DEM spatial resolution –
what does this mean
for flood modellers?

It’s widely accepted that the digital elevation model (DEM) data, or terrain data, used for flood modelling significantly influences the flood data results. But what role does the DEM data’s spatial resolution serve in creating good flood maps? Is it simply, “the higher the resolution the better”? What other considerations can tip the scales for or against the use of high resolution? Such weighty questions are often discussed amongst those immersed in the flood modelling and risk management industries. 

Through scientific literature and JBA flood modelling expertise, this blog considers some cost-benefits of using higher or lower resolution terrain data, when lower resolution can be helpful and what advantages investing in higher resolution affords.

What is spatial resolution?

DEM data are used in flood models to represent a physical land-surface. The spatial resolution of a DEM refers to the area of land being represented by a single grid cell. So, a spatial resolution of 10 metres means one grid cell is representing a 10 x 10 metre area of physical land. Low (or coarse) resolution and high (or fine) resolution are relative terms. ‘Higher’ resolution implies comparatively greater preservation of land features, while a ‘lower’ resolution dataset generally smooths over topographic details (Environmental Systems Research Institute, Inc., 2016).

DEM resolutions used in the hydrological and hydraulic modelling industry typically range from 1000m to 2m or less (Vaze, Teng, & Spencer, 2010). Here, we limit our discussion to the resolution of JBA’s national-scale flood maps, which range from 30m- 5m.

Figure 1: Example of higher (left) to lower (right) spatial resolution representing the same land area polygon.

The spatial resolution trade-off

Flood modellers face plenty of important decisions throughout the flood modelling process. We study complex natural processes and investigate the assumptions used to represent them; we source, vet and assimilate input data; we design and test models and critically analyse the results. Central to this applied and uncertain science is the DEM. Good quality DEM data are important because we use these to estimate how water interacts with the environment and identify where flooding is most likely to pose challenges to people and properties. The accuracy of water depth predictions is linked to the accuracy and spatial resolution of the DEM (Vaze, Teng, & Spencer, 2010).

Yet, although higher spatial resolution and accuracy often translate to improved results, these benefits cannot be considered in isolation, especially when we consider how numerical flood models work. Numerical models rely on multiple simulations to represent the physical processes of flooding and these simulations can number into the hundreds of thousands or more depending on an area’s extent. The more complex models are (i.e., the higher the spatial resolution), the greater the demands on computational power, processing time, file size, data storage and cost (Environmental Systems Research Institute, Inc., 2016). On the other hand, using comparatively lower resolution data means lower feature spatial accuracy but faster processing and smaller file sizes. Thus, depending on a client's objectives and requirements, lower resolution may fit the bill.

The benefits of 30m resolution

These trade-offs highlight the advantages that lower resolution data can offer. Consider this analogy: in a standard health check-up, most people accept insightful, albeit generalised, information about their health, such as heart rate, blood pressure and joint reactivity. In a short amount of time, key information is summarised. However, if the results triggered an alarm regarding a person’s blood pressure, reasonably he would seek out more detailed, timely and costly – i.e., better resolution- analyses for that specific health concern.

In the same way, flood modellers seek to use the level of complexity reasonable for the task at hand. JBA’s suite of flood maps includes global national-scale maps at 30m resolution. These national-scale flood maps enable users to compare flood data at different locations across countries or continents for a consistent view of river, surface water, and in some regions, coastal flooding. They can be used to indicate areas that may benefit from a more detailed flood risk assessment based on higher resolution data, guiding high-level decision making and balancing the requirements of the task at hand with the complexities mentioned above.

Figure 2: Higher resolution data (left) define flow paths of water better. Lower resolution data (right) help highlight flood susceptible areas that may benefit from further analysis. 

The benefits of higher resolution

The higher the DEM resolution, the greater the preservation of the topographical terrain features. This provides better definition of the floodplain, small streams, roads and other narrow conduits of flow which can significantly impact results. While the availability of high-resolution terrain data is on the rise, it’s still not available everywhere. Often these data are limited to countries and regions with high concentrations of population and/or economic development. However, where there is a client and market need for higher resolution flood data and good quality and quantity of data, it is appropriate to invest in these.

JBA provides high-resolution flood maps for the UK and Republic of Ireland, Continental Europe and the US at 5m resolution. Figure 3 illustrates the impact using higher resolution data has had on flood maps for downtown Miami. Although both images indicate the general areas at risk from flooding, the 5m map captures movement of water in complex urban areas which is not captured in the 30m map. The validation of our modelling outputs with information about historical and current flood events confirms an improved, more detailed view of flooding from high-resolution maps.

For more information on our 5m US Flood Map, see our blog Why does the US need another flood map? 

Figure 3: JBA 30m flood map (left) vs JBA 5m flood map (right) for downtown Miami, where flooding in narrow walkways is more realistically represented in the 5m map.

Achieving the right balance

Modelling flood is a complex science and it's the combination of a variety of factors that enable us to meet the different requirements of clients. While resolution is not the only metric for determining the quality of a dataset, in geographical areas where we are confident in the quality and quantity of the input data, using higher resolution resources has made an important difference to assessing flood risk.

Ultimately, the most informed decisions can be made by those who have access to the right expertise. Speak to us about the range of data we provide, how it can be used, bespoke consultancy options we offer and identifying the best option for you.


Environmental Systems Research Institute, Inc. (2016). Cell size of raster data. Retrieved from ArcMap:

Saksena, S., & Merwade, V. (2015, November). Incorporating the Effect of DEM Resolution and Accuracy for Improved Flood Inundation Mapping. Journal of Hydrology, 530, 180-194.

Vaze, J., Teng, J., & Spencer, G. (2010). Impact of DEM Accuracy and Resolution on Topographic Indices. Environmental Modelling and Software, 25, 1086-1098.

News &

News International Women’s Day 2024

JBA women talk about their achievements, support for other women in the workplace, and ideas to #InspireInclusion for International Women's Day 2024.

Learn more
News Mind the physical-risk due diligence gap

A failure to use good quality data and sophisticated climate change intelligence to understand the impact of flood on physical assets could be putting investors' portfolios at risk.

Continue reading
News JBA seals flood data partnership with Old Mutual Insure for South Africa

JBA and Old Mutual Insure have signed a new deal, providing them access to our cutting-edge flood maps for South Africa.

Continue reading
Blog Modelling and Uncertainty - Climate Sensitivity

JBA shares insights from its climate change experts about uncertainty and modelling future flood risk.

Continue reading