BT2 Road network at risk of flooding

Three separate factors were considered in the development of these metrics:

1. The source of flooding – fluvial, pluvial and coastal
2. The social and economic value of the road network (with trunk roads used as a proxy for  value) 
3. The vulnerability of communities relying on the network (with rurality as a proxy for vulnerability)

Separate figures are given for a) the road length directly at risk of flooding and b) the road network that would be affected by such flooding.  The latter helps to account for the networked nature of road infrastructure in a more meaningful way. A 0.15m flood depth threshold is used throughout as a measure of the depth at which driving may become unsafe.

Current figures are given in Table 1, and Figure 1 shows the road network in areas at risk of flooding, both for the road network as a whole (dark blue) and for the trunk road network only (light blue).

Table 1: Proportion of road network at risk of flooding (1:200 year event)

Figure 1 Proportion of road network located in areas at risk of flooding

Notes: Flood risk assessed for a 0.5% probability (1:200 year) flood event. Dark blue columns represent the road network as a whole; light blue columns represent the trunk road network only. The figure shows the both the directly at risk sections, and the sections that would be affected / impacted by this flooding

Figures 2 and 3 show the road network at risk of flooding for two example areas: the Central Belt and Lochaber and Mull. The figures show how flood risks to road infrastructure are spread across the areas depicted, reflecting the spatial distribution of watercourses and their catchments (fluvial flood risk), roads and other less permeable features (pluvial flood risk) and low lying coastal areas (coastal flood risk).

Figure 2 Distribution of fluvial, pluvial and coastal flood risk to the whole road network: Central Belt

Note: Trunk roads are shown as red lines. The map shows sections of the network at flood risk where flood depth is modelled to be 0.15m or greater.

Figure 3 Distribution of fluvial, pluvial and coastal flood risk to the road network:  Lochaber and Mull

Note: Trunk roads are shown as red lines. The map shows sections of the network at risk of flooding where flood depth is modelled to be 0.15m or greater.

Risk to the road network in very remote rural areas

Road closures due to flooding in very remote rural areas warrant separate consideration. This is because communities in rural locations may be more vulnerable to road closures as rural roads are generally less networked – i.e. there are fewer alternative routes. In line with SEPA (undated), the use of very remote rural areas provides a proxy assessment of vulnerability, whereby vulnerability is construed as the rurality of the road’s location^[2]. (Note that the definition of ‘very remote rural areas’ is taken from the Scottish Government 8-fold urban rural classification (Scottish Government, 2014)). The road network in very remote rural areas is shown in Figure 4.

Figure 4 The road network in very remote rural parts of Scotland

Figure 5 shows the extent of flood risk to the road network in very remote rural parts of Scotland, while Figure 6 shows its spatial distribution. The network in these parts of Scotland has similar levels of exposure to fluvial and pluvial flood risk.  It is notable that the greatest risk is posed by pluvial source flooding which has the potential to impact 42.4% of the trunk road network.

Figure 5 Road network within very remote rural Scotland located in areas at risk of flooding

Note: Dark blue columns represent the road network as a whole; light blue columns represent the trunk road network only. The figure shows the both the proportion of directly affected sections, and the proportion of the network that would be directly affected (impacted) by this flooding

Figure 6 Road network at risk of flooding in very remote rural parts of Scotland – spatial distribution of flood risk to the road network as a whole

Note: Trunk roads are shown as red lines. The map shows the road network at risk of flooding where flood depth is modelled to be 0.15m or greater.

The UK Climate Change Risk Assessment (HR Wallingford et al, 2012a; Thornes et al, 2012) assessed changes in flood risk to road infrastructure as a result of anticipated climate changes. Whilst this assessment was only undertaken for England and Wales (due to data availability) it provides a broad indication of what might happen in Scotland in the future given anticipated climate changes. Given this, transport specific aspects from the UK CCRA (ibid) have been used in conjunction with general aspects from the Scotland CCRA (HR Wallingford et al, 2012b) to understand how the risk of road closures from flooding might change in the future.

The Scotland CCRA (ibid) considered climate projections and impacts on biophysical systems to undertake a climate change risk assessment for Scotland. Climate change in Scotland is expected to result in increased winter rainfall and potentially increased incidence of intense summer rainfall events. In terms of impacts on biophysical systems, these projected climate changes are expected to lead to greater volumes of runoff, higher river flows and increased catchment wetness / waterlogging problems, particularly during winter months. These factors are expected to cause increased flooding from both rivers and surface water (note that that this could also be caused by an increase in intense summer rainfall events). Clearly the scale and magnitude of current and potential future flood risk depends on the configuration of the road network relative to floodplains and low lying coastal areas as well as the flood alleviation / road design measures in place, all of which will vary between the devolved countries, meaning that the data for England and Wales from the UK CCRA below can only be used as a broad indication of possible changes.

The CCRA for England and Wales indicated that the projected length of road at significant likelihood of flooding (where significant is defined as a 1.3% annual probability) would be between 13,000km and 16,000km by 2020 compared with a baseline of about 12,000km (Thornes et al, 2012). This equates to increases of between 8.3% and 33.3% from baseline with the range reflecting the different climate change scenarios considered in the assessment. The CCRA also highlighted how in addition to an increase in the overall length of infrastructure that could be affected, the frequency of flooding of infrastructure that is already located in the floodplain is expected to increase also (ibid). These projections do not account for any actions that could be taken to alleviate flood risk to roads (e.g. flood defence infrastructure, enhanced maintenance regimes) which may help to reduce overall flood risk (noting that these types of actions are picked-up in related indicator BT6). In summary the following projected changes are anticipated to take place in Scotland in the future:

• The proportion of the road network located in areas at risk of flooding is projected to increase.
• Road infrastructure that is already located in the floodplain is expected to be affected by flooding more frequently.
• Increased incidence of intense rainfall events may result in more frequent pluvial (surface water) flooding

There are several key limitations to the assessment as summarised below:

1. Roads in the floodplain are often raised above the ground surface on embankments. The difference in elevation afforded by these embankments is not always identified in flood modelling and mapping (Thornes et al, 2012). As such it may be the case that flood risk to road infrastructure is over estimated – i.e. where the embankment would raise the road out of the inundated area. This is not reflected in the modelling due to the granularity of SEPA’s flood hazard modelling process.
2. Future consideration of road closures are based on English and Welsh data from the UK CCRA, which only provides a broad indication of possible future risks, and so should be interpreted with caution when applied to Scotland.
3. This assessment is based on an early version of data that was generated by SEPA’s Baseline Appraisal flood hazard modelling Appraisal (SEPA, 2015^b).  The approach to and assumptions underpinning this model may be subject to refinement and changes in the future.
4. The SEPA Baseline Appraisal (ibid) data that underpins the assessment includes some inaccuracies when classifying roads. As part of the proxy assessment of the value element of risk, a search term^[1] is used in all trunk road specific metrics to extract Baseline Appraisal outputs that relate to trunk roads only (see Table 4). The Baseline Appraisal’s classification of trunk roads is not entirely in agreement with Transport Scotland’s official list of trunk roads (Transport Scotland, 2015) meaning that the assessment for trunk roads only is a slight over estimate.
5. SEPA’s flood hazard modelling and the associated Baseline Appraisal data (SEPA, 2015) does not capture flood duration – i.e. in effect it presents a ‘snapshot’ of flood events based on flood extent, depth and velocity criteria. Whilst road infrastructure is susceptible to pluvial (surface water) flooding in that it is only designed to accommodate low return period / high probability flood events (5 year events) without flooding of the carriageway, drainage design for roads is such that surface water flooding should dissipate quickly, meaning therefore that flood duration should be relatively small (Highways Agency et al, 2004). This characteristic of flooding is not captured in SEPA’s flood hazard modelling and data and, as such, the assessment of pluvial metrics may paint a ‘worse picture’ than is actually the case.
6. The social and economic effects of flooding are not fully reflected in this assessment which would require a more in-depth understanding of the duration of the event (see Point 4 above),  the number vehicles affected by a flood event, the journey purpose of those affected, the availability and the length of any diversion route amongst other factors.
7. This assessment was only undertaken for 0.5% probability (1:200 year) modelled flood events.  These are medium probability events.  Higher probability events (e.g. 1:10, 1:50) can be expected to affect a smaller extent of the network than 1:200 year events but on a more frequent basis.  It should also be noted that the flood hazard modelling is based on historic data.  As a result of climate change, the magnitude of a 1:200 year event will almost certainly be greater than represented in this analysis.