Wilby (2012) concludes that ‘pluvial studies generally report greater increases to multi-day precipitation totals, and proportionately greater changes to extreme single-day events’. As urban drainage systems (piped and surface storage services) and pumped catchments have a fixed capacity to accommodate pluvial events, any increase in the severity of rainfall events and run-off will bring into question the ability of these systems to cope. Changes to the duration, sequence and spatial extent of rainfall (including towards less intense long-duration rainfall) will also influence antecedent conditions, run-off and the moisture content of structures and green spaces and hence the design performance of the drainage network (Dunne and Black, 1970). In both cases, the higher rates of surface run-off may increase debris recruitment (leaves, wood and anthropogenic debris) and subsequent transport and blockage of surface drains (Streftaris et al., 2012).

Understanding the changes in the variability of the climate (e.g. the potential for extreme storms) is perhaps more important to river infrastructure than changes in mean flows. This is because changes in the morphology of the catchment is most responsive to individual storm events and any associated increased erosion (including scour around bridges, embankments, bends, etc.) or accretion (leading to loss of channel section or blockage) can undermine the performance of FCERMi. Persistently saturated soils also act to decrease the stability of embankments, and issues were highlighted as a concern during the winter of 2013/14. These issues are discussed in more detail below.

Despite few infrastructure having the sole purpose of managing groundwater floods, groundwater flows are an important consideration in understanding the performance of raised defences. For example, higher than design groundwater flows can (i) bypass a raised defence and flood the land behind (Macdonald et al., 2012), (ii) exacerbate scour (Loveless et al., 1996), (iii) drive progressive erosion and piping of the embankment or foundation soils (Schweckendiek et al., 2014), and (iv) destabilize soil slopes and cliffs, increasing the chance of a catastrophic slip (Iverson and Major, 1986). In urban areas, recent discussions have also focused on the relationship between increased groundwater levels and ingress in to piped drainage systems by means of below-ground pathways (but limited evidence exists).

During extended periods of lower-than-average rainfall, low groundwater levels can lead to differential settlement and resulting instability (Wols and van Thienen, 2014), with significant impacts on urban infrastructure, including FCERMi assets (Foster, 2001). Lower groundwater levels at the coast can also lead to saline intrusion, exacerbating the corrosion of engineered infrastructure and reducing the natural infrastructure capital of coastal freshwater water and brackish lagoons (Hiscock et al., 2011).

Although the interaction between groundwater and climate processes is poorly understood (Taylor et al., 2013). There is a general consensus that groundwater levels tend to be slow to respond to driving rainfall conditions. Any changes in the temporal sequencing and spatial coherence of rainfall events are therefore likely to be important.

Infrastructure at the coast has perhaps the greatest sensitivity to climate change of all FCERMi. Sea level rise (the strongest of climate change signals) acts to reduce the depth-limiting effect of near-shore waves (Sutherland and Woolf, 2002; UKMMAS, 2010). In turn, this leads to increased overtopping and the potential for larger wave impact forces (and subsequent structural damage and increased breach potential). Larger waves (or more persistent storms) are also likely to drive coastal morphology change and, particularly where backshores are constrained, lower beach level, further exacerbating the impact of sea level rise. Over the medium to long term, any growth in offshore wave heights is therefore likely to be expressed at the coast (Hall et al., 2006). These issues are discussed in more detail below.

Concrete structures, like other infrastructure, are likely to deteriorate faster if they experience more frequent and extreme periods of freeze-thaw (Auld et al., 2007; Environment Agency, 2013). Prolonged hot dry periods are likely to accelerate desiccation of surface soils on earth embankments (Figure 1). Extreme hot and cold temperatures can act to restrict or even stop mechanical and electrical assets from operating (Rowan et al., 2013; Sayers and Dawson, 2014).

Increasing working with natural processes to help manage flood risk (in a way that protects, restores and emulates the natural regulating function of catchments, rivers, floodplains and coasts) is being recognised as a legitimate and important component of FCERMi (Demuzere et al., 2014; POST, 2011). Central to the idea is working with the river and coastal processes (and flooding) rather than against them. ‘Soft path’ infrastructure (e.g. green roofs, wetland storage, shelter belts, urban ponds, floodplain reconnection) and ‘hard path’ infrastructure measures (e.g. bypass channels, controlled storage) are both important aspects of modern flood management and, in many cases, if managed well, offer some degree of natural resilience to change. This does not mean their performance will be unchanged by climate change. High temperatures and drought can, for example, influence the performance of ‘soft path’ measures, significantly reducing the ability for infiltration, altering the mix of the vegetation and/or encouraging the formation of standing water and associated undesirable outcomes such as disease or increased mosquito population (Armitage et al., 2012).

Although often overlooked, the vegetation, microbes and nutrients present within marine and freshwater systems are important components of the FCERMi system. Vegetation within watercourses needs to be managed to maintain conveyance and avoid blockage; marine vegetation can provide important buffers against erosion at the coast, and nutrients and microbes can attack concrete and steel structures (Gu et al., 2011). For example, accelerated low water corrosion (ALWC, the attack of concrete and steel structures by nutrients and microbes in the marine and estuarial environment) is an important influence on the performance of flood defence structures (Melchers, 2014). Infrastructure in tidal and brackish water, such as the Thames Estuary, are particular susceptible to ALWC and can experience rates of corrosion exceeding 1 mm/side/year (CIRIA, 2005), a rate that is expected to increase with higher temperatures (Stewart et al., 2011).

Conveyance of river channels, afflux at structures and the stability of flood defences can also be influenced by invasive species such as Japanese Knotweed (Defra, 2013). The preferential growth and survival of such species can be influenced by their adaptation to conditions of high temperatures or drought. Internationally, climate change has been associated with the potential increase in more aggressive, non-native, animal burrowers that undermine the stability of flood defences, although there is currently no evidence to suggest this is occurring in the UK.

Retrofitting or modifying infrastructure at a range of scales will be an important feature of future modifications. The Adaptation Sub-Committee (ASC) (2011a) noted that ‘buildings are a priority area for adaptation, because decisions concerning the design, construction and renovation of buildings are long lasting and may be costly to reverse’. Over longer timeframes, it is realistic to retrofit at the landscape scale, not just upsizing traditional urban drainage systems but transforming the built and rural environments to be more water sensitive. This might include a range of measures that mimic or enhance natural processes, including permeable paving, green infrastructure, small-scale storage ponds, and using urban features such as streets and parks as temporary flood pathways. Individually, these measures may offer limited benefits, but given sufficient change over long periods, their cumulative benefits should become clear.

Various simple examples of adaptable design exist, for example, purchasing land in the lee of an embankment to facilitate future raising or widening, or designing foundations that anticipate a heightened embankment in the future (see Figure 3). Such options often demand greater upfront expenditure than perhaps would be the case if future change had been ignored; there are seldom, true win-win situations. Flexible solutions are, however, likely to be more cost-effective over the longer term. For example, beach nourishment is often promoted as a flexible solution in that the amount of fill placed on the beach can be modified from one nourishment campaign to the next, in the light of improving understanding of beach behaviour and changing objectives with respect to risk reduction.

Figure 3 Adaptive design keeps future options open without incurring unnecessary additional expenditure (source: Sayers et al., 2012b).

There is, of course, a close connection between flexibility and robustness. Flexible solutions tend to be robust to uncertainty as they are able to be adapted to perform well under a wide range of possible conditions, assuming that future adaptation decisions are taken in a timely way. Existing FCERMi appraisal guidance (Environment Agency, 2014) is in line with the Treasury Green Book (HM Treasury, 2011) and does include some advice about accounting for future climatic change and uncertainty. Decision makers and planners have been slow to take up these approaches; therefore, practical methods for decision tree analysis and the application of real options analysis have started to emerge (e.g. Woodward et al., 2013). Despite this, there remain few examples where FCERMi has been purposefully designed to facilitate future modification in response to observed changes in climate, and significant engineering innovation will be required to make a real transformation in infrastructure provision.

Typically, FCERMi will be designed to protect against a single ‘design storm’ (which may be described in terms of storm surge level, high flow rate, or rainfall duration and intensity). The 2007 floods in the UK highlighted that a single spatially coherent event could affect large parts of the country simultaneously with severe knock-on impacts for supply chains and critical service provision (Pitt, 2007). The 2013/14 winter floods further exposed inadequacies in this existing approach, highlighting the fundamental difference between a ‘single event’ and a prolonged sequence of events upon an increasingly saturated land and progressively weakened FCERMi (as witnessed in the collapse of the main railway line in Dawlish in 2014).

Uncertainties about future climate are factored in by addition of a precautionary allowance onto the design-loading conditions (Environment Agency, Undated). For example, moving gate weirs (such as radial and buck gate weirs) are normally designed so that the underside of the gate(s) can be raised above an anticipated peak water level. Changes in spatial coherence and temporal sequencing of storms events, although widely recognised as important, are not a standard consideration in design. Appropriate infrastructure choices, and design details, will only start to emerge as the full richness and potential impacts of climate change are recognised. This will include recognising that the single design storm is ‘dead’.

Local authorities face difficult trade-offs when planning future development. The cost to the local economy of constraining development in areas at risk from flooding or erosion (now or in the future) can be significant. Often, either opposing demands to develop brownfield sites or the lack of alternative sites mean there is little choice. However, planning for the long term is crucial. FCERMi is often a central feature in Catchment Flood Management Plans, Shoreline Management Plans and Strategies, but it is unclear if these studies are sufficiently innovative and sufficiently influential to modify local authority development plans. ASC (2011a) noted that despite there being evidence of long-term, strategic planning for adaptation (e.g. Shoreline Management Plans), it was unclear how influential these initiatives were on local development plan policies and actual development decisions. They also found limited evidence that local authorities were factoring in long-term costs when making decisions on the strategic location of new development in their Local Plan. Local authorities should take a strategic approach to managing vulnerability at the scale of communities as well as at the property level. This will require explicitly weighing up the long-term costs of climate impacts against social and economic benefits from development that are more immediately released.

FCERMi is not constructed for its own intrinsic value but to provide wider benefits to society through the protection it provides and the functions it supports. As such, FCERMi is inextricably linked with ecosystem services (Sayers et al., 2015), the provision of critical infrastructure and associated services (e.g. ICT, energy, health). The impact of a failure within FCERMi can extend far beyond the footprint of the physical floodwaters. Understanding the spatial cascade of impacts through infrastructure networks (Brown and Dawson, 2013) and the economy (Crawford-Brown et al., 2013), and the measures implemented to manage these risks, should be considered to make a legitimate contribution to FCERMi.

Property owners can act to increase the demand they place upon system-scale infrastructure or reduce it (Defra, 2012). For example, by improving their resilience to future changes in climate (by installing property-scale resilience measures or making provision to protect the owner business functions from flooding or erosion risks), they can also impact positively on the infrastructure upon which they depend. A review by the ASC (2011a) highlights that there is limited evidence of the uptake of adaptation measures in the retrofit or repair of existing properties, despite a number of reasonably low-cost measures for existing buildings. Householders and developers require the right incentives to take action. The ASC’s second progress report found instances where there is either a lack of or misaligned incentives, both of which lead to an inefficient adaptation outcome. Levers other than regulation, such as insurance incentives and better information, may be more important for existing homes (ASC, 2011b).

New approaches to the design and construction of flood defences are seeking to provide multiple uses and benefits (Anvarifar et al., 2013; Kundzewicz, 1999). Crucial to FCERMi is the role of soft infrastructure, such as green spaces in urban areas, wetlands, channels, beaches and mudflats. Such features provide a wide range of benefits, including habitat creation and aesthetic value. These systems make important contributions to the management of flood and erosion risks (e.g. Sayers et al., 2015). Uncertainties associated with their performance have, in part, been responsible for hindering their uptake as it is difficult to reliably compare them against more traditional engineered infrastructure. This will need to change if we are to maximise the multiple benefits and inherent adaptability such systems provide.

Similarly, there are a wide range of propositions for redesigning houses and towns (Casey, 2012) from raising houses onto stilts through to floating artificial island communities. These measures all intend to help society ‘live with water’. Pervasive sensors and the move towards real-time monitoring of condition and loads could provide a significant contribution to more targeted maintenance and emergency response (Pengel et al., 2013). Ultimately, this might help promote the update of adaptive management and lead away from a bias towards design conservatism and precautionary allowances that can lead to maladaptations.