 2015 – 2040

Page F24

SWW Water Resources Management Plan 2015 – 2040

Page F25

SWW Water Resources Management Plan 2015 – 2040

Page F26

SWW Water Resources Management Plan 2015 – 2040

2 Summary of method used for the derivation of monthly flow sequences at surface water sites

As part of our pre-consultation process, we shared with the Environment Agency our analysis of the two approaches, Approach 1.2 (UKCP09 flow factors) and Approach 1.4 (Future Flows), which we could use within our Water Resources Management Plan (WRMP).

We concluded the monthly UKCP09 flow factors are reasonably representative of the range of possible climate change scenarios and in our view are appropriate for our current water resources planning work.

A copy of our report is shown below, with the exception of the Appendix which is solely a reproduction of parts of the Water Resources Planning Guideline2.

2 Ibid. 1

Page F27

SWW Water Resources Management Plan 2015 – 2040

Page F28

SWW Water Resources Management Plan 2015 – 2040

Page F29

SWW Water Resources Management Plan 2015 – 2040

Page F30

SWW Water Resources Management Plan 2015 – 2040

Page F31

SWW Water Resources Management Plan 2015 – 2040

Page F32

SWW Water Resources Management Plan 2015 – 2040

Page F33

SWW Water Resources Management Plan 2015 – 2040

Page F34

SWW Water Resources Management Plan 2015 – 2040

Page F35

SWW Water Resources Management Plan 2015 – 2040

Page F36

SWW Water Resources Management Plan 2015 – 2040

Page F37

SWW Water Resources Management Plan 2015 – 2040

Page F38

SWW Water Resources Management Plan 2015 – 2040

Page F39

SWW Water Resources Management Plan 2015 – 2040

Page F40

SWW Water Resources Management Plan 2015 – 2040

Page F41

SWW Water Resources Management Plan 2015 – 2040

Page F42

SWW Water Resources Management Plan 2015 – 2040

Page F43

SWW Water Resources Management Plan 2015 – 2040

Page F44

SWW Water Resources Management Plan 2015 – 2040

Page F45

SWW Water Resources Management Plan 2015 – 2040

Page F46

SWW Water Resources Management Plan 2015 – 2040

3 Climate Change impacts on groundwater Deployable Output

It was agreed with the Environment Agency that groundwater sources are not the dominant source of supply in any of our Water Resource Zones (WRZs). As a consequence analysis of climate change impact was tailored to provide a level of assessment appropriate to each source using the most suitable techniques and the data available. The Water resources planning guideline3 on the assessment of climate change was closely followed but with some variations agreed with the Environment Agency.

Our consultants, AMEC Environment & Infrastructure UK Ltd, were commissioned to carry out the analysis. The majority of our groundwater sources abstract from the Otter Sandstone aquifer in East Devon in the Wimbleball WRZ. These sources were assessed for climate change impacts by consideration of our Dotton 1, 2, 3 and 7 boreholes and our Otterton 1A borehole using a combination of lumped spreadsheet models4 and the newly developed Otter Valley Groundwater Model. The lumped spreadsheet model was developed by AMEC (then ENTEC UK Ltd) during AMP4. It was adopted into the methodology for climate change impact assessment as part of the PR09 WRMP planning process.

One borehole (Bovey Lane) and two spring sources (Wilmington and Hook & Cotley) abstract from the Upper Greensand aquifer which occurs outside the Otter Valley. These sources were assessed using the lumped spreadsheet model approach alone.

For each of these sources, GR2 spreadsheets were constructed and calibrated against heads and flows. The input rainfall and Potential Evaporation (PE) were then perturbed using the 20 factors taken from the UKWIR09 rapid assessment report5. The perturbed rainfall and PE gave a series of perturbed groundwater levels and spring flows.

The Otter Valley Groundwater Model was then used to confirm the validity of the results from the GR2 spreadsheets. This was done by identifying the driest UKCP09 scenario which was used to perturb the historic rainfall and PE utilised in the model to generate modelled water levels.

In addition the model was also used to assess how sea level rises may affect water levels in the Otter Sandstone. The Otterton 1A borehole lies adjacent to the coast and its operation is controlled in such a way as to prevent saline intrusion. Any sea level rise due to climate change has therefore the potential to restrict source output and requires specific consideration.

The predicted impact from sea level rise was combined to the predicted decline in groundwater level identified from modelling. The combined impact was assumed to have an equivalent impact on the saline monitoring borehole such that the trigger level for abstraction control was revised from 0.5 mAOD to 0.69 mAOD. Analysis of the performance of Otterton 1A in relation to the trigger level during low groundwater level condition in 2012 revealed the likely level of impact on Deployable Output (DO).

The results of the analysis are provided in the tables below.

3 Ibid. 1 4 UKWIR, “Effects of climate change on river flows and groundwater recharge: A practical methodology – recharge and groundwater level assessment” (06/CL/04/8), 2007 5 UKWIR “Assessment of the Significance to Water Resource Management Plans of the UK Climate Projections 2009” (09/CL/04/11), 2010

Page F47

SWW Water Resources Management Plan 2015 – 2040

Table F.1: Climate change impacts on Wimbleball WRZ groundwater sources

Source Water level Reduction in Comments or flow deployable reduction output (Ml/d) Dotton boreholes 0.25 m 0.0 Water quality and pump constraints Otterton 1A 0.19 m 0.9 DO impact assessed from impact on S1 monitoring ole Bovey Lane 0.38 m 0.0 Source pump constrained Wilmington Springs 0.91 Ml/d 0.29 Base DO 1.2 Ml/d Hook & Cotley Springs 0.93 Ml/d 0.32 Base DO 1.25 Ml/d

A detailed report describing the climate change impact assessment undertaken by AMEC on our Wimbleball WRZ groundwater sources has been submitted to the Environment Agency.

Consideration of a water level decline due to climate change at other Otter Valley sources suggests their DO would not be reduced due to groundwater level not being the constraining factor. These sources are shown in the table below.

Table F.2: Constraints on Otter Valley groundwater sources

Borehole(s) Constraint Harpford boreholes 4, 7, 8 & 9 Abstraction licence constrained Otterton 4 Abstraction licence constrained Colaton Raleigh 2 & 4 DO already considered to be zero Kersbrook 2P Water quality constrained Greatwell boreholes 1, 2, 3, 4 & 5 Water quality and abstraction licence constrained

Groundwater sources outside of the Wimbleball WRZ make up only a small percentage of the resources compared to rivers and reservoirs and there are no solely groundwater-fed treatment works. Nevertheless some consideration of climate change impact has been undertaken with the conclusions laid in the table below.

Page F48

SWW Water Resources Management Plan 2015 – 2040

Table F.3: Climate change impacts on groundwater sources outside Wimbleball WRZ

Source WRZ Reduction in Comments deployable output (Ml/d) Duckaller Borehole Roadford 0.0 0.25m* water level impact not sufficient to affect DO Vennbridge Borehole Roadford 0.0 0.25m* water level impact not sufficient to affect DO Littlehempston Roadford 0.0 Sources sensitive to river level. Groundwater Climate change impact on river Dart not sufficient to reduce source DO

* Duckaller and Vennbridge abstract from Permian sandstone considered similar to Otter Sandstone for climate change assessment purposes. A water level decline of 0.25m as estimated at Dotton has been assumed for these sources.

Page F49

SWW Water Resources Management Plan 2015 – 2040

Appendix G Summary of strategic environmental assessment

SWW Water Resources Management Plan 2015 – 2040

South West Water Strategic Environmental Assessment of Water Resources Plan 2009 Summary Document

Hyder Consulting (UK) Limited 2212959 5th Floor The Pithay All Saints Street Bristol BS1 2NL United Kingdom Tel: +44 (0)870 000 3003 Fax: +44 (0)870 000 3903 www.hyderconsulting.com

South West Water Strategic Environmental Assessment of Water Resources Plan 2009 Summary Document

Author Gemma Blackler

Checker Nicola Hartley

Approver David Hourd

Report No DV53405/NTS3

Date 16/03/2009

This report has been prepared for South West Water in accordance with the terms and conditions of appointment for Strategic Environmental Assessment of Water Resources Plan 2009 dated March 2009. Hyder Consulting (UK) Limited (2212959) cannot accept any responsibility for any use of or reliance on the contents of this report by any third party.

(both positive and negative) of the Plan and its reasonable alternatives. Introduction The findings of the SEA are presented in the Environmental Report 1. Hyder Consulting Ltd. was appointed by South West Water to carry out a Strategic Environmental Assessment (SEA) of its Water Resources Plan The South West Water WRP (WRP). The Water Act 2003 introduced a legal requirement into the Water Industry The water industry, through UK Water Industry Research Ltd (UKWIR), Act 1991 for water companies to prepare, publish and maintain WRPs. recognises that WRPs may be subject to SEA under the requirements of These new provisions are contained in sections 37A to 37D of the Water the European Directive 2001/42/EC ’on the assessment of effects of certain Industry Act and came into force in April 2007. plans and programmes on the environment’ (the SEA Directive). The SEA Directive has been transposed into UK legislation as Regulations. In The WRP outlines how South West Water proposes to meet the essential England this is the Environmental Assessment of Plans and Programmes water supply needs of its customers through to the year 2035 in a Regulations 2004 (Statutory Instrument 2004, No. 1633). sustainable manner. The WRP covers all of South West Water’s water supply area, which covers Cornwall and Devon and small parts of Somerset South West Water has chosen to undertake an SEA to ensure that and Dorset, shown in Figure 1 below. environmental issues are considered throughout the development of the WRP. Figure 1: South West Water Strategic Supply Areas (SSAs)

An Environmental Report of the Draft WRP 2008, detailing the SEA process and outcomes, was prepared and then consulted upon in May 2008. Following the consultation, a Second Draft WRP and Environmental Report were produced before the WRP was finalised. This Summary Document provides a non-technical summary of the information provided in the Environmental Report of the Final WRP 2009.

The Purpose of SEA The primary aim of the SEA process is to provide for a high-level of protection of the environment. By ensuring the integration of environmental issues into the preparation of plans and programmes, SEA encourages sustainable development.

SEA is a decision-support tool, providing information on the environmental 1 South West Water (2009): Strategic Environmental Assessment of effects of the WRP. The output of the SEA process informs both the Plan makers and interested parties of possible significant environmental effects Water Resources Plan – Environmental Report . Report Number DV53405/ER3

1

The water resources planning process requires a variety of studies to be Preferred options were chosen from the list of feasible options in carried out in order to establish the supply and demand balance in water consideration of the security of supply issues, economic factors and supply within all the South West Water SSAs. Where deficits are identified, environmental impacts (including the findings of the SEA). potential supply and demand management options to meet the shortfall are drawn up. Consultation is a key component of the water resource planning process. South West Water invited comments on the WRP and Environmental To arrive at the options detailed in the WRP, a range of demand and supply Report at various stages in the plan development process. forecasting calculations were carried out. These calculations highlighted those SSAs that are in, or are predicted to fall into, deficit i.e. demand for Habitats Regulations Assessment Screening water will be higher than available supply. A wide range of supply and European Union Directive 92/43/EEC (the ‘Habitats Directive’) 2 requires demand management options were then considered to offset the that any plan or programme likely to have a significant impact upon a determined deficit for each zone. These options are referred to as the unconstrained options . South West Water considered a range of Special Area of Conservation (SAC), candidate Special Area of Conservation (cSAC), Special Protection Area (SPA), potential Special unconstrained options for their WRP from the following four categories: Protection Area (pSPA) or Ramsar site, which is not directly concerned with • Resource Schemes - Options which increase the available water output the management of the site for nature conservation, must be subject to an through the gaining of additional water supply (such as new boreholes Appropriate Assessment. All of these sites are of European/international abstractions or increased river abstraction). importance. • Customer Side - Measures which optimise customer water use efficiency through education, advice, metering and other means. A separate screening exercise was undertaken in parallel to the SEA process to determine if the WRP options were likely to result in significant • Distribution Management - Measures which improve the efficiency and effects on these valuable ecological sites. The Report concluded that the flexibility of the distribution network, such as leakage management and new pipelines. WRP is unlikely to have significant effects upon these sites but that further review of some of the options should be undertaken in the future when they • Production Management - Measures used at the production stage to are brought forward for implementation. improve capacity and efficiency such as blending, treatment, pumping regimes etc.

These unconstrained options were then narrowed down to a list of feasible options by South West Water using criteria which included environmental, social, economic and practical reasons . The feasible options consisted of generic options, for example, improved water efficiency measures that could be applied anywhere across the plan area and site-specific options that are only appropriate in certain locations. 2 Council Directive 92/43/EEC of 21 May 1992 on the conservation of natural habitats and of wild fauna and flora

2

The SEA Process • There is a need to protect and where possible enhance the condition of designated protected areas (e.g. Sites of Special Scientific Interest The SEA for the WRP has been undertaken in a number of stages as (SSSIs), SPAs and SACs); shown below: • A number of rivers have poor water quality. There is a need to enhance • Stage A – Setting the context and objectives, establishing the baseline this where possible; and deciding on the scope; • Climate change poses a long-term threat and there is a need to adapt • Stage B – Developing and refining alternatives and assessing effects; to the risks it poses; • Stage C – Preparing the Environmental Report; • Seasonal variations exist in both groundwater and surface water flows; • Stage D – Consulting on the draft plan or programme and the • There are high quality landscapes within the WRP plan area which Environmental Report; and includes two National Parks and six Areas of Outstanding Natural • Stage E – Monitoring the significant effects of implementing the plan or Beauty; programme on the environment. • There are wide-ranging recreational opportunities and Public Rights of The Environmental Report is the key output of the SEA process. It details Way; and, the SEA process for the WRP and presents information on the effects of the • There is a need in the WRP plan area to reduce energy use and Plan. improve energy efficiency.

Stage A: Policy Context, Environmental Baseline and Key Issues Stage B: Assessing the WRP A number of policies, plans and programmes have been identified and reviewed that set out a range of environmental themes (e.g. water, climate SEA Objectives change, biodiversity, landscape, sustainable development, heritage, health The SEA objectives were developed during Stage A. The SEA objectives and well-being). Environmental baseline data has also been collected in provided a framework for assessing and improving the environmental order to establish trends and the current state of the environment. This performance of the WRP; ensuring maximum synergy with existing policies review highlighted key issues relevant to the WRP and these relate to the and plans. The SEA objectives are as follows (not in any order of priority): following: 1. Protection and enhancement of biodiversity, key habitats and species • Predicted rise in population and seasonal fluctuations from tourism could have effects on water supply and demand; 2. Protection and enhancement of the cultural, historic and industrial heritage resource

3

3. Protection and enhancement of the quality and quantity of the surface Preferred options were selected through consideration of the security of water environment and the groundwater resource supply, economic factors and environmental impacts (including the findings of the SEA). 4. Ensuring the appropriate and efficient use of land

5. Limiting the causes, effects of, and adapting to climate change Preferred options The primary objective of the WRP is to ensure that all South West Water’s 6. Ensuring sustainable use of water resources customers have a secure supply of water through to 2035, whilst having 7. Protection and enhancement of landscape character regard to economics and the environment.

8. Protection and enhancement of human health Preferred options and their predicted effects are shown in Table 1.

The above SEA objectives were used to test the WRP options. SEA Scoring System ++ Major positive Consideration of Alternatives As mentioned previously, South West Water has considered a wide range + Positive of options (unconstrained options) for their WRP under four categories or ? Uncertain ‘strategic alternatives’

O Neutral • Resource Scheme • Customer Side - Negative • Distribution Management - - Major negative

• Production Management Due to the complex nature of WRPs, it is not possible to assess one It is possible for options to score a combination of these ratings, for strategic alternative against another, e.g. metering is not necessarily better example, there may be potential positive and negative aspects for the same or worse than repairing leaks. The WRP has considered options from each objective. of these strategic alternatives. In the table, S, M and L refers to short, medium and long term effects. Feasible Options The feasible options were assessed against the SEA objectives. Feasible options included a combination of existing projects already initiated, and new schemes. Each feasible option was assessed for its potential impact on each of the SEA objectives in the short, medium and long term.

4

Table 1 – Summary of Preferred Option Effects

Options SEA Objectives

Biodiversity Cultural Heritage Surface and Ground Land Use Climate Change Sustainable use of Landscape Human Health Water water

S M L S M L S M L S M L S M L S M L S M L S M L

Sophisticated Conjunctive Management O O O O O O + + + O O O + + + + + + O O O + + +

Compulsory metering

O O O O O O O + + O O O + + + + + + O O O O O O

Changes to existing measured tariffs O O O O O O O O + O O O +/O + + +/O + + O O O O O O

Targeted water conservation information O O O O O O O O + O O O O + + O + + O O O O O O

Advice & information on direct abstraction & irrig ation O O + O O O O O + O O O O + + O + + O O O O O + techniques

Advice & information on leakage detection & fixing O O O O O O + + + O O O + + + + + + O O O O O O techniques

Water saving devices

O O O O O O O + + O O O O + + O + + O O O O O O

Recycling & reuse

O O O O O O O O O O O O + + + + + + O O O O O O

5

Options SEA Objectives

Biodiversity Cultural Heritage Surface and Ground Land Use Climate Change Sustainable use of Landscape Human Health Water water

S M L S M L S M L S M L S M L S M L S M L S M L

Other water efficiency initiatives O O O O O O O + + O O O O + + O + + O O O O O O

Customer supply pipe leakage reduction -/? O O -/? O O -/? O O -/? O O + + + + + + -/? O O -/? O O

Leak detection

O O O O O O O O + O O O + + + + + + O O O O O O

Pressure reduction programme O O O O O O O O O O O O + + + + + + O O O O O O

Advanced replacement of infrastructure for leakage -/? O O -/? O O -/? O O - O O + + + + + + -/? O O -/? O O reasons

Distribution capacity expansion -/? O O -/? O O O O O O O O O O O O O + -/? O O -/? O O

Diagnostic studies

O O O O O O O O O O O O + + + + + + O O O O O O

Improved leakage detection & reduction on raw water -/? O O -/? O O -/? O O O O O + + + + + + -/? O O -/? O O mains

Domestic water efficiency project O O O O O O O O + O O O + + + + + + O O O O O O

6

Options SEA Objectives

Biodiversity Cultural Heritage Surface and Ground Land Use Climate Change Sustainable use of Landscape Human Health Water water

S M L S M L S M L S M L S M L S M L S M L S M L

Small and medium enterprises project O O O O O O O O + O O O + + + + + + O O O O O O

Waste water efficiency at WWTW O O O O O O O O O O O O + + + + + + O O O O O O

Porth catchment clean up and replacement for Rialton O ++ ++ -/? -/? -/? O ++ ++ - -/+ -/+ O + + O ++ ++ -/? -/? -/? O ++ ++ WTW

Reintroduce abstractions at -/+/ Boswyn & Cargenwyn -/? -/? O -/? O O -/? O O O O O O O O O O O -/? O O ? + +

Restormel licence variation -/+/ -/+/ -/+/

? ? ? O O O + + + O O O + + + + + + O O O + + +

Northcombe WTW output increased capacity to 60 O O O O O O O O O O O O -/+ -/+ -/+ O O O O O O + + + Ml/d

Roadford/Northcombe -/+/ pumped storage from - - - O O O ------/+ -/+ -/+ + + + - O O ? + + Gatherley

7

Environmental impacts have to be balanced against economic and security Mitigation Measures of supply issues to meet the needs of the region over the next 25 years. South West Water is committed to environmental protection and South West Water has sought to make the best use of the water that is enhancement and recognises the need to avoid and to mitigate adverse already available rather than developing new resources wherever possible. effects on environmental resources as far as possible. Prior to undertaking Many of the preferred options score as neutral against most of the SEA any works, South West Water will ensure that all appropriate projects are objectives and there are clearly many potential benefits as demonstrated by reviewed from an environmental perspective prior to any site works being the number of pluses/green boxes, particularly in terms of ‘Climate Change’ initiated and that appropriate mitigation measures are implemented. The and ‘Sustainable Use of Water’. A particularly beneficial option is ‘Porth highest levels of environmental protection will be given to those catchment clean up’. It scores several major positives as the option environmental resources of international and national value, whilst also involves the clean up of a polluted catchment which has beneficial effects recognising the value of locally designated sites and interest features. on biodiversity, surface and groundwater, the sustainable use of water Essentially, the environmental sensitivity of all projects will be considered resources and human health and recreation. on a case by case basis. In addition to more specific mitigation measures, a number of general mitigation measures are suggested that are summarised Some of the options have the potential to have negative effects as indicated below: by the minuses/orange boxes. In many cases this is as a result of potential construction impacts that would be largely short-term and could be
Ecological studies to be undertaken, particularly if works are to be effectively mitigated through good working practices. Some of the preferred carried out in an area with designated sites or BAP Priority habitats options would require new abstraction licences to enable abstraction from and site specific mitigation measures to be developed including good surface or groundwater sources. The potential effects on biodiversity environmental codes of practice and appropriate protected species resources including designated sites and Biodiversity Action Plan (BAP) mitigation as necessary. Priority habitats and species were considered as part of the SEA. It is
Archaeological studies to be undertaken where works are to be carried considered unlikely that there would be significant negative effects on out in an area of cultural heritage or historical value. biodiversity resources. Furthermore, all abstraction licences would have to be subject to a licence consent issued by the Environment Agency and
Avoid impacting upon the setting or integrity of any scheduled during this process, the Environment Agency has to consider potential monuments or World Heritage Sites when undertaking site works. effects on environmental resources. Without licence consent South West
Any fuel and oil storage on site for the purposes of operating Water would not be able to proceed with some of the site-specific options. machinery would comply with the Control of Pollution (Oil Storage) (England) Regulations 2001 (Oil Storage Regulations). The implementation of the recommended mitigation measures by South West Water means that the options selected are not expected to have any
Applications to be submitted for licence variations and new licences significant adverse environmental impacts. as appropriate.

Replacement and/or repair of pipes should minimise disruption and must take into account any sensitive or designated sites, historic or cultural heritage resources, biodiversity and key habitats and species 8

(as identified in the environmental baseline) and try to avoid affecting monitoring process, South West Water will monitor abstraction rates, rates the public’s opportunities for recreation where possible. of flow and groundwater levels for all options where there is an abstraction from a surface or groundwater source in order to ensure that it is within
Where new pumping stations/WTWs are to be built, investigate Environment Agency licence conditions. South West Water also monitors potential brownfield sites as an alternative to using greenfield sites. leakage, compliance with drinking water standards, carbon emissions and energy consumption (including percentage from renewables).
Consideration of energy efficiency and including increasing use of energy from renewable sources. It will be necessary for the monitoring framework to be reviewed and updated on an ongoing basis, particularly in view of the long time span of Stage C: Preparing the Environmental Report the plan. The results of the SEA process have been presented in the Environmental Report and summarised in this Summary Document.

Stage D: Consultation Provisions Consultation is a key component of both the WRP preparation process and the SEA process, ensuring that the views of key stakeholders are appropriately incorporated at an early stage and in an effective manner.

Consultation on the Draft WRP and Environmental Report took place between May and August 2008 . Comments received during this time were incorporated into the Second Draft WRP 2009 and Environmental Report. The WRP and Environmental Report were finalised in March 2009.

Stage E: Monitoring A requirement of the SEA process is to monitor potentially significant environmental effects predicted.

Monitoring is expected to draw upon existing monitoring programmes (or proposed monitoring programmes) undertaken centrally by the Government, and other organisations, rather than set out to collect a full set of plan specific data, for example, Natural England Condition Assessments of designated sites and Environment Agency river quality data. However, it is the responsibility of the plan-maker to ensure that the data collated is relevant to the significant effects identified through the SEA process and can be used to monitor the environmental effects of the plan. As part of the

9

Appendix H Water resource zones

SWW Water Resources Management Plan 2015 – 2040

1 Our water resource zones

We defined our Water Resource Zones (WRZs) in accordance with the Water resources planning guideline1. We use three WRZs – Colliford, Roadford and Wimbleball - for planning and managing our water resources.

Each WRZ is served primarily, but not exclusively, by its strategic reservoir (ie Colliford, Roadford or Wimbleball Reservoir) and associated sources. For example, customers in the Roadford WRZ have a very high proportion of their water supplied by Roadford Reservoir and its associated sources, but a small proportion of their water is provided by transfers from Wimbleball WRZ.

The main transfers between each WRZ are shown schematically within our report to the Environment Agency as shown within this Appendix below. These include the:

. Bude transfer from Colliford WRZ to Roadford WRZ (potential maximum capacity of the order of 0.5 Ml/d)

. Saltash transfer from Roadford WRZ to Colliford WRZ (potential maximum capacity of the order of 3 Ml/d)

. Imports / exports from Wimbleball WRZ to Roadford WRZ near Exeter (potential maximum capacity of the order of 12 Ml/d)

. Tiverton to North Devon transfer from Wimbleball WRZ to Roadford WRZ (potential maximum capacity of the order of 4 Ml/d)

. SWW to Wessex transfer from Wimbleball WRZ to Wessex Water (potential maximum capacity of the order of 0.05 Ml/d)

Further information about each WRZ is given below.

1.1 Colliford WRZ

The Colliford WRZ covers most of Cornwall except the north east of the County.

We use Colliford Reservoir conjunctively with local reservoirs, two disused former china clay pits and river intakes to form Colliford WRZ. These sources are supplemented by a bulk transfer from Roadford WRZ of up to the order of 3 Ml/d. The storage of Colliford Reservoir can also be supplemented by pumped transfers from Restormel.

Colliford Reservoir is both a river regulation and a direct supply reservoir and supports supplies in three ways:

. releases to the River Fowey for abstraction and treatment at Restormel Water Treatment Works (WTW) . pumping water direct to De Lank and Lowermoor WTWs . supplying water, via a gravity pipeline, direct to St Cleer WTW.

1 Environment Agency, Ofwat, Defra and the Welsh Government, “Water resources planning guideline”, October 2012

Page H1

SWW Water Resources Management Plan 2015 – 2040

The Colliford WRZ comprises a number of supply sub-systems, based on WTWs which are supported either directly by Colliford Reservoir or indirectly by Colliford Reservoir via Restormel WTW and the Cornwall Spine Main.

Brief descriptions of these sub-systems are given below:

Restormel Abstractions from the natural flow of the River Fowey at System: Restormel intake and WTW, in Mid Cornwall, are supported by river regulation releases from Colliford Reservoir and Siblyback Reservoir. Most of the water treated at Restormel is pumped up to a high level service reservoir from where it gravitates throughout the Cornish peninsula via the Cornwall Spine Main.

Drift System: Drift Reservoir in the far west of Cornwall supplies Penzance. Abstractions are made from the reservoir direct to Drift WTW. Supplies can be supplemented by the Cornwall Spine Main.

Wendron Wendron WTW supplies the Lizard peninsula. The abstractions System: for the works are supported by pumped transfers from Stithians Reservoir when required. Water can also be transferred to the Lizard peninsula from Stithians WTW.

Argal System: Argal and College Reservoirs, in South West Cornwall, supply Falmouth. Water can also be transferred into the zone from Stithians WTW.

Stithians Stithians Reservoir can directly or indirectly supply much of west System: Cornwall either by direct supply to Stithians WTW or by pumped augmentation to the River Cober.

De Lank and De Lank intake and WTW, on Bodmin Moor, and Lowermoor Lowermoor WTW, serve Camelford, St Minver, Bodmin and St Columb Systems: Major. Abstractions at De Lank are supported by a pumped transfer of water to De Lank WTW directly from Colliford Reservoir. Lowermoor WTW receives its water from Crowdy Reservoir supplemented by a pumped transfer from Colliford Reservoir. Stannon Lake also supports this system.

Bastreet Bastreet intake and WTW, on the Withey Brook, supplies System: Launceston and Torpoint. Abstractions are supported by pumped augmentation of the Withey Brook from Siblyback Reservoir.

St Cleer St Cleer WTW, in East Cornwall, is supplied both from the River System: Fowey via Trekeivesteps intake supported by releases from Siblyback Reservoir, and directly from Colliford Reservoir and Park Lake. St Cleer WTW supplies Looe, Torpoint and Saltash. The Saltash area can be supported by a bulk transfer from Roadford WRZ.

The abstraction licences at Park Lake and Stannon Lake are time-limited to 2016. As required by WRMP planning guidelines it has been assumed in this Plan that both licences will be renewed.

Page H2

SWW Water Resources Management Plan 2015 – 2040

A list of sources within the Colliford WRZ is given in Table WRP1a.

1.2 Roadford WRZ

The Roadford WRZ covers a large part of Devon, from Plymouth, the South Hams and Torbay in the south to Bideford and Barnstaple in the north. It also includes parts of North East Cornwall. The area is served primarily by Roadford Reservoir operating conjunctively with other impounding reservoirs; river intakes and other sources.

The most important single source in the area is Roadford Reservoir on the River Wolf, a tributary of the River Tamar. We use Roadford to augment the River Tamar for abstraction downstream at Gunnislake and also for direct supply to parts of North Devon (via Northcombe WTW).

We can pump our abstractions from Gunnislake to Crownhill WTW for use in Plymouth and the surrounding area or transfer them, via the South Devon Spine Main, to Littlehempston WTW at Totnes for use in the South Hams.

Burrator Reservoir on the River Meavy is a direct supply reservoir which supplies water to Crownhill WTW and Littlehempston WTW. We can also use Burrator to support Dousland WTW which is primarily fed by the Devonport Leat (a transfer from the headwaters of the River Dart). The other important source of water for Crownhill WTW is the abstractions from the River Tavy at Lopwell.

Littlehempston WTW is primarily fed directly from the River Dart, riverside boreholes and radial collectors in addition to the transfers from Burrator and Gunnislake. The south of Devon is also supplied by a number of direct supply reservoirs on Dartmoor including Kennick, Trenchford, Tottiford, Fernworthy, Avon and Venford.

In addition to the Roadford water at Northcombe WTW, North Devon is supplied by a variety of local sources including Meldon Reservoir, Upper Tamar Lake and Wistlandpound Reservoir.

A list of sources within the Roadford WRZ is given in Table WRP1a.

1.3 Wimbleball WRZ

Wimbleball Reservoir was constructed by South West Water Authority, the predecessor organisation of South West Water, with part of the financing costs being paid by Wessex Water Authority (WWA). We use the reservoir principally for making augmentation releases to the River Exe for subsequent abstraction near Tiverton and Exeter. These releases support abstractions from the natural flow of the River Exe.

Wessex Water, the successor to WWA, uses the reservoir for direct supplies. Historically, Wessex Water has abstracted close to its annual licensed quantity and therefore this abstraction from the reservoir has been assumed in calculations of the Deployable Output available to South West Water.

Page H3

SWW Water Resources Management Plan 2015 – 2040

The Wimbleball WRZ is also dependent on the significant groundwater resources of East Devon. The geology of East Devon includes Permian breccias and sandstones, Triassic conglomerates and sandstones, and sandstones of the Lower Cretaceous Upper Greensand. Some Chalk blocks in the extreme east of the catchment have also been exploited historically. Of these strata the Triassic Sandstone Group (the Otter Sandstone and Budleigh Salterton Pebble Beds) of the southern part of the Otter valley is the most significant in terms of public water supply. The typical method of abstraction in this area is from groups of boreholes, such as at Dotton and Otterton.

Two key groundwater abstraction licences in the Otter Valley are time-limited to May 2013. In line with Environment Agency guidelines, there is a presumption that these licences will be renewed.

Elsewhere our public groundwater supplies in the East Devon area are more widely dispersed, tapping a range of strata but mostly the Upper Greensand, and Permian sandstones, where significant yields occur as a result of natural fissuring and fracturing. In these cases, the method of abstraction is normally from single boreholes or spring sources.

In order to minimise the need to pump river Exe-derived treated water to supply the eastern fringes of the catchment our strategy is to optimise East Devon groundwater abstractions.

A list of sources within the Wimbleball WRZ is given in Table WRP1a.

2 Copy of our report on WRZ integrity submitted to the Environment Agency

A copy of our report covering WRZ integrity which has been sent to the Environment Agency is shown below with the exception of the Appendix which is solely a reproduction of parts of the Water resources planning guideline2 .

2 Ibid. 1

Page H4

SWW Water Resources Management Plan 2015 – 2040

Page H5

SWW Water Resources Management Plan 2015 – 2040

Page H6

SWW Water Resources Management Plan 2015 – 2040

Page H7

SWW Water Resources Management Plan 2015 – 2040

Page H8

SWW Water Resources Management Plan 2015 – 2040

Page H9

SWW Water Resources Management Plan 2015 – 2040

Page H10

SWW Water Resources Management Plan 2015 – 2040

Page H11

SWW Water Resources Management Plan 2015 – 2040

Page H12

SWW Water Resources Management Plan 2015 – 2040

Page H13

SWW Water Resources Management Plan 2015 – 2040

Page H14

SWW Water Resources Management Plan 2015 – 2040

3 Information from the Environment Agency

We held discussions with the Environment Agency concerning our WRZ integrity and received the following letter confirming our approach is appropriate (see below).

Page H15

SWW Water Resources Management Plan 2015 – 2040

Appendix I Consultation

SWW Water Resources Management Plan 2015 – 2040

1 Consultation

In accordance with the Water resources planning guideline1, we have consulted with regulators, other stakeholders and customers when preparing our Water Resources Management Plan (WRMP).

1.1 Pre-consultation

Before preparing our draft Plan, we consulted the following: . Environment Agency . Ofwat . Secretary of State for Environment, Food and Rural Affairs . Wessex Water . customers, through the Customer Challenge Group.

1.2 Neighbouring companies Contact Plan

In accordance with the Water resources planning guideline2, we published a Contact Plan to inform neighbouring companies of any „need‟ or „availability‟ with regard to future bulk supplies. We consulted directly with Wessex Water and Bristol Water regarding possible future bulk supplies. Details on future bulk supplies are given in Section 7.5.2 of this Plan.

1.3 Public consultation on the draft Plan

South West Water submitted the draft Plan to Defra on 28 March 2013 and we were given permission to publish it for public consultation in accordance with the Water resources planning guideline3.

We notified the following stakeholders of the publication of our draft Plan for consultation: . Secretary of State for Environment, Food and Rural Affairs . Environment Agency . Ofwat . Consumer Council for Water Western Region . Drinking Water Inspectorate . Cornwall Council . Devon County Council . Exeter City Council

1 Environment Agency, “Water resources planning guideline”, 2012 2 Ibid. 1 3 Ibid. 1

Page I1

SWW Water Resources Management Plan 2015 – 2040

. Plymouth City Council . Somerset County Council . Torbay Council . East Devon District Council . Mid Devon District Council . North Devon District Council . South Hams District Council . Teignbridge District Council . Torridge District Council . West Dorset District Council . Dartmoor National Park Authority . Exmoor National Park Authority . Natural England . English Heritage . Salmon & Trout Association . South West Rivers' Association . Westcountry Rivers Trust . Dartmoor Preservation Association . Wessex Water . Tiverton Canal Company Ltd . London Fire Brigade Water Team

Our draft Plan could be downloaded from our website. Alternatively, a paper copy could be viewed at our offices at Peninsula House, Rydon Lane, Exeter, EX2 7HR.

We produced a summary version of our draft Plan which could be downloaded from our website. Printed versions of this summary were available on request.

In accordance with the Water resources planning guideline4, following the consultation period we published a Statement of Response to the comments we received on our draft Plan. The Statement of Response outlined the changes we made to the Plan following the consultation. In January 2014 we provided further information to Defra and published it on our website.

In accordance with the Water resources planning guideline5, we have notified the stakeholders listed in Section 1.3 of this appendix of the publication of our Plan.

4 Ibid. 1 5 Ibid. 1

Page I2

SWW Water Resources Management Plan 2015 – 2040

Appendix J Business Planning and Price Review Process

SWW Water Resources Management Plan 2015 – 2040

The diagram below shows how the Water Resources Management Plan (WRMP) sits within the overall business planning and price review process.

Page J1

SWW Water Resources Management Plan 2015 – 2040