External Review of Sewer Flooding Risk Registers. Report. October 2008

Centre City Tower 7 Hill Street Birmingham B5 4UA External Review of Sewer Flooding Risk Registers Report October 2008 Demeter House Station Road Cambridge CB1 2RS UK Tel : 44 (0)1223 463500 Fax : 44 (0)1223 461007

External Review of Sewer Flooding Risk Registers Report Issue and Revision Record Rev Date Originator Checker Approver Description A 04/08/08 A Heather M Surendra P Chadwick First issue for comment B 22/08/08 M Surendra A Heather Revisions and corrections C 15/10/08 A Heather P Chadwick P Chadwick Final corrections This document has been prepared for the titled project or named part thereof and should not be relied upon or used for any other project without an independent check being carried out as to its suitability and prior written authority of Mott MacDonald being obtained. accepts no responsibility or liability for the consequence of this document being used for a purpose other than the purposes for which it was commissioned. Any person using or relying on the document for such other purpose agrees, and will by such use or reliance be taken to confirm his agreement to indemnify for all loss or damage resulting therefrom. accepts no responsibility or liability for this document to any party other than the person by whom it was commissioned. To the extent that this report is based on information supplied by other parties, accepts no liability for any loss or damage suffered by the client, whether contractual or tortious, stemming from any conclusions based on data supplied by parties other than and used by in preparing this report.

List of Contents Page Chapters 1 Introduction 1-1 2 Findings from Review 2-1 3 Summary of Main Observations 3-1 3.1 Comments from CCWater on At Risk Registers 3-1 3.2 Company Understanding of the At Risk Registers 3-3 3.2.1 Understanding of the Purpose and Issues 3-3 3.2.2 Understanding and Interpretation of the Nomenclature 3-4 4 Main Process Steps 4-1 4.1 Overview of the process 4-1 4.2 Developing the Register 4-1 4.2.1 On First Contact 4-1 4.2.2 During Clean-up Activity 4-2 4.3 Adding Properties to the Register 4-6 4.3.1 After Initial Investigation 4-6 4.3.2 After Detailed Investigation 4-9 4.3.3 During Investment Planning and Delivery 4-9 4.3.4 Addition of Properties Estimated as being at Risk 4-11 4.4 Keeping the Register Up to Date 4-11 4.4.1 The Link with Information Systems 4-11 4.4.2 Update Frequency 4-12 4.4.3 Recording Changes 4-12 4.5 Removing Properties from the Register 4-13 4.5.1 On First Contact 4-13 4.5.2 During the Clean-up Activity 4-13 4.5.3 After a long time without Flooding 4-13 4.5.4 After Company Action 4-13 4.5.5 As a Result of Better Information 4-13 5 Comments on the At Risk Register Approach 5-1 5.1 Register Name 5-1 5.2 Return Periods, Risk and Sub-registers 5-1 5.3 GSS Payments 5-1 5.4 Assigning Properties to the Register 5-2 5.4.1 First Time Flooding 5-2 5.4.2 Repeat Flooding 5-2 6 Recommended Process 6-1 7 Summary of Recommendations 7-1 7.1 Short-term recommendations 7-1

7.2 Longer term recommendations 7-2 Figure 4.1: Proportion of companies asking customers about the severity of flooding, if neighbours are known to be flooded, and if it is raining. 4-2 Figure 4.2: Adding Additional properties 4-3 Figure 4.3: Numbers of companies having exclusion rules for internal and external flooding. 4-4 Figure 4.4: Approaches to GSS payments and definitions of severe and extreme weather 4-6 Figure 4.5: Choice of registers for first-time flooding is reported, and treatment of severe weather 4-8 Figure 4.6: Proportion of companies claiming late additions as outputs 4-10 Figure 6.7: Link between flowcharts 4 Figure 6.1: Best practice process flow chart identification 6-1 Figure 6.2: Best practice process flow chart initial investigation 6-3 Figure 6.3: Best practice process flow chart GSS payments 6-5 Figure 6.4: Best practice process flow chart Adding to the register 6-6 Figure 6.5: Best practice process flow chart detailed investigation 6-8 Figure 6.6: Best practice process flow chart removal from register 6-9

1 Introduction Ofwat s At Risk Registers record properties that have flooded from sewers and are at risk of flooding again. There are separate registers for internal and external flooding. Each register is split showing the risk of flooding twice in ten years (2:10), once in ten years (1:10) and once in twenty years (1:20). The 1:20 years register, generally used to record properties flooded for the first time, has been introduced recently compared with the 2:10 years and 1:10 years registers that have been in place for many years. Companies target investment to remove properties from the 2:10 years and 1:10 years registers if there is a cost beneficial case for doing so or to mitigate the risk of flooding 1 if not. Investment is not normally targeted at properties on the 1:20 years register, but if properties suffer repeat flooding, they are moved to one of the other registers and solutions to resolve the problem are developed as appropriate. Properties flooded in severe weather (rare events) are recorded by the water companies but do not go on to the registers. The At Risk Registers are maintained by each of the ten water and sewerage companies in England and Wales. Each year the companies report details of the numbers of properties and areas at risk of sewer flooding, together with the number of incidents that year, to Ofwat through the annual June return in tables 3 and 3a. Two major outputs are monitored by Ofwat: The reduction in size of the 2:10 and 1:10 internal flooding register (the net reduction ). The number of problems resolved. A further measure is the number of properties subjected to flooding, which affects the wastewater Overall Performance Assessment (OPA). Ofwat publishes June Return Guidance which sets out information requirements and definitions of the data required in each line of the June return tables. The manual aims to standardise reporting definitions but it has become apparent, from comments by water companies, that information requirements are interpreted differently by the various companies, and that these differences might affect the number of properties added to the register. To help understand the differences in approach and their effect on company At Risk Registers, Ofwat commissioned to conduct a short study of each company s approach to reporting. The aims of the study were to: Identify and set out the various processes and procedures used to identify flooding, record it, allocate properties and areas to the appropriate register and eventually remove properties from the At Risk Registers as the risk of flooding is reduced by company action. Establish whether a flood would be allocated to the same At Risk Register irrespective of which company s methodology was used. 1 Mitigate the risk of flooding means relatively minor investment that reduces the flood risk without providing a permanent solution. Mitigation is appropriate where a permanent solution is too expensive to justify. It is also used to provide a temporary solution to severe flooding when a permanent solution will take a long time to deliver. Mitigation does not necessarily provide full protection from flooding and so properties protected by it are reported as a subset of the At Risk Register. 1-1

Check alignment with reporting guidelines, identify the most common approach, and recommend amendments to the guidance and/or processes to ensure consistency and promote good practice. Establish the impact on the At Risk Registers if Ofwat adopts any recommendations arising from the study. The study took the form of a structured interview with each of the ten water and sewerage companies, limited to a maximum of one day per company during June 2008. At this time most were compiling June Return data and preparing business plans for their August 2008 submission to the 2009 periodic review. Interviews were conducted in water company offices so that information systems and flood records could be seen. The authors would like to acknowledge the cooperation of the ten water and sewage companies in this review, as well as the contribution of the Consumer Council for Water to the review. It must be stressed that this was not an audit of company procedures. The field work was limited to one visit per company and although companies presented procedures and individual records of sewer flooding it was not possible to check that the procedures are followed in the field. It is also possible that our understanding of company procedures, and hence our view on differences in approach, might result from the different roles and knowledge of the people interviewed in different companies. The discussion topics were developed as the main differences emerged between companies. Therefore some items were not discussed at the earlier interviews. However, those items are mentioned where more than one approach was seen, as this provides evidence of differences between companies. Detailed quantification of the impact of differing approaches on the At Risk Registers was not undertaken and in many cases would not be feasible. A general indication of whether different approaches would have an upward or downward impact on the number of properties on the register is made in the report where appropriate. In this report, flooding means sewer flooding, unless stated otherwise. 1-2

2 Findings from Review All companies understand the At Risk Register and the nomenclature, but in matters of detail there are some differences of interpretation between companies. All companies consider that they are adhering to the June Return reporting requirements, but there are differences in interpretation of the Reporting Requirements. A flooding incident would not necessarily be allocated to the same register irrespective of which company s methodology was used. For example the availability and choice of data and analytical technique influences the assignment of return period and hence the choice of register. Since solutions are usually implemented for properties on the 2:10 years and 1:10 years registers, selection for a solution may depend on which company s area the flooded property is in. In most companies, the At Risk Register is compiled by automated reporting directly from job management systems. In others, it is compiled as an off-line database or spreadsheet, using more manual approaches to reporting from customer relationship management systems or job management systems. Effective examples of both approaches were seen. On reliability of the approaches, companies pointed out that both depend on the source data having a correct classification of the incident. Hence all companies conduct manual checking of individual records to check that the classification aligns with descriptions of the findings. Clarification of some aspects of the Reporting Requirements would help to improve the comparability of company registers. Areas for consideration should include: o The definition of, and approach to identifying, severe weather. Differences in approach can affect whether flooded properties are added to the register, or excluded on the grounds of severe weather. There are three aspects to this: clarifying the definition of severe weather, strengthening the guidance on best practice assessment of severe weather events, and ensuring that all companies apply best practice methods of assessment and report using the same definition of severe weather. o Clarifying the definition of internal flooding, especially for cellars and garages. o Clarifying the approach to flood return period analysis and the addition of first-time flooding to the At Risk Registers. o Clarifying the approach to assessing repeat flooding, especially where some instances are the result of severe weather and others are not. o Clarify the approach for recording properties when flooding is discovered at a much later date (through detailed investigations or anecdotal evidence from customers). Current practice varies between companies as to whether such properties are added to the register and count towards outputs. o Separating the monitoring of properties that have mitigation measures installed, to better reflect their true risk of flooding. There is agreement that the term at risk register is misleading and flooding incident register or unresolved flooding incident register would be a better description. There is also support from some companies for the register to record the risk of individual properties 2-1

flooding (frequency x severity), compared with the current approach which only records internal or external flooding in large frequency bands. There is general concern from companies about the potential impact of at risk registers on home information packs (HIPs) and property values. The inclusion of a measure of flood severity as well as frequency might help alleviate this, as it would enable property owners and buyers to distinguish between large risks and small ones. It would also enable the benefits of mitigation measures, as well as network enhancements, to be recognised. The approach to dealing with severe weather if there is repeat flooding differs between companies. In some cases, all events are counted once it has been established that a property floods when the weather is not severe. In others, severe weather is always excluded. The former approach is simpler for customers to understand and may better reflect the risk to the property, while also increasing the frequency recorded for that property. Sewer flooding is identified from physical signs at the site, such as the location of flood water, debris, movement of access covers etc. Customer reports and sometimes photographs also provide evidence of the origin of the flood water. All companies check if neighbouring properties have flooded at the same time as a reported flood, but the extent of the check varies between companies. The evidence suggests that more rigorous searches during the initial investigation will identify the greatest number of flooded properties. Flooding is attributed to overloaded sewers if no other cause can be found. Where hydraulic overload is suspected, the majority of companies conduct CCTV surveys after the incident to check that there are no hidden obstructions in the pipe. However, the subsequent removals of properties from the register, on the grounds of better information, suggest that this is not a precise science. Some companies continue to operate exclusion rules despite Ofwat having stated that no incidents should be excluded (for example the time taken to clean up after external flooding, or damp patches on cellar walls). Even where a company does not have a formal exclusion rule, it is clear that at the time of the initial investigation a judgement is made as to whether to conduct further investigation or to close the case, and hence on whether to report the flooding. Properties are removed from the At Risk Register when companies have taken action to resolve the problem; when it is found that the problem is not associated with water company assets; or sometimes if the property does not flood again for a specified length of time. Properties may be transferred between registers or removed when better information about the cause becomes available, as may happen during detailed network investigation after the flood. 2-2

3 Summary of Main Observations 3.1 Comments from CCWater on At Risk Registers As part of the review, we met a representative of the Consumer Council for Water (CCW) to obtain its views on companies procedures with respect to sewer flooding and at risk registers from a customer s perspective. The views of several CCW regions were collated and are summarised below: (i) Acceptability CCW research confirms that sewer flooding is not acceptable to customers. (ii) Customer reporting There is a perception that customers concerns about potential property blight leads them to underreport sewer flooding incidents. Therefore, CCW expects companies to pay special attention to identifying flooded properties, so that robust solutions can be specified. (iii) Internal flooding definition CCW is concerned that some companies define some types of flooding such as toilet overflows, flooding of garages with no internal doors and below-floor flooding as external flooding. CCW would like clearer definitions to ensure that theses are classified as internal flooding events. (iv) GSS payments Some companies are going beyond the statutory minimum requirement for GSS payments, which CCW welcomes. CCW also welcomes prompt payment, provided the company also pays due attention to preventing flooding. (v) Hydraulic overloading and risk assessment CCW feels that severity and frequency might be better measures of flood risk than the nominal internal/external and banded return-period registers (some companies also suggested this/agreed with this comment). The level of risk is not currently meaningful to customers and CCW feels that companies hide behind the technical classification. It is difficult for customers to understand why some flood events count and some do not. (vi) Cost benefit analysis CCW feels that cost benefit analysis should be used as a prioritisation tool rather than for the ultimate selection/rejection of a property for a solution. 3-1

(vii) Mitigation CCW suggests that if the result from implementing mitigation measures for a property is a lower level of flood risk than neighbouring properties, then perhaps it does not need to remain on the register. Companies should count properties where sewer flooding was only prevented by customers own mitigation actions. (viii) Garages and sheds CCW believes that if garages and sheds are in regular use and they are flooded, they should be counted as internal flooding. There is concern that companies are counting too many garages as external flooding. (ix) Design standards CCW notes that different companies have different design standards with respect to designing their networks to prevent flooding and wonders whether this is this equitable. (x) Communications CCW supports regular communication with customers on the register (for example, some companies send an annual letter). Communications need to be sensitive and show what is being done, with delivery dates. (xi) Information flow CCW contends that there is generally poor information flow between local authorities and water companies. Customers often report flooding to the wrong agency and information is not always passed on to the right one 3-2

3.2 Company Understanding of the At Risk Registers 3.2.1 Understanding of the Purpose and Issues The At Risk Register enables Ofwat to monitor company performance in managing its network and informs customers about network capacity issues. However, there were some comments on the purpose of the register: The classes of risk (2:10, 1:10, and 1:20) may be considered as a measure of the need for a solution, but in most companies, more sophisticated risk assessments or cost-benefit assessments are now being used. In some companies the three frequency bands and the differentiation for internal and external flooding are used for reporting purposes only. Some companies advocate a more continuous risk scale based on an assessment of the frequency, severity and scale of likely flooding, but not all companies would want to change from the existing approach. The register has been considered a technical device used between the companies and the regulators. Many companies stated that the introduction of Home Information Packs has made house sellers and buyers more aware of flooding. They are concerned that reporting of 1:10 and 2:10 flooding through property searches might introduce housing blight. Companies feel that this changes the emphasis on the At Risk Register and that great care should be exercised in placing properties on the register. All companies said that they make the decision on what properties to enter and they do not accept householder requests to be excluded from it, when the company has evidence that the property is subject to flooding. Dealing with properties where mitigation reduces the risk but the properties remain on the register. Companies generally agreed that there are benefits to maintaining records of which properties have had mitigation measures installed. However there are situations where the mitigation may reduce the flood return period to less than that of neighbouring properties which are not on the register. For example, mitigation measures for a house on the 1:10 register may reduce the flood risk to 1:30, whereas the risk for neighbouring properties is 1:25. Yet searches by a prospective house buyer would show the house was at risk of flooding because it is on the 1:10 register but not the neighbouring house. This could therefore mislead a house buyer as to the true risk and also blight the value of the house. Although the severe weather exemption is a necessary acknowledgment of the design limitations of the network, it can be confusing for customers, especially when their property is flooded several times under differing weather conditions. Most companies acknowledge the differences in interpretation of the reporting requirements, and would welcome more prescriptive guidance on reporting flooding. Adoption of private sewers: In the event that private sewers are transferred to the water companies, reporting could become much more complicated. For example, there are currently three frequency classes for both internal and external flooding, and if private sewer flooding is reported separately this will result in a total of 12 register categories. This may reduce the certainty of allocation and thus lead to more changes as better information is gathered during the asset management process. Companies do, however, acknowledge the need to monitor the impact of adopting private sewers on network management costs. 3-3

3.2.2 Understanding and Interpretation of the Nomenclature There was some debate about the meaning of At Risk Registers. Many companies pointed out that the registers do not list all properties at risk of flooding, but rather are flooding incident registers, because they list properties that have flooded. Several companies suggested that the term at risk register is misleading because customers may believe that their property is only at risk of flooding if it is listed on the register. In practice, the register is mostly restricted to properties that have flooded and which remain at risk of flooding in the future. Other properties may also be at risk of flooding due to hydraulic overload as a result of changing demand or weather patterns, but a suitable trigger event may not yet have happened. Flooding Incident Register or Unresolved flooding incident register may therefore be a better name for the At Risk Register, although any description would need to explain that the properties are listed because they remain at risk. A further area of confusion relates to the term sewer flooding. Ofwat s guidance defines a flooding incident as an event of flooding from a public sewer, whether foul, combined or surface water. It should perhaps be clearer that this definition excludes flooding incidents from sewers that are caused by surface run-off, fluvial flooding, some types of highway drainage or other sources of flooding where they are the responsibility of other agencies. Ofwat wishes to ensure that water customers only pay for sewer flooding schemes that address problems arising from flooding that water companies are clearly responsible for. Despite this lack of clarity, companies do generally exclude flooding events where the cause is clearly the responsibility of others, for example flood events arising from private sewers. The problem arises where the cause of flooding is unclear at the point of flooding (perhaps because there is a mix of overland flow and sewer surcharge) or the cause is remote from the point of flooding (for example, where high river levels prevent downstream overflows from operating properly and sewage backs up in the network). Companies assessments of such causes is generally limited to what the clean-up team can observe at the flooding site, with limited follow-up analysis to confirm the cause and identify the principal owner of the flooding problem. As companies adopt more holistic solutions in partnership with other members of the drainage community, the identification and sharing of flood risks between the responsible agencies will become more important. Properties are not usually added to the register unless flooding has occurred, although one company does add adjacent properties that it assesses to be at risk of flooding, even if they have not flooded at the time of an incident being investigated. 3-4

Conclusions and recommendations: At risk register is confusing terminology and does not describe what the register is. Although properties on the register remain at risk of further flooding, the ARR is not a comprehensive list of properties at risk of flooding. Therefore Flooding incident register or Unresolved flooding incident register better describes it. The at risk registers record frequency, but not severity of flooding, so they do not record risk (probability x consequence). This can make it difficult to distinguish between small internal flooding problems and large external flooding problems, for example. The impact extends to home information packs (HIPs), where the presence of a property on a register can have the same impact irrespective of its flooding risk, as it is conventionally understood. An alterative approach that could be considered would be a measure of the frequency, severity (depth, area) and location of flooding, rather than location and frequency alone. The result would be a continuous risk scale that would more closely align with the cost:benefit approaches now in use in most companies, although for reporting the risk would be grouped into a few (e.g. five) categories. A risk-assessment approach also applies to mitigation, where interventions reduce the frequency of flooding but do not eliminate the risk of flooding. Such properties currently remain on the register after mitigation has been delivered (although they are reported as having received mitigation). This can cause anomalies with neighbouring properties that are not on the At Risk Register but have a greater flood risk than those with mitigation in place. A risk measure could therefore include mitigation measures, to demonstrate the benefit that they deliver. If properties on the registers had a full risk rating, it would enable a distinction to be made between frequent small flooding incidents and infrequent large flooding incidents. It would also enable the benefits of mitigation to be recognised, where a property would remain on the register, but its risk rating would change. This would also reduce the potential impact of the register on property value and recognise the efforts made by companies to manage flood risks, when increasing the capacity of the network is not always cost-beneficial. Some companies already record flood severity information, (such as depth and extent of flooding and duration) to inform their prioritization methods and flooding cost-benefit assessments. Consistent methods of assessment would need to be agreed if this was taken forward. We recommend that Ofwat considers reviewing these comments with the industry and other stakeholders, to establish support for developing full risk registers, and what form the measures should take. It would be important, for example, that the risk categorisation was simple and easily understood by customers and property owners, and would require careful consultation. We also recommend that a definition of sewer flooding is provided with a clear statement of what it includes and what it excludes. This could also be the subject of consultation with the industry and other organisations with responsibility for drainage. The aim would be to ensure that all stakeholders in the management of flooding agree on a common definition and understand the boundaries between water company responsibilities and their own. This could help initiate or reinforce conversations that companies are already having with stakeholders in the context of integrated urban drainage. Once a definition for sewer flooding is agreed, companies would need to review their procedures for assessing and recording the cause of flooding and flood ownership. This is likely to develop as integrated urban drainage becomes established as common practice. 3-5

4 Main Process Steps 4.1 Overview of the process This section reviews the main activities that create and manage the At Risk Register in four sections: Developing the register Adding properties to the register Keeping the register up to date Removing properties from the register. There is an implication of a sequential process but in practice but the various activities are carried out as they are required for each flooding case. 4.2 Developing the Register 4.2.1 On First Contact All companies record all customer contacts through contacts centres and central call handling systems. All companies attempt to identify the main causes of flooding and whether the problem is with a privately owned sewer. In some companies the call centre staff can see electronic maps of the sewer networks and this improves their ability to differentiate between company and private sewers. When flooding is caused by private sewers, the customer is normally told that it is their responsibility, if it is clear that the sewer is blocked or collapsed. Sometimes a problem with the public sewer including hydraulic overload results in egress from a private sewer. Companies said that if there is any doubt about ownership or the cause then they will visit the customer. It is also common to ask customers if it is raining, the severity and location of the flood, and whether neighbours are affected (see Figure 4.1). This helps in prioritising and allocating the best response. For example, if a householder feels sure that no other properties are affected and it is not raining, then the problem is more likely to be blockage or collapse, than if a house floods during heavy rain or several properties are affected. This is not a definitive guide at this stage and in most companies is used to inform the operational response rather than for full classification of the event. Call handling software and the capability to forward information to field staff are variable. Although it is possible that jobs might not be forwarded at this stage, we found no evidence of this during our visits to companies. In some companies the call handlers have to learn the questions to ask customers, with scripts or prompt sheets on their desks. In others case based reasoning is used, in which the questions are presented to the call handler on screen. Subsequent questions are presented by the software, dependent on the preceding response. This decision-tree approach is highly consistent and results in a specific set of actions appropriate to the reported incident. It appeared to be efficient and in some companies it is directly linked to the job management system, GSS payment systems, and At Risk Register reporting. The use of GIS maps linked to the call handling software allows cal centre staff to see the location of recent calls, which helps in the early identification of wide area flooding. 4-1

Figure 4.1: Proportion of companies asking customers about the severity of flooding, if neighbours are known to be flooded, and if it is raining. Ask about severity of flooding Ask if neighbours affected Ask if it is raining The uncertainties at the time of the initial call mean that most companies investigate more flooding than their assets cause. Depending on circumstances they are probably also clearing some blockages of, and in some cases carrying out minor repairs to, customers pipes. At this stage in the process companies have an unverified list of reported floods and their nature. This is the list which eventually becomes the basis of the at risk register, after investigation of the problem, validation of the cause, and estimation of the likely frequency of flooding. Best practice Case-based reasoning offers advantages over manually read or memorised scripts for consistent and reliable evaluation of customer calls about sewer flooding, where establishing responsibility for the problem is important. Visibility of electronic maps of the sewer networks improves call handlers ability to differentiate between company and private sewers. By showing the location of recent calls it also helps to identify emerging area-wide problems as calls come in. Simple questions about whether it is raining and are neighbours affected give early indication of the nature of the problem. Direct linkage between the call handling software and the job management system and GSS payment system is an efficient and reliable means of data handling. 4.2.2 During Clean-up Activity All companies have targets for response times to investigate and clean up flooded properties. During the visit the details recorded by the customer contact centre are reviewed to confirm that they are correct. All companies provide for field staff to add comments and correct the details about the scale and location of the flooding, and ownership of the assets. At this stage it is common to identify blockages and collapses and to clear such problems immediately. Where no cause can be found it is assumed that the hydraulic capacity of the sewer has been exceeded. 4-2

In demonstrating their quality control procedures, some companies showed examples where field staff had classified flooding as hydraulic overload, and subsequently noted the removal from the pipe of an obstruction such as a brick. It might be that the obstruction was seen as resulting in reduced capacity and hence hydraulic overload or that staff felt under pressure to attribute the problem to underlying capacity rather than operational problems. All companies have a quality control step after the initial investigation, in which all entries are reviewed and their classification is checked. In the example above, the classification would be changed to blockage or collapse, depending on what other information was available about the source of the brick. The extent to which neighbouring properties are checked appears to vary significantly between companies. In some cases leaflets are left with neighbours but in others neighbouring properties are only checked if the clean-up staff can see evidence that they have flooded. This latter practice is likely to lead to fewer properties being reported as flooded at this stage, with a potential benefit to companies OPA measures. Figure 4.2: Adding Additional properties Search for Neighbouring Properties Addition of Neighbouring Properties Extensive Depending on evidence Minimal To current job To new job All companies were able to show examples of how additional properties are added to their database following initial clean-up. Some companies add them to special fields in the same job as the original flood, others create new entries specifically for neighbouring properties. One company mentioned anecdotal evidence that adding neighbouring properties to the original job requires less work and is therefore more reliably implemented. (i) Exclusion Rules Ofwat s reporting requirements state that all flooding must be recorded, and all companies do record details of all incoming calls about flooding. When deciding whether to record flooding, some companies operate exclusion rules, including a description of surcharge if external flooding can be cleaned up within 15 minutes and the exclusion of weeping manholes not causing a nuisance. Some companies reported difficulty in deciding how to handle damp patches on cellar walls where there is insufficient water to take a sample, and many will not record these as flooding until a sample can be taken. In practice most companies are likely to use judgement in deciding whether to record a flood, especially during heavy rain when surface water flooding is prevalent. If all discharges, however small, are reported then the number of external floods is likely to rise significantly. 4-3

The June Return Reporting Requirements already state that all flooding incidents should be recorded and exclusion rules do not apply. The guidance lacks clarity in its definition of sewer flooding (as described earlier) and although companies check that the problem arose from a public sewer not a private sewer, checks on root causes and ownership of the flooding problem are variable. There are also differences in the interpretation of internal flooding. Two companies consider garages to be external if they do not have an internal door connecting them to the property, even when they are integral to the building s construction (see Figure 4.3). We consider this to be a misinterpretation of the reporting guidelines it is certainly not a common interpretation which reduces the number of internal flooding incidents being reported. Although garages tend to have low door sills and therefore be more likely to flood than other parts of a building, we consider this is unlikely to make a big difference to the size of register. Classifying integral garage flooding as external might make it more difficult to justify investment for those properties where they flood. The June Return Reporting Requirement provides a definition of internal flooding but it appears that two companies consider internal garages with no connecting door to be external. The guidelines could be modified to state: However, garages forming an integral part of a property are classed as part of the building, even where there is no internal door to the house, and are included, even if their prime purpose is storage, etc. Some companies noted that customers often consider flooding of garages and cellars to be internal flooding. It is therefore possible that the current June Return definition does not align with customer preferences. This could be checked by inclusion in future customer perception or stated preference surveys. Figure 4.3: Numbers of companies having exclusion rules for internal and external flooding. Exclusion Rules for External Flooding 10 8 6 4 2 0 Overw helmed by w ater course Time to clean None Exclusion Rules for Internal Flooding 8 6 4 2 0 No internal door in garage Damp patch on cellar wall None No internal door in garage Damp patch on cellar w all None 4-4

Best practice All flooding incidents should be recorded, as required by the guidance. Adoption of robust quality control procedures is essential to check that the call-centre details are correct and to check for anomalies in work-order entries that might imply incorrect identification of the cause of flooding. Prioritising the early capture (generally within two weeks of an event occurring) of information about the number of neighbouring properties affected and other data on the causes and severity of flooding is important to prevent loss of data. Information captured long after flooding has taken place is likely to be more anecdotal and less certain of being correct. The allocation of flooded neighbouring properties to the same work order as the original flood ensures a better data audit trail. This may also be achieved by a database link to the original work order. Identification of the root causes and ownership of flooding problems is necessary to establish responsibility for resolving them. With the exception of identifying ownership of private sewers, practice in this area is variable. Flooding of garages should be categorised as internal flooding where the garage is an integral part of the building, even if there is no internal door. This aligns better with customers perceptions of internal flooding. (ii) Link to GSS payments and the Estimation of Flood Return period Some companies allocate GSS payments regardless of the flood return period, except in cases of wide area flooding where it is obvious that extreme weather is likely to have contributed. Such companies make more GSS payments than they are required to. Companies cited the difficulty in obtaining reliable return period data and benefits of showing goodwill in a difficult situation as the main reasons for this behaviour. Some company areas are covered by high resolution rainfall radar and in these areas companies are able to identify severe weather more quickly, therefore reducing the number of GSS payments. One company fully models the sewer event return period, even though this might mean missing the GSS payment deadline and having to make an additional penalty payment. Another company purchases weather analysis where it considers that unusually high rainfall might have happened, this approach balancing the cost of acquiring weather data and the cost of overpaying GSS. Anther calculates the return period of rainfall in each 1km grid square evry night, so return period analysis is available the following day. In addition to its effect on GSS payments, the assessment of flood return periods determines the register to which the flood will be allocated. The approach to assessing flood return period varies and this leads to differences in the allocation of properties to registers (see Figure 4.4.). Furthermore there is some confusion between the definition of extreme weather in the GSS guidelines and severe weather in the June Return reporting requirements for At Risk Registers. 4-5

Figure 4.4: Approaches to GSS payments and definitions of severe and extreme weather Definition of Exceptional Weather Definition of Severe Weather GSS and severe weather 1:10 yr storm 1:20 yr storm 1:40 yr storm Not know n 1:10 yr storm 1:20 yr storm 1:40 yr storm Normally paid before weather data av ailable Get w eather data if severe weather suspected Always get weather data first Note: Not known means that the subject was not discussed with company, or the contact for the study was not involved in the GSS process and was not certain of the company approach. We understand that Ofwat is currently consulting on the definition of exceptional weather for use with GSS regulations. We recommend that when this is complete Ofwat clarifies the severe weather and exceptional weather definitions. This would help to ensure that service is described in the same terms for all customers, regardless of where they live. There might also be potential to align GSS definitions and entry of properties on the at risk registers. 4.3 Adding Properties to the Register 4.3.1 After Initial Investigation In most companies the initial investigation is carried out at the time of clean-up. It is common for minor repairs to be carried out at the same time, for example clearing roots or repairing collapsed pipes. Most companies conduct an initial investigation within a fortnight of the event occurring, except in extreme cases of wide area flooding. When the clean-up and initial investigation is complete, companies have a database of reported flooding events, with follow-up details that confirm which appear to result from hydraulic overload, and which are from other causes. In many companies this subset of the customer contact or job management database forms the basis of the At Risk Register, although at this time the return period of the sewer event might not have been estimated. In some companies the relevant records are extracted and added to a separate flooding register, in a spreadsheet or local database. Companies having integrated information systems generally referred to a live register, where those with off-line approaches tended to update their registers monthly or annually. The definition of severe weather is a source of confusion and there are many approaches to determining if the flooding has resulted from severe weather, in part because of differences in the source data that are available in different regions. The issues are: The determination of flood return periods is complex and requires detailed network data and modelling. 4-6

Some companies use rainfall return period as a surrogate for flood return period, despite specific guidance from Ofwat that DG5 measures the frequency of flooding incidents and not the return period of the storm that causes the flooding. Some companies use the interval between repeat floods as a surrogate for flood return period. Some companies use weather radar data, but radar coverage is variable and not universally available. (Water companies do not operate weather radar systems.) Some companies use Environment Agency local rain gauge data, which may also be combined with data from other sources, but rain gauges are not necessarily near enough to the event being investigated to give confidence in the results. There is some evidence that even Met Office analysts differ on the return periods of individual rainfall events. The analysis of return periods is also complicated by the effects of climate change, which means that a weather event with a 1 in 20 return period today may have a shorter return period (i.e. be more frequent) in the future. In addition, urban creep has led to a greater proportion of rainfall entering the drainage system, exacerbating the effects of climate change. Simple methods of analysis are likely to miss these effects and events assessed as severe based on today s data may become the norm in the future, increasing the number of emerging flooding incidents. If the at risk registers are to be comparable between companies and geographical areas, a consistent approach is required. That consistency does not exist at present. At this stage the list will be classified for internal or external flooding, severity, and an initial assessment of the return period. It forms the basis for reporting in the At Risk Register although it will be developed over time as more information becomes available. At this stage two issues influence the choice of register and lead to differences between companies: the approach to the assessment of return period and the extent to which repeat flooding takes account of severe weather. The true return period of a flood is not generally evident from a single event. Companies typically use either the estimated return period of the rainfall event as a surrogate for the flood return period, or add the property to the 1:20 register. The former approach allows properties to be added directly to the 2:10 register if the rainfall return period is estimated as less than 2:10 years and tends to increase the relative size of the 2:10 and 1:10 registers. Most companies will take account of customer statements that previous flooding has not been reported, which may lead to the property being entered to the 2:10 or 1:10 register instead of the 1:20. One company builds hydraulic network models to assess the return period of every internal flood event, and chooses the register on the result of that modelling rather than the rainfall return period. The treatment of severe weather in repeat flooding events differs between companies. Flooding events arising from severe weather are normally excluded from water company assessments, but this can cause confusion and dissatisfaction among customers for whom flooding is frequent and also occurs under normal weather conditions. One company includes repeat flooding under severe weather conditions in its frequency assessment, once it knows that the property has flooded under normal conditions. Another will retrospectively include previous flooding that happened under severe weather conditions, if a repeat flooding occurs under normal conditions. Both these approaches probably better reflect the actual flood risk, but they also increase the reported flood frequency and hence change the register choice, compared with the companies that do not adopt this approach. 4-7

The number of companies adopting each approach to the two issues is illustrated in Figure 4.5 Figure 4.5: Choice of registers for first-time flooding is reported, and treatment of severe weather Choice of Register on First Time Flooding 6 4 2 0 1:20 unless Rainfall return 1:20 unless evidence of previous flooding 1:20 register Rainfall return period Modelled Severe Weather and Choice of Register 10 8 6 4 2 0 Severe weather is Does not count severe Severe w eather is counted in repeats Severe w eather counted w hen previous flooding under normal condition Does not count severe w eather at all Best practice A rapid GSS payment to sewer flooding victims is commended, and whilst the difficulties of analysing extreme weather events are acknowledged, payments should be made as quickly as practicable. A range of techniques is applied to the assessment of the return period for sewer flooding incidents. No one method stands out as best practice, as the type and availability of data varies between companies. There is also inconsistency in the way that climate change and urban creep (a gradual increase in drained area) are accounted for in the assessments. In the absence of a consistent approach or clear best practice for identifying return periods, we recommend that properties should be added to the 1:20 register on first flooding, unless there is credible evidence that previous flooding has occurred but not been reported. Allocation to 1:10 or 2:10 registers should be on the basis of observed flood frequency. We recommend that the assessment of return periods should be subject of an industry-wide study to share information and establish consistent yet workable methodologies which companies can apply in practice in their areas. This will enable a consistent approach for allocation to be developed, when companies should then use the method to assign properties to registers. The benefit will be to ensure that severe weather is identified consistently, and that the selection of registers for repeat flooding is consistent between companies. We recommend clarification of the meaning and use of severe weather and exceptional weather. This would enable service to be described in the same terms for all customers, regardless of where they live. 4-8

4.3.2 After Detailed Investigation The approach to detailed investigation differs between companies and may be formed of one or several steps. Where hydraulic overload is suspected, it is common to carry out more detailed review and analysis of the network, including camera surveys and hydraulic modelling. This is commonly done to determine the most appropriate long-term solution to the problem and leads into capital programming. Companies reported issues with the use of cost-benefit analysis at this stage. In some cases it is difficult to justify capital intervention on cost-benefit terms and yet the companies believe they will be subject to criticism or legal action if a solution is not implemented. Many companies reported that some schemes are not truly cost-beneficial but that they believe they have a moral obligation to deliver a level of service. If the level of service is described as a social choice rather than an economic case, it is likely that benefits have not been expressed at their true value. Some companies commented that this valuation is especially difficult where willingness to pay is taken into account, as customers do not always value the service in line with its costs. On a straightforward economic assessment this could lead to an increase in unresolved sewer flooding until customers valued the service sufficiently to pay for it, but those companies discussing the subject felt that would be the wrong approach. The detailed investigation step identifies errors in classification and leads to updating of the At Risk Register categories, typically reducing the number of floods attributed to hydraulic overload. The changes result from wider surveys of the network identifying blockages and collapses, and from hydraulic modelling showing adequate capacity. Some companies carry out more detailed rainfall investigation at this stage, leading to further exclusions where severe weather caused the hydraulic overload. Some companies carry out additional investigation of neighbouring properties at this stage (although most do this either with the initial investigation or with the investment delivery). Searching for additional flooded properties often results in new additions to the register. Hence the timing of this investigation is important: if it is carried out early in the process it may affect OPA measures, but at the same time it improves the prioritisation of the register for investment and increases the benefit in cost-benefit assessment. 4.3.3 During Investment Planning and Delivery All companies carry out more detailed assessment in the design of capital solutions to flooding. The scope of this step depends on the extent of previous investigational steps. At this stage companies generally check for additional flooded properties in the neighbourhood, and those that have not done a substantial early search typically find between 20% and 60% additional properties. During our visits to companies, some commented that flooding events were becoming more localised, with only a few properties affected each time, rather than the larger groups benefiting from previous big schemes. This is consistent with a mature flood risk register, where companies have been prioritising the bigger problems for some years. We have not collected information on the number of properties per event, so cannot comment on whether this applies to all companies. 4-9

If companies are experiencing a diminishing number of properties/event, this makes it more difficult to make a cost-beneficial case for increasing the capacity of the network and more likely that mitigation measures will be proposed instead. Mitigation solutions might be effective for less than their design life if the underlying problem results from gradually increasing drained area or climate change, because increasing flow may render mitigation ineffective. In such situations more strategic investment would be indicated in order to prevent the problem returning and possibly affecting more properties. Risk assessments and the associated cost-benefit analysis therefore need to take account of the rate of change of demand so that appropriate long-term solutions are selected. The scope of this project did not include a review of company procedures for designing network investment and hence we are unable to comment on.examples of best practice. Any additional properties identified during pre-design investigations may be added to the At Risk Register in the year of delivery of the solution the same year in which they are removed. They add to the number of properties removed from the register as a result of company action, but do not support the output of net reduction in the At Risk Register. One company notes such additions separately and does not claim them in its outputs (although they would be included in the cost-benefit assessment), preferring to measure outputs in line with the original quantities assumed in its business plan (see Figure 4.6). The addition or removal of large numbers of properties from the register at this stage implies a poor quality of initial investigation and reduces confidence in any year s register values. Reporting progress of the properties originally on the register i.e. not allowing late additions to be counted within the company s outputs might provide a more balanced incentive for better investigation and addition of all flooded properties at the time of a flooding incident. But it might also encourage companies to add too many at the initial stage and lead to large numbers of removals through better information later in the process. Hence we recommend reviewing the links to other incentive mechanisms before changing the reporting requirements. Figure 4.6: Proportion of companies claiming late additions as outputs Adding Neighbouring Properties Claim outputs Don't claim outputs 4-10

4.3.4 Addition of Properties Estimated as being at Risk Most companies only add properties to the register when flooding has been experienced. One company adds nearby properties if on investigating a flood, it considers they are at risk of flooding in the future, even if they have not flooded at the time. Best practice Detailed investigations, including CCTV and network modelling are essential to confirm the initial investigation, assess the root cause of flooding and flood ownership and to inform assessment of the flood return period. This needs to be done as soon as possible after the flood, to ensure that all relevant information is captured whie it is available. Companies having good coverage of their sewer network with verified models are in a good position to undertake this quickly. The use of cost benefit analysis to justify expenditure on network enhancements to resolve flooding problems is a relatively new requirement and some companies are questioning it. We recommend a review of how companies have applied cost benefit analysis to sewer flooding problems following PR09 to share experience and best practice. This should extend to the use of CBA for mitigation schemes. If flood severity is introduced, (see above), then the value of mitigation in reducing flooding risks will be more visible. Some companies are reporting a diminishing trend in properties per flood event. It would be helpful if more facts were available on this to assess if it is an industry-wide experience. Late additions should be included in the analysis. In assessing the benefits of a scheme to enhance network capacity to resolve flooding problems, account should be taken of the effect of climate change and urban creep on future numbers of properties affected should nothing be done. It is not clear that this is being done consistently and thus benefits may be under-stated. The addition or removal of large numbers of properties from the register at the design stage implies a poor quality of initial investigation and reduces confidence in any year s register values. One company notes late additions at the design stage separately and does not claim them in its outputs (although they are included in its cost-benefit assessment), preferring to measure outputs in line with the original quantities assumed in its business plan. Extending this practice to the rest of the industry might provide a more balanced incentive for better investigation and addition of all flooded properties at the time of a flooding incident, when confidence in the data is more assured. 4.4 Keeping the Register Up to Date 4.4.1 The Link with Information Systems The quality and degree of integration of information systems varies significantly between companies. In some companies the information systems are highly integrated and the current status of the At Risk Register may be interrogated at any time. In others the systems are not integrated and the register is compiled off-line. We observed that companies with well integrated information systems tended to have most confidence in demonstrating their At Risk Register process, and that the results would compare with those in the June Return, although some companies with off-line processes also had good data management processes and archives of the registers. In the companies with integrated systems, the At Risk Register was seen as a simple report from the database, rather than a separate entity. 4-11

There did not appear to be a link between the approach to information management and the number of better information changes to the register entries over time. This suggests that these changes are a function of the quality of the investigation processes rather than the approach to data management. 4.4.2 Update Frequency Companies with integrated information systems can report the status of individual properties on the register at any time during the reporting year, although most companies review the status of flooding on a monthly basis. Companies without integrated information systems review new flooding that occurs in each month through internal reports, but the regulatory At Risk Register may be compiled on an annual basis. Companies generally aim to settle on a confident register allocation by the June Return reporting deadline for the year, but achieving this depends on the number of floods that occur in the last month of the reporting year. However, the extent of additions to, and removals from, the registers as a result of better information is significant in some companies, suggesting low confidence in the reported allocations in those companies. It would be expected that company confidence grades for the At Risk Register reports should reflect the proportion of retrospective changes that companies are making to their registers. 4.4.3 Recording Changes Most companies retain a full audit trail of the changes that are made to records of flooding and the progress of an individual property from entry to the register to ultimate removal. Companies having information systems that do not allow the full data history to be retained stated they were developing new systems that will retain a full audit trail. Best practice Integrated information systems provide greater confidence that data has not been lost and is not subject to human error from manual handling of data between off-line storage systems and applications and corporate systems. The status of the at risk registers should be reportable at any time and updated on a monthly basis at least. A full audit trail should be available for properties added and removed from the register. The extent of additions to and removals from the registers as a result of better information is significant for some companies. It might be expected that the confidence grades for the at risk register reports for these companies would reflect this. 4-12

4.5 Removing Properties from the Register 4.5.1 On First Contact If it is clear during the initial customer call that the flooding has arisen on a private pipe, as a result of a problem with that pipe, then the property will not be entered on the At Risk Register. 4.5.2 During the Clean-up Activity During the initial visit and clean-up, the causes other than hydraulic overload are commonly identified. These will be entered on the job record and the property will not be entered onto the At Risk Register. 4.5.3 After a long time without Flooding The Flood Estimation Handbook provides for removal of properties from the At Risk Register if they have not flooded for a pre-determined time. All companies were aware of the time-out rule but it is not in widespread use. Its most common use is in companies where properties assessed as being at risk were added to the register, but have not flooded, and one company uses a shorter approach than the normal 95% confidence duration. Although there are differences in approach, relatively few properties are removed from At Risk Registers using the time-out rule. 4.5.4 After Company Action The majority of properties are removed from At Risk Registers as a result of company action to improve network capacity. 4.5.5 As a Result of Better Information All companies remove some properties from the At Risk Register as a result of better information about the cause of flooding. In some cases this information only becomes available during preparation of capital solutions. Where large numbers of properties are removed through better information, it may be inferred that early investigations are insufficient to provide confidence in the entry onto the register. 4-13

5 Comments on the At Risk Register Approach Companies were invited to make their own suggestions of ways to improve the At Risk Registers. 5.1 Register Name The term At Risk Register may be associated with other flood risk assessments e.g. in surface water management plans, when the issue is a risk of flooding in extreme weather conditions such as 1:100 storms, during which the underground system is effectively irrelevant. A better title for the ARR might be Flooding Incident Register or Unresolved Flooding Incident Register ; the description would state that it s a register of properties that have flooded, and which are listed because they are at risk of flooding again if the circumstances are repeated. 5.2 Return Periods, Risk and Sub-registers Some companies commented that the current division of the register is not particularly useful. For example, there might be an overlap between customer perception of minor internal flooding (e.g. water coming up in a shower tray but not leaving the shower tray) and severe external flooding. Likewise the return period is hard to determine accurately and the exclusion of a flooding event for severe weather for a property that has already flooded in the past under normal conditions is distressing for some customers. An alternative suggested by some companies would be to list properties according to a measure of observed risk i.e. the frequency and severity. This would then be reported in categories such as high, medium, and low risk. The definitions for property owners would, for example, be that a high risk would be liable to flooding that causes significant disruption, on a regular basis. Other companies disagreed with that approach, suggesting that high would effectively be the same as internal 2:10. When sewer flooding is identified, an emergency job is started. Most companies use field service contractors and if internal flooding is being investigated the company s sewerage manager or supervisor will also attend. Companies stated that this demonstrates ownership to the customer. It was also mentioned that the company manager provides a point of contact for the customer after the cleanup is finished. All companies check the extent of flooding at this stage, and all said that for internal flooding they knock on neighbours doors if they believe the flooding might affect those properties. Some companies have exclusion rules, which provide guidance for staff on what to class as flooding. These rules apply to external flooding and include puddle size, puddle depth, time to clean up. Since Ofwat amended its reporting guidelines to include any escape from sewers, most companies have discontinued these exclusions. However, where there is no exclusion rule, companies agreed that staff have to use their discretion at the time of the visit. The extent of this discretion is likely to influence the number of external flooding incidents that are reported. 5.3 GSS Payments All companies reported that except in the case of wide area flooding, it is generally difficult to identify the flooding return period at the time of the customer call or initial investigation. Therefore GSS payments tend to be authorised as soon as the site visit confirms that the flooding has taken place. This means that companies are often paying more in GSS payments than they would have to if they were able to make a strict application of the rules. Where a strict application of the rules is applied, companies find it difficult to complete within the 20 day limit. 5-1

5.4 Assigning Properties to the Register Companies do not all follow the same approach to assigning properties to registers and this affects the number that are added to each register. Some of the main differences are described below: 5.4.1 First Time Flooding Ofwat s June return reporting guidelines state: When a previously unreported property or external area is flooded, it should normally be considered to be at risk and added to the 1 in 20 category unless: - investigation clearly shows that it is at risk of flooding more frequently than once in ten years, when it should be included in the one in ten year category; - investigation clearly shows that it is at risk of flooding more frequently than twice in ten years, when it should be included in the twice in ten year category; - the storm was exceptionally severe and investigation shows that it is clearly not at risk of flooding as frequently as once in twenty years and the severity of the storm can be verified (e.g. by the Meteorological Office); or - the cause was a blockage, etc The most common approach used by companies is to add first-time flooding to the 1:10 register, but one company adds them to the 1:20 register, and one uses the rainfall return period to determine the register. Another conducts a full hydraulic analysis of the local network response to rainfall and thus derives a modelled return period for the flood event. These differences affect the proportion of properties on different registers rather than the total number reported. The use of rainfall data leads to many additions directly to the 2:10 register. Using rainfall data alone will add more properties directly to the 2 in 10 register simply because there are many more rainfall events at 2 in 10 year return period than 1 in 10 year return period. However the effect on the sewer network should be considered when allocating risk. It is not clear why a sewer network would become overloaded from a 2 in 10 year return period storm. 5.4.2 Repeat Flooding All companies review a property s position on the register when a repeat flooding incident occurs. The period between flooding incidents is frequently used to determine if a property should be moved to a higher risk category. Some companies mentioned that it is not clear how they should handle multiple flooding when some instances are caused by severe weather. Some companies ignore the severity of the weather once it has been identified that flooding occurs under normal conditions. In other companies, repeat flooding is not counted if it is the result of severe weather. Companies commented that the latter approach is difficult to explain to customers and can lead to an underestimate of the risk of flooding. 5-2

6 Recommended Process We set out the approaches approach to managing DG5 registers using a common generic process flow-chart that allows comparisons to be made. From our assessment of company processes and taking a comparative view, we have suggested a best practice approach in Figures 6 to 11. The six flow-charts cover the following areas of practice: Identification of a flooding event from first contact with the customer. Initial investigation of a flooding event, establishing the cause and the number of properties affected. Making decisions on GSS payments. Adding properties to the at risk registers. Undertaking detailed investigations to inform the design of capital solutions. Processes for removing properties from the registers. We have categorised processes as preferred (green), non-preferred (grey) and optional (yellow). The latter applies particularly to the assessment of flood-event return period, where a number of techniques are used according to the availability of local data and weather analysis. We believe that this is an area of developing practice that is worthy of industry-wide study to share information and establish consistent yet workable methodologies that companies can apply in practice in their areas. Commentary on the recommended processes and the potential impact of their adoption on the at risk registers is given after each figure. 6-1

Figure 6.7: Link between flowcharts Chart A: Identification Chart A: Identification Chart B: Initial Investigation Chart B: Initial Investigation Chart D: Adding to Register Chart D: Adding to Register Chart C: GSS Chart C: GSS Chart E: Detailed Investigation Chart E: Detailed Investigation Chart F: Removal from Register Chart F: Removal from Register 6-2

Figure 6.1: Best practice process flow chart identification Chart A: Identification Record details of all calls Call Handling Can see asset register Manual Scripting Check if private sewer Yes Case Based Reasoning End Can see GIS maps No Ask about severity of flooding Ask about location of flooding Legend Check if repeat flooding Preferred process Optional processes Non-preferred process Ask if neighbours affected Check if it is raining Automatic creation of work order Manual creation of work order Telephone customer to confirm visit time At same time as first contact Dispatch investigation/clean-up team 6-3

Principal points: Case-based reasoning preferred to manual scripts. Access to GIS and asset register helps call-handlers identify private sewers. Simple questions establish nature and extent of problem early. Automatic set-up of work-order using integrated IT preferred to manual methods. Inform customer of expected time of arrival of clean-up team on first contact. Impact on ARRs: Limited direct impact, but integrated systems likely to lead to fewer data losses and improved audit trail for subsequent processes. 6-4

Figure 6.2: Best practice process flow chart initial investigation Chart B: Initial investigation Telephone customer to confirm visit time Unless already done on first contact Supervisor or engineer attends as well as field contractors Confirm not private sewer Clean up and repair damage Internal CCTV External Identify cause End Blockage/Collapse Hydraulic Overload Equipment failure End Exclude surcharges Exclude nonsewer flooding Exclusions (none) Exclude non-integral garages Exclude 1 damp patches (internal) Complete Work Order (WO) 1= Where samples cannot be taken and patches may arise from groundwater Check all neighbours potentially flooded Check neighbours only if evidence of flooding Leave forms if unoccupied Additional properties added to new WO Additional properties added to same WO Validate WO entries Update WMS Update via paper form Write to customer Pay GSS 2 2 = For companies who pay GSS without reference to severe weather analysis 6-5

Principal points: Supervisor attends as well as clean-up team/contractors to ensure that procedures are followed. CCTV undertaken for all flooding events unless information already available. Exclude non-sewer flooding (eg surface-water flooding) where responsibility for resolving the problem lies wholly with others. Exclude damp patches if samples cannot be taken and cause could be groundwater. Undertake detailed investigation of neighbouring properties affected, including those in the area that are unoccupied. The intention is to capture and report this information once and it is not expected to be amended months later through better information. Add any properties subsequently confirmed to have been affected to the same work order record to ensure clean data audit. Validate work-order entries through robust quality control procedures and update digital records (not via paper forms). Make GSS payments where they apply, unless it is an exceptional wide-area flooding event clearly resulting from severe weather. Inform customer in writing of the findings from the initial investigation and the company s proposed actions. Impact on ARRs: Reporting all sewer-flooding incidents may increase the number of properties on the registers for those companies applying exclusion rules in the past. Emphasising the need to capture information on properties affected at the initial investigation stage should reduce the number of late additions through better information and provide more confidence that the ARRs are correct and up to date. Validating work-order entries through quality control procedures and application of integrated IT will ensure improved data audit trails. This should be reflected in improved data confidence grades. Early payment of GSS payments is encouraged as a goodwill gesture, accepting that some customers may be over-paid by the strict interpretation of the extreme-weather rules. Some companies already do this. It is only proper that companies write to customers to inform them of the findings from the initial investigation and the company s proposed actions to reassure them that their problem is being addressed. It is acknowledged that this may raise expectations that a capital solution or mitigation measure will be delivered when the company may not always be able to make a cost-beneficial case for doing so. 6-6

Figure 6.3: Best practice process flow chart GSS payments Chart C: GSS Flood confirmed by field team Automatic ID GSS Manual ID GSS High resolution radar data Pay if < 1:10 Weather Analysis Check if extreme weather Pay if < 1:20 Aligned with severe weather Met Office data Confirm GSS payment Pay if < 1:40 Check GSS paid (audit) Write to customer confirming actions Principal points: Automatic identification of customers potentially eligible for GSS payments linked to the flooding incident report and work management systems ensures that eligible customers are not missed, which is a risk with manual systems. A range of techniques for estimating the return period of the flooding event can be applied, according to available data. Best practice in this area needs to be defined through separate study. The criterion for extreme weather should be reviewed to consider alignment with that for severe weather used to assess the inclusion of properties on the at risk register. The criterion proposed is that adopted by most companies, which is 1 in 20 years return period. GSS payments do not apply if the flooding event is more extreme than this. Companies that have adopted a higher standard than this in the past (e.g. 1 in 40 years) may opt to continue to apply this to GSS payments (though not to the ARR, to ensure national consistency). 6-7

Impact on ARRs: Clarification of the severe weather definition to rarer than 1 in 20 years will affect the number of properties on registers for those companies that have adopted a different standard in the past. Figure 6.4: Best practice process flow chart Adding to the register Chart D: Adding to Register 1:20 Check if severe weather 1:40 Does customer report previous flooding? YES NO First time flooding If confirmed by company records Accept Ignore If not confirmed Anecdotal evidence from customers flooded in the past informs register Consistent rules determine register Rainfall intensity determines register Event specific network analysis determines register Model output for standard storm events informs register Add neighbouring properties assessed as being at risk From results of initial investigations Add to register 2:10 1:10 1:20 Keep record of severe weather flooding Repeat flooding Anecdotal evidence from customers flooded in the past informs register Interval determines register Rainfall intensity determines register Event specific network analysis determine register Model output for standard storm events informs register Severe weather is counted in repeats Severe weather not counted in repeats 6-8

Principal points: A >1 in 20 year return period is used to check for severe weather, in accordance with Ofwat s guidance. Information from customers about previous flooding events can be accepted although there should ideally be evidence to corroborate the information. It should be clear that it resulted from hydraulic overload and not other causes. Anecdotal evidence from customers of past flooding should not otherwise be considered in determining the appropriate register to place the property. A consistent set of rules should be applied when assessing the return period of a flooding event to decide which register to record the affected properties against. This should not be determined by the return period of the storm alone. The decision may be informed by the results from verified network models from simulating standard storm events. Such information may already be readily available. Companies may also wish to model the storm event using verified models to try to replicate the flooding event. This is more onerous and may not be justified for all flooding events. Most first-time flooding events are added to the 1 in 20 years register, unless there is compelling reason why it should be added to the other registers. Records should be kept of severe weather flooding events, even though the effected properties are not added to the register. For repeat flooding, the interval between flooding events informs which register to allocate the properties to. This may be supported with network modelling. Where one of the flooding events results from severe weather, the severe weather event should count towards the assessment of the register allocation. This represents a change to Ofwat s guidance (which excludes all events arising from severe weather) but aligns better with customer perceptions. Impact on ARRs: The use by all companies of a 1 in 20 year return period for assessing the occurrence of severe weather will bring consistency in the number of properties excluded from the ARR for severe weather. Disallowing anecdotal evidence of past flooding from customers unless company records support it should reduce the number of new additions for some companies; effectively, if customer claims it was flooded, but there is no record on the DG5 register of past event, then assume it must be due to other causes or it never happened. Focusing on the return period of the flooding event, rather than that of the storm, will probably result in a reduction in the number of first-time flooding events to the 1 in 20 years and other registers for some companies. Including severe weather events in the assessment of repeat flooding may increase the number of properties on the 2 in 10 and 1 in 10 registers slightly for some companies, though by definition, severe weather events are rare. 6-9

Figure 6.5: Best practice process flow chart detailed investigation Chart E: Detailed investigation Detailed rainfall analysis Check severe weather To confirm analysis undertaken when properties were added to the register and to assess future weather patterns CCTV To obtain additional information to that gathered for the initial investigation where required Hydraulic modelling To confirm analysis undertaken when properties were added to the register and to assess the impact of future weather patterns on solution design Confirm cause Further checking of neighbouring properties optional. This should have been established at the initial investigation stage Check neighbouring properties door to door Check neighbouring properties letter survey Claim additional properties in outputs Update register Principal points: The purpose of the detailed investigations is to capture additional information to that captured at the initial investigation stage to inform the development of capital solutions to flooding problems. Further assessment of severe weather determines future weather patterns as well as current to inform the design and cost-benefit assessment. Additional CCTV surveys (if required) and network modelling studies are undertaken to confirm the initial analysis and to assess the impact of future weather patterns and growth in flood volumes on proposed solutions. Further checking for neighbouring properties affected can be undertaken at this stage, but this is expected to have been completed at the initial investigation stage. Late additions should not be allowed in outputs, which should reflect the original register values, unless there are compelling reasons to the contrary. 6-10

Impact on ARRs: The emphasis on identifying all affected properties at the initial investigation stage and only allowing the claiming of late additions in outputs by exception should reduce the number of late additions through better information. Figure 6.6: Best practice process flow chart removal from register Chart F Removal Full audit trail of decision to remove Remove as a result of company action Separate register for mitigation Demote via time-out rule Remove via time-out rule Better Information Remove if strong evidence that analysis undertaken when properties were added to the register was inaccurate (eg discovered to be other causes). This should only occur on rare occasions End 6-11

Principal points: Full audit trail of decisions to remove properties from the register. Generally, properties are only removed as a result of company action. Removal for better information remains possible, but should only occur on rare occasions. Separate register kept for properties protected through mitigation measures. Large numbers of properties would not normally be removed through time-out rules. Impact on ARRs: A separate register for mitigation measures will not affect numbers on the ARR, but will record the benefits that such measures bring. 6-12

7 Summary of Recommendations 7.1 Short-term recommendations Clarification of some aspects of the Reporting Requirements would help to improve the comparability of company registers. Areas for consideration should include: o The definition of, and approach to identifying, severe weather. Differences in approach can affect whether flooded properties are added to the register, or excluded on the grounds of severe weather. There are three aspects to this; clarifying the definition of severe weather, strengthening the guidance on best practice methods of assessing severe weather events and ensuring that all companies report using the same definition and applying best practice methods of assessment. o Clarifying the definition of internal flooding, especially for cellars and garages. o Clarifying the approach to flood return period analysis and the addition of first-time flooding to the At Risk Registers. o Clarifying the approach to assessing repeat flooding, especially where some instances are the result of severe weather and others are not. o Clarifying the approach for recording properties that are discovered at a much later date, (through detailed investigations or anecdotal evidence from customers), that are likely to have been flooded. Current practice varies between companies as to whether such properties are added to the register and count towards outputs. o Separating the monitoring of properties that have mitigation measures installed, to better reflect their true risk of flooding. The name At risk register is confusing and it should be named the Flooding incident register or \Unresolved flooding incident register. The definitions of severe weather (sewer flooding) and extreme weather (GSS) should be aligned, to a 1:20 year return period. When a property floods under severe weather and normal weather conditions, the severe weather flooding should be included in cost-benefit analysis (if it has a return period that should be met by normal design standards), as this better reflects the customer experience and hence the benefit of investment. The definition of sewer flooding should be strengthened with guidance on what is included and excluded. This would reduce the range of interpretation and help to provide a more consistent service between sewage company regions. Companies should not exclude minor flooding from the sewer flooding register (for example through the use of exclusion rules or judgement). Proper checks should be in place to ensure that all flooding is recorded. Full checks for flooding of neighbouring properties should be made at the time of the initial investigation. Although it is appropriate to add more neighbouring properties through better information at the time of investment, the addition of significant numbers of properties at that 7-1

stage suggests that companies are not making adequate checks at the time the flooding occurs, and may indicate poor service to some customers who experience flooding. A common approach to the assessment of flood return periods (or flood frequency if a full risk scale is used) is required. His should ideally be developed by an industry-wide study or by consensus of the sewerage companies. When assigning June Return confidence grades, companies should take account of the number of additions to and removals from the ARR. (Where there is a significant addition or removal rate through better information, companies cannot have high confidence in the original data.) 7.2 Longer term recommendations The flood frequency (2:10, 1:10, 1:20) and severity (internal, external, other flooded areas) categories should be replaced with a risk scale taking account of the frequency and severity of flooding. This would be better aligned with the benefit-cost analysis now common in planning solutions. For reporting purposes, risk would be reported in categories. If a risk scale is adopted, the impact of mitigation on the flood risk should then be taken into account and used to re-assess the risk once mitigation is in place. This would provide clearer information to customers on the impact of mitigation. 7-2