Mapping Flood Disadvantage in Scotland 2015: Methodology Report

5. Indicators of flood hazard-exposure

The Flood Risk Management (Scotland) Act 2009 (FRM Act) requires the publication of High (10-year return period), Medium (200-year return period) and Low (1000-year return period) probability flood hazard information. SEPA published new national Flood Maps for these scenarios for coastal, river (fluvial) and surface water flooding (incorporating rainfall and sewer model outputs) in 2013. The Flood Maps were based on the consistent application of modelling methodologies across Scotland.

The SEPA flood hazard maps were developed for a suite of flood scenarios ranging from high probability to low probability events and considering climate change. In addition, they considered 'defended' and 'undefended' scenarios. A 'defended' scenario is where the underlying models specifically consider formal flood protection schemes (e.g. walls, embankments). This information is taken from the Scottish Flood Defence Asset Database (SFDAD) which is the best source of information on defence structures. However, the SFDAD is incomplete and the information varies in quality across Scotland. Furthermore, certain necessary assumptions are applied in the application of defence data, appropriate to the strategic nature of the models. Consequently, the defended runs do not provide a precise, explicit assessment of flooding at each defence location. Threfore, a specific assessment of a particular defence structure is likely to indicate a more accurate reflection of the flood pattern and protection offered.

An 'undefended' scenario is one that does not explicitly consider formal flood protection structures. Both scenarios will include implicit reference to other structures (e.g. road embankments) where they are components of the underlying Digital Terrain Model (DTM).

Coastal, river, and surface water flooding were all considered separately in this assessment, in addition to flooding from any of these sources. This study uses the SEPA v1.1 Flood Maps, which were updated and made publicly available in March 2015. The mapping of flooding is a dynamic process and the Flood Maps are subject to change as input data, methodologies and techniques are developed. Thus, the Flood Maps are subject to continual improvement and the study represents the current knowledge on flood hazard as of March 2015. It may not account for more recent changes or fully reflect flood hazard in certain locations (e.g. Forres and Elgin).

In consultation with the Steering Group, three return periods were selected for each source of flooding to provide a range of outcomes of the probability of flooding (Table 4). The return periods covered 1 in 25 or 1 in 30 years (4% or 3.3% annual probability); and 1:200 years likelihood of flooding, together with a low scenario which incorporates consideration of future climate change projections. For surface water flooding, the depth of 0.1 metres was considered, since even shallow water can cause significant damages and repair costs, thus making it difficult for people to recover after flooding (Kazmierczak and Cavan, 2011). The defended extents were used to include the presence of flood defences.

Analysis of the Flood Maps (v1.1) revealed some inconsistencies between the extents for different return periods, for example, in some areas, the 1 in 200 year extents did not incorporate all of the 1 in 25 year extents, as would be expected (the lower the probability of flooding, the greater its magnitude and the larger the area affected). Table 5 presents the number of data zones affected by these inconsistencies. Given the inherent uncertainty in flood modelling (i.e. resulting from simplifications necessary to reflect highly complex natural processes) there are areas of the country where the methods work more effectively than others. In these other areas (e.g. for very small, urban watercourses at lower return period scenarios) there is therefore relatively lower confidence in the outputs. For the statutory scenarios, SEPA sought to make amendments in areas of lower confidence prior to publication. However, further work is required to address some of the intermediate return periods not published but used in other areas of flood risk management (e.g. 30, 50, 100-year return periods). Thus there may remain locations where there are inconsistencies between the published return periods and those other scenarios.

SEPA is considering these inconsistencies in the improvement plan for the Flood Maps so that anomalies and inconsistencies between scenarios will be addressed and overall confidence improved with more focussed modelling assessments. In this assessment, we considered return periods for a given type of flooding seperately to avoid making assumptions, for which of the return periods the flood extents were more accurate. The 'any type of flooding, 1 in 200 years incuding the impacts of climate change' flood extent accounts to some extent for the inconsiencies by combining all flood extents (all types of flooding, all return periods).

Table 4: Flood hazard maps used in the project (SEPA, version 1.1, March 2015)

*Depth 0.1m, incorporating rainfall and sewer model outputs

Table 5. Number of data zones affected by the inconsistencies in flood risk maps

The index of flood hazard-exposure represents the percentage of residential addresses exposed to flooding in each data zone. Calculating the percentage of residential addresses within the flood extents presents a more accurate account of the proportion of the population exposed than simply using the percentage of the data zone area exposed to flooding, because the extent of flooding may include unpopulated areas. Furthermore, whilst Flood Maps are of a strategic nature and not appropriate for property level assessment, they are intended for use at a community level, and therefore, presenting our findings at the data zone level is more appropriate.

Residential addresses were obtained from Ordnance Survey AddressBase, supplied by the Scottish Government (version: 11 April 2015). OS AddressBase classifies each address into residential or commercial (Figure 3). Only residential addresses were considered in this assessment, and therefore, as previously noted, it is different to the NFRA, which also considers flood risk to commercial properties.

Processing conducted in ArcGIS v10.2 included Select by Location to calculate the number of residential properties within each flood extent, and a Spatial Join to summarise the information by data zone. Table 6 shows the number and percentage of residential addresses at risk from all types of flooding and extents for the whole of Scotland.

Figure 3. AddressBase over OS MasterMap Topography Layer (OS, 2015)

Table 6: Percentage of residential addresses exposed to flooding in Scotland (total number of residential addresses: 2475709*)

* It should be noted that an additional 1465 address points were located outside of the Data zone extents. These properties have not been considered in this assessment. It is likely that they are mobile properties such as boats, but this level of information was not available in OS AddressBase (only AddressBase Plus or Premium products).

The percentage of residential address points within each flood extent, and all extents combined (any flooding) was calculated for each data zone. For each of the extents separately, and for all data zones (containing residential addresses, see section 3) the percentage of address points potentially exposed to flooding was standardised to develop the hazard-exposure indicator (see Figure 1).

The standardised percentage of residential addresses within all extents combined (thus residential properties exposed to any flooding) has been used to calculate the main flood disadvantage index.