Studying effects of climate change
UKWIR's Peter Spillet and Brian Wilkinson, HR Wallingford's Richard Kellagher and MWH's David Balmforth explain how changing weather patterns can effect sewer systems performance
Although there is still considerable uncertainty over the magnitude of future climate change, it is generally accepted future rainfall patterns in the UK are likely to be significantly different from the present day (Kerr et al, 1999, Futter and Lang, 2001). Changes in climate will differ from one part of the UK to another and climate will be less predictable year on year.
Winters are likely to become wetter overall and although summers are predicted to become drier, the less frequent rain events are likely to be much more severe, with greater rainfall intensity. Since rainfall is the primary input to sewerage systems, it follows that increased flooding and increased spills from combined sewer overflows (CSOs) can be expected.
Increases in mean sea level will effect tide locked outfalls and decreases in summer river base flows will worsen impacts from continuous and intermittent discharges. Over the past two and a half years a team from HR Wallingford, the Meteorological Office, MWH and Imperial College have been evaluating the potential effects of climate change on the performance or sewerage systems, under a research project for UK Water Industry Research (UKWIR Project CL10). This project aimed to:
Climate predictions were undertaken by the Meteorological Office's Hadley Centre and were based on the UKCIP 1998 scenarios. The medium-high emissions scenario was chosen as a conservative estimate, but recognising any of the four emissions scenarios was equally likely. Future rainfall amounts for the year 2080, for varying return period and duration, were predicted and compared with present day rainfall. The results were expressed as simple ratios and represented as scaling contours on maps of the British Isles.
In all, 49 such maps were produced by HR Wallingford, based on the Hadley Centre's climate modelling (Dale, Gallani and Hollis, 2002). These maps allow current design rainfall to be rescaled so an engineer can predict the likely future effects of climate change on sewerage system performance. The maps show the greatest changes are in south-east England and in Scotland, but that the patterns of change are quite different for different rainfall events.
A stochastic rainfall generator tool was built by Imperial College (Onof et al, 2002 a and b) to generate 100-year 5min time series suitable for use in sewer network modelling. The series was used to analyse CSO spill frequency and volume and assess receiving water impacts. The time series was also used to predict changes in infiltration and treatment works flow.Sewer Systems
To analyse the impact of climate change, five drainage areas were chosen to represent a variety of drainage types, topography and size. Three were inland and two coastal.
Four climate regions were chosen to represent the range of climate conditions across the UK. Climate regions were combined in different ways with the catchments. For example, for the flooding analysis all four regions were tested with each of the five catchments, making 20 combinations. This allowed the effect of catchment characteristics to be separated from climate effects in the analysis.
For flooding analysis the ten-year and 30-year return period design storms were used with the duration and season chosen to give the worse case for each catchment (Balmforth, 2002). For the coastal impact analysis only the two coastal catchments were used. A representative ten-year time series record was selected from the 100-year synthetic series generated for each climate location.
This was used to generate annual and bathing season CSO spill frequencies and volume. A similar analysis was undertaken on the three inland catchments to generate annual spill frequency and volume from the CSOs. In addition, a representative two-year rainfall series was selected and used in a full UPM analysis on one of the inland catchments using fundamental intermittent standards (FIS) to analyse river impact. The impact analysis was repeated, with the river base flow reduced by 20% to demonstrate this separate effect of climate change.
Finally the effects of climate change on sewer dry weather flow was analysed using regression models from treatment works' flow gauges in two different catchments (one urban and a second urban/rural). The present day and future ten-year records were then used to simulate dry weather performance and estimate its effect on treatment works flows.Coastal CSOs
Figure 1 shows the increase in spill volume and frequency for coastal CSOs in the two coastal catchments. Two sets of results are given - for the annual series to test shellfish waters compliance and for the bathing season to test bathing waters compliance.
Storage solutions were developed to meet current shellfish and bathing waters spill frequency standards for present day and 2080 rainfall series. The results of these are shown in Figure 2.
Figure 2 shows the storage provision to meet the ten spills per annum for shellfish waters and three spills per bathing season for bathing waters. For increasing rainfall, the required storage is proportionally very much greater than the proportional increase in rainfall. For example, a 20% increase in rainfall generates an increase of around seven-fold in storage requirement. The scatter is because catchment and sewer network characteristics have some influence on CSO discharge. A regression analysis for increasing rainfall yielded the correlation equation:
The results show significant investment would be needed to meet spill frequency standards for both shellfish and bathing waters. It may therefore be more sustainable in the longer term to move to an impact assessment approach to coastal CSO discharges.Inland CSOs
A similar analysis of inland CSOs again showed using CSO spill frequency as a surrogate for receiving water impact tends to over-assess the consequential effects of climate change. To understand the potential impact more fully a full UPM impact assessment approach was used with one of the inland catchments.
A representative two-year rainfall time series was selected for the analysis. FIS were used for the assessment. Figure 3 shows the results of the assessment. Also on Figure 3 are the results of a similar assessment but with the river base flow reduced by 20% to allow for the potential of climate change on this factor. The results show changes in rainfall due to climate change have a commensurate effect on future storage requirements. However, by analysing the actual impact, the disproportionate effect of trying to manage spill frequency is avoided.
Storage requirements lie within a manageable range similar to present day volumes. Of greater significance is the potential effect of climate change on river base flow. Here, a 20% reduction in river base flow, leading to a reduction in assimilative capacity, accounts for increases in storage requirements from 25-100%.Scaling Up
Having determined the likely effects of change on the test catchments, the research team was then faced with the challenge of scaling up the impacts to UK values. Various approaches were attempted. At first it was thought the most promising approach would be to scale up on the basis of length of sewer or population. However, it proved impossible to find any meaningful trend between the observed effects and these parameters. It was finally decided the most reliable method for scaling results up to area-wide values was on the basis of proportional increase and climate region.
The four climate regions were overlaid on the respective sewerage undertakers areas and an average value for scale up was determined, based on the regression equations 1 and 2 above. The scale up accounted for population distribution, which was found only to be significant for Thames Water's area. Towards the end of the project the UKCIP 2002 climate modelling results became available. Using an approximate method of scaling based on the ten-year, six-hour rainfall, a new set of scaling factors was produced for each sewerage undertakers' area.
These factors showed changes of between -15% and +11%, compared with the earlier model, though this averaging process masked a greater degree of local and seasonal variation between the two climate models. The results of the study have major implications for the UK water industry. The study has produced:
- rainfall maps which enable engineers to determine design rainfall events that allow for the effects of climate change,
- a stochastic rainfall generator that can produce time series rainfall data for future climates, suitable for sewer network modelling. The research has also shown that:
- for many areas of the UK, climate change may result in an increase in rainfall depths in excess of 1.4 x current values with a subsequent doubling of flood frequency and volume,
- as a result of this increase in rainfall, storage volumes to prevent internal property flooding may need to be increased by more than two-fold,
- climate location has a much greater influence on flooding performance than drainage area type or size,
- increases in storage volume to meet spill frequency standards for coastal CSOs for the 2080 scenario are substantial with up to a six-fold increase for shellfish waters and a seven-fold increase for bathing waters,
- water quality analysis indicates minimal impact on receiving waters in the south (assuming no change in river base flow) with the likelihood of impacts increasing to the north,
- the effect of climate change in reducing river base flows is likely to have a significant additional detrimental effect on river quality requiring addit-ional provision of around 25-100% storage volume for a 20% reduction in river flow for example,
- changes to the infiltration from groundwater into sewers are closely related to the change in average annual rainfall. The results show the effects due to climate change are minimal and the resulting impact on WwTWs inflows small