Assessing future flood risk at BGS and NERC observatory sites : summary report

UK Research and Innovation (UKRI) recognises the problems posed by climate change, its impact on society, and the need for positive action to address the environmental sustainability challenges we now face. By 2040, UKRI aspires to be ‘net-zero’ for its entire research undertaking, which includes reducing and mitigating all carbon emissions from UKRI owned operations (UKRI, 2020). Surface water flooding can cause disruption to people’s daily activities, businesses, and societal functioning, consequently increasing the pressure on natural resources. UKRI aims to understand the risk of flooding to its properties to act where possible to enhance climate resilience. This Summary Report describes work undertaken by the British Geological Survey (BGS) in partnership with the Natural Environment Research Council (NERC) to investigate the risk of flooding to the BGS Keyworth and BGS Edinburgh sites, and to four NERC observatory sites (at Capel Dewi, Eskdalemuir, Hartland, and Herstmonceux). Flood risk was assessed under both ‘current’ and ‘future’ climate conditions. After reviewing existing assessments of the risk of flooding at these locations, additional flood analyses and modelling were undertaken for the sites that have been mapped as being at risk of fluvial or pluvial flooding. These sites are BGS Keyworth, BGS Edinburgh, and the National Centre for Atmospheric Science (NCAS) Capel Dewi Atmospheric Observatory (CDAO). This report summarises the findings from the analyses and hydraulic modelling studies of the three sites. It is accompanied by a second report, which provides more detailed technical information (Nagheli et al., 2022). Flooding due to direct heavy rainfall (pluvial flooding) or due to overflowing surface water features (fluvial flooding) could cause water to inundate areas of the sites investigated, potentially resulting in business disruption and damage to infrastructure. The risk of this is assessed by evaluating whether a feature would be affected by surface water or not, and if so, how often it would be expected. The UKCEH Flood Estimation Handbook (Institute of Hydrology, 1999) methodology was used to obtain profiles of rainfall over time for design storms (see Glossary). The ReFH2 software (the Revitalised Flood Hydrograph rainfall-runoff method version 2; Kjeldsen, 2006) was used to estimate the corresponding surface runoff hydrographs for catchments above points of interest. The HEC-RAS flood modelling software (US Army Corps of Engineers, 2022) was used to simulate fluvial flooding. The SWMM modelling software (Storm Water Management Model; US EPA. 2022) was used to simulate pluvial flooding and to assess the capacity of drainage infrastructure (for BGS Keyworth only). The assessment of how flood risk will change in the future makes use of climate change ‘uplift’ factors. These factors have been used to shift historical design storms. Uplift factors have been estimated using the latest UK Met Office Hadley Centre climate projections—the UKCP18 projections—by the UKRI-funded FUTUREDRAINAGE project (Chan et al., 2021). Factors are only available for a ‘worst case’ atmospheric greenhouse gas concentration trajectory (referred to as a Representative Concentration Pathway or RCP)—the RCP8.5 pathway. Based on these uplift factors, Table 1 summarises how flood risk at each of the sites is predicted by the modelling to change between the historical period (1961–1990) and the two future time horizons considered: the 2050s (2041–2060) and the 2070s (2061–2080). The following findings and recommendations (see also Appendix 2) are presented for the three sites considered: BGS Keyworth • The site is not at risk of flooding from rainfallrunoff causing the water level within the channels running along the north-west and north-east of the site to rise and inundate parts of the site. • The critical storm duration (see Glossary for definition) for BGS Keyworth was calculated to be seven hours. • There are three culverts in the channel along the north-west of the site. If we adjust the historical 7-hour duration, 100-year return period summer storm to account for climate change, then the modelling indicates that the culverts in the drainage channel along the north-west of the site will surcharge but not result in inundation of any parts of the site. (Summer and winter storms are treated separately statistically by flood hydrologists because summer storms are more intense). • Considering the same storm as described in the previous bullet, then if it is assumed that the bottom half of the culverts become blocked, the modelling predicts that the Platt Lane entrance to the site will be inundated by approximately 20 cm of water. No other part of the site would be affected. • Again, considering a 7-hour storm with a return period of 100 years (calculated using data for the period 1981–2020), analysis of the UKCP18 climate projections for RCP8.5 suggests that the frequency of this event will change to: » 1 in 20 years over the period 2021–2040 » 1 in 10 years over the period 2061–2080 • BGS facilities team should inspect the culverts at least annually and arrange for any debris to be cleared by the appropriate authority, if necessary. • BGS should make Nottinghamshire County Council, the Lead Local Flood Authority (LLFA) for Keyworth, aware of this work, given the potential vulnerability to flooding of the new homes recently built on the northern side of Platt Lane, and of Severn Trent Water’s sewage pumping station at the corner of Platt Lane and Nicker Hill. • There has not been sufficient information about the site’s drainage network to assess the risk of water appearing on the ground surface when the drainage network becomes surcharged. Furthermore, the development of a model to do this would be a complex task. Consequently, we have modelled the capacity of the subsurface drainage pipes and used this as a proxy to indicate which parts of the system are more likely to cause water to pond on the surface. Those pipe sections that have been simulated to surcharge, or exceed 90% of their capacity, during a 30-minute storm, need further investigation. The model simulates that 6% of the network’s pipes exceed 90% of their capacity during a 30-minute, 10-year return period storm, which increases to 9% during a 30-minute, 75-year return period storm. First, the slopes and lengths of the problematic network sections should be measured accurately, and the modelling exercise repeated to confirm the findings of this study. Updating and rerunning of the model would be relatively quick. After confirming the fidelity of the model, several potential solutions could then be reviewed, and their costs and benefits evaluated against the level of risk that NERC BGS are willing to accept. Solutions could include replacing small diameter pipes with larger pipes, increasing the slopes of the pipes, optimising the size of catchment areas generating runoff by altering the direction of surface flow paths/directions. It is important to maintain the drainage infrastructure to avoid surcharging of the network and flooding. BGS Edinburgh • The levee and flood gates constructed along the Murray Burn in 2020 have enhanced the protection of the Lyell Centre. However, our modelling predicts that the Lyell Centre would still be affected by flood water under a 20-year return period storm. We conclude that the levee is not sufficiently high at its downstream end and, based on our new drone-based LIDAR survey of land surface elevations, flood water overtopping the levee here flows towards the Lyell Centre. If it is considered that the degree of flood protection is currently insufficient, we recommend that NERC and Heriot Watt University discuss what the options are for increasing the level of protection to the Lyell Centre. For example, this could include extending the levee downstream and increasing its height, or potentially increasing the crosssectional area of the channel. • The critical storm duration for BGS Edinburgh was calculated to be seven hours. Considering a 7-hour storm with a return period of 100 years (calculated using data for the period 1981-2020), analysis of the UKCP18 climate projections for RCP8.5 suggests that the frequency of this event will change to: » 1 in 20 years over the period 2021–2040 » 1 in 7.1 years over the period 2061–2080 • Our modelling has shown the potential for flooding of other buildings on the Heriot Watt campus, e.g. the Energy Academy and the buildings north-east of the Lyell Centre on the opposite side of the Murray Burn and Research Avenue South. This report should be shared with the Heriot-Watt estate management department to make them aware of the risks to the occupiers of these buildings, and to allow them to consider any necessary actions. NCAS Capel Dewi Atmospheric Observatory (CDAO) • The south-east corner of the site was flooded on 21 January 2018. Measurements of rainfall every 10 minutes during this day have been made available by the CDAO’s Project Scientist. Comparison against long-term historical observations of rainfall has indicated that the design storm that most closely matches the peak rainfall intensity and total rainfall of the observed storm has a 7-hour duration and 30- year return period. • Land surface elevation data for the site are only available on a relatively coarse, 5 m grid. Because of this, there is significant uncertainty about the cross-sectional shape, and slope, of the Afon Peithyll, which flows east to west along the south of the site. The results of the modelling must, therefore, be considered as ‘indicative’. • For a 7-hour, 30-year return period design storm the current model simulates flooding that was more extensive than that observed in January 2018. However, it does indicate the area of the facility that is at higher risk—the south-east and east of the site, which is consistent with the observations. • Simulation of the influence of the culvert (approximately 300 m downstream of the site) and whether it is partially blocked or not, suggests that it has little impact on the flood risk of the site. • The critical storm duration for the site was calculated to be four hours. The modelling suggests that a 4-hour storm with a return period of seven years will initiate out of bank flooding at the south-east corner of the site. • Considering a 4-hour storm with a return period of 100 years (calculated using data for the period 1981–2020), analysis of the UKCP18 climate projections for RCP8.5 suggests that the frequency of this event will change to: » 1 in 20 years over the period 2021–2040 » 1 in 10 years over the period 2061–2080 • A survey of the Afon Peithyll and its floodplain is needed to define the dimensions and slope of the channel accurately and improve confidence in the model. • A number of engineering options are listed that could be considered to protect the site from flooding; their viability would depend on the characteristics of the site, cost, and possible environmental impacts. • Consideration could be given to the feasibility, and costs and benefits of moving infrastructure located in the south-east of the site, where flood risk is higher, to another part of the site