• Sign Up or Sign In

Understanding the spatial risk from lead

Statistical and GIS-based €˜hotspot' techniques offer the best route to the implementation of a risk-based lead strategy, says Oliver Parsons-Baker, asset management consultant, Black & Veatch

Figure 1: Visualisation outputs of point and kernel density analysisFigure 1: Visualisation outputs of point and kernel density analysis

Revision of the drinking water standard for lead, from 25µg/l to 10µg/l in December 2013 is forecast to have a significant impact on water company key performance indicators for water quality. With this change on the horizon, the cost of removing all lead within distribution systems and the uncertainty over achieving compliance using some current strategies, there is an even greater need for utilities to build on measures to mitigate lead risk developed for AMP5.

Together with the continued maintenance of supply system-wide plumbosolvency control, future intervention measures for lead will potentially require focusing resources on the most vulnerable people in communities where the risk from lead is highest. Ultimately risk reduction and improved compliance is likely to be achieved via a wide ranging Integrated Management Strategy that includes a number of tailored options outlined in Table 1.

Implementation of more quantitative, risk-focussed lead ‘hotspot’ techniques has the potential to deliver a number of clear benefits, especially for companies looking to maximise the power of available datasets for decision making support. These are:

  • Cost-effective, auditable, transparent, repeatable, rapid and analytically robust approach to identifying lead ‘hotspots’
  • Greater evidence-based and visual approach to AMP6 lead risk assessment and investment planning

‘Hotspots’ or ‘hotspot areas’ are spatial clusters of point events, which appear over time, within a limited geographical area. Hotspot analysis aims to assist in the identification of locations with unusually high concentrations of point events.

Commonly with hotspot analysis it is hard to identify clear boundaries due to the continuity of point distribution.

This will be the case in some water supply areas, especially where sampling data is sparse or phosphate dosing has been introduced. In reality it will be a gradient, continuous over an area, being higher in some parts and lower in others.

However, using hotspots to analyse and reduce large data sets into meaningful subgroups of events remains an important part of any risk-based approach to lead. There are dozens of techniques for spatial point pattern analysis.

Suitable statistical and GIS-based techniques for hotspot analysis of lead are overviewed below. It is recommended that these techniques are examined using segregated assessments for pre- and post-phosphate dosing to understand the effect that this control measure has on hotspot delineation and outputs.

Nearest-Neighbour Hierarchical Clustering (Nnh)

  • Nnh clustering identifies groups of events that are spatially close
  • Nnh defines a threshold distance and compares the threshold to the distances for all pairs of points. Only points that are closer to one or more other points than the threshold distance are selected for clustering
  • The routine then conducts subsequent clustering to produce a hierarchy of clusters
  • Nnh allows different strategies to be adopted within an Integrated Management Strategy for Lead and provides a coherent way of approaching ‘hotspots’

Multi-Distance Spatial Cluster Analysis (MDSCA)

  • MDSCA is a scan-type of clustering algorithm which searches for, and identifies, the densest clusters of events based on the scattering of points on a map
  • The ability of the user to control the approximate size of the clusters allows a broad search for ‘hotspot’ areas across, say, a whole city - and a second search concentrating on a smaller area to derive more focused ‘hotspots’

Point and Kernel Density Analysis

  • The uneven distribution and nature of point datasets such as lead sampling data means density analysis can provide a powerful tool for the visualisation of ‘hotspots’ through methods such as surface ‘heatmaps’ (Figure 1)
  • Point and Kernel techniques take known quantities and spread them across the landscape and so provide density estimates for all parts of a region. The data can be displayed either by surface or contour maps that show the intensity at all locations
  • The main advantage of these surface representations is that they can provide a more realistic continuous model of lead ‘hotspot’ patterns, reflecting the changes in density as a result of the use of lead in building practices over time
  • An additional benefit of point and kernel density analysis is ‘coldspot’ analysis through highlighting areas of low lead sampling density normalised by population. The results could be used to direct operational sampling programmes for lead
  • Black & Veatch has found density analysis to be the most valued by clients because of the clear visualisation the surface representations/’heatmaps’ provide. However, some other techniques may provide better longevity because of the less intensive useof GIS.

Combining clustering approaches and overlaying outputs such as density-based ‘heatmaps’ with Nnh ellipses can also be completed to increase the robustness of any clustering approach.

Due to the advantages and limitations of different techniques, an incremental and phased approach to the development of the hotspot analysis can be taken. Outputs from different techniques can then be reviewed and fed into the development of a finalised approach.

The quantitative hotspot analysis techniques above can help prioritise areas for the implementation of intervention measures and also provide the justification to regulators that efforts are targeted at highest ‘risk’.

However, the heavy reliance on datasets such as compliance results within the analysis does not necessarily adequately demonstrate risk when multiple factors are involved. Nor is it likely to be a good predictor of future performance in areas where sampling or asset surveys have not been undertaken.

The development and generation of spatial lead risk models, used for other water quality parameters such as discolouration, can help overcome data confidence and scarcity issues.

The spatial risk models are based on a combination of assessing:

  • Past performance: data such as customer-requested lead pipe replacement, opportunistic lead communication pipe replacement, asset surveys and long-term regulatory sampling data are reviewed to assess the past susceptibility for, and spatial distribution of, lead
  • Causal mechanisms: the future vulnerability of each area is assessed by considering the various potential factors related to lead through the development of relationships based on the analysis of indicator data

The factors involved in the above processes are assessed separately using indicator data and then the probabilities combined to predict the overall vulnerability. This is managed by the analysis of characteristics related to lead in drinking water and associated environmental data which, together with company work order information, can be used to produce lead failure rates for all communication pipes, properties or areas.

These failure rates can then be used to create cohorts which result in similar outcomes with regards to lead concentrations. The cohorts can then be applied company-wide and used to predict the likelihood and location of potentially elevated compliance risks, or health issues, if combined with geo-demographic segmentation data. Hence, the analysis can have a greater predictive capability and extend to areas or properties where sampling has not been undertaken.

The benefits include the ability to identify, and then group, high risk areas on sensible spatial scales for the implementation of an Integrated Management Strategy. Remedial measures could include collaborative lead pipe replacement at specific streets and a communication campaign across a localised area.

The core visualisation aspect of spatial risk ‘heat maps’ can also form key tools to communicate and engage a wide stakeholder base, a key area encouraged by the Drinking Water Inspectorate to support tackling the issue of exposure to lead pipes.

Enhancing lead strategy

Spatial risk assessment models can also be designed to integrate with the cost benefit model developed as part of the recent UKWIR project Alternatives to Phosphate for Plumbosolvency Control. The combined use of these tools can further develop and enhance AMP6 lead strategy and provide the clear, auditable evidence-base required by regulators for PR14 submissions. Significant economic and logistical barriers exist for the complete removal of lead within distribution systems and internal pipe work.

These, together with problems associated with division of responsibility for the supply pipe between water companies and property owners, mean that compliance and health risks associated with lead in drinking water will continue for many decades yet.

While CIWEM recently called for the development of a national strategy for the removal of lead pipes, in the short/medium-term water companies will continue to be required to demonstrate due diligence in achieving compliance with the lead standard and in tacking the health risk of exposure to lead. The techniques detailed in this article provide water companies with a range of sustainable, Drinking Water Safety Plan aligned hotspot and lead risk assessment approaches as a basis for the development and implementation of an evidenced based Integrated Management Strategy for AMP6 and beyond.


Sign up today for your daily news alert and weekly roundup

© Faversham House Group Ltd 2018. WWT and WET News news articles may be copied or forwarded for individual use only. No other reproduction or distribution is permitted without prior written consent.

Cookie Policy   |   Privacy Policy