Identifying the most suitable low-carbon heating technology for neighbourhoods in Great Britain

To decarbonise home heating, we will need to find the right heating technology for different homes. The Climate Change Committee is clear that heat pumps will be the most important technology for clean heating, but there are many types of heat pump to choose from and the right choice often depends on the area you live in. Some areas will be more suited to shared low-carbon heating systems such as heat pumps with shared ground loops (SGL) or heat networks (HN), other areas may be suited to individual systems such as air source heat pumps (ASHP) and ground source heat pumps (GSHP), and some areas will require a mix. Sometimes the factors that determine suitability may be simple, for example sparsely populated areas will only be suitable for individual systems. However, in many areas the factors that determine suitability are more complex - something which we will explore further later on in this data story.

Over the last year, we have been curating an open dataset which allows us to make predictions about heat pump suitability across Great Britain at a neighbourhood level (lower layer super output areas - LSOAs). This data can help decision makers, whether at an individual, local authority, or national level, to identify the most appropriate low-carbon heating technology options for a region.

In this article we take a deeper dive into our dataset, pulling out the types of insights we think will be most impactful for local energy planners. These findings are all powered by our heat pump suitability dataset which allows us to map the suitability of different heating technologies across Great Britain. All scores are from Nesta’s suitability scoring criteria - which can be found on our open Github repository.

So what insights did we find?

1. London is most suitable for shared heating networks, and rural Wales is most suitable for individual heating technologies

In general, the higher an area’s suitability is for individual technologies such as air source or ground source heat pumps, the lower the suitability will be for a heat network or shared ground loop, and vice versa. At the extremes of this trend, local authorities in rural Wales are generally most suitable for individual technologies and least for shared, whereas local authorities in London are most suitable for shared technologies and least for individual systems.

This is due to the various factors behind our suitability scoring. For example, properties in Britain particularly suitable for air source heat pumps might be detached homes with big gardens, whereas flats will be more suitable for heat networks. Therefore, dense urban areas will on average have suitability scores which are higher for heat networks and lower for air source heat pumps.

In the chart below you can see where different local authorities sit along this trend. Some local authorities are slightly off trend or outliers. For example the Shetland Islands and the Orkney Islands both have very high proportions of off-gas properties, making them particularly suitable for air source heat pumps. The City of London is another outlier where there is a particularly high proportion of flats, making heat networks very suitable and air source heat pumps unsuitable.

High level findings like these could help regional energy planners find clear target areas for streamlined heating decarbonisation initiatives, such as outreach about how to get an air source heat pump.

2. Suitability goes beyond density of households

As seen above, areas with very high or very low household densities tend to mean high and low suitability for shared technologies respectively. However, when we look at the medium population density scores we find no such correlation.

Because our metrics for suitability are calculated using household density as well as related factors such as rural/urban categorisation, garden size, and proportion of flats, we would expect a correlation. However, for the LSOAs with medium household density values we can see the influence of more factors, such as grid capacity, the presence of anchor properties, or whether or not the property is in a building conservation area (England and Wales) or a World Heritage Site (Scotland).

This means that for some neighbourhoods a more multi-faceted discussion about low-carbon heating is needed. For example, this might involve considering the specific housing stock on individual streets to determine the most suitable technologies. Our rich dataset can help identify such neighbourhoods to support these conversations.

3. Some regions are more variable in their suitability scores than others

We found that some local authorities are particularly suitable for air source heat pumps across most of their neighbourhoods. This might be accounted for by homogeneity in air source heat pump friendly housing stock. For example in Caerphilly only 13% of properties in our dataset are flats and 2% are in protected areas, whereas across Great Britain these numbers are 31% and 7% respectively.

On the other hand, some local authorities have quite high suitability for a particular technology, but also high variability across neighbourhoods. For example, Argyll and Bute in West Scotland has a similar average air source heat pump suitability score to Caerphilly but much more variation. This is likely because LSOAs in Argyll and Bute are approximately 50/50 split between urban and rural categorisations, and half of properties are off-gas.

We can look at the variability in scores for each local authority to compare trends across regions, for example we see local authorities in the East Midlands (Blaby and Bolsover) and South Wales (Caerphilly and Rhondda Cynon Taf) are particularly suitable for air source heat pumps across most of their neighbourhoods. Conversely, neighbourhoods in several London local authorities, especially City of London, Kensington and Chelsea, and Camden, are generally all very suitable for heat networks.

Making decisions about which technology is most suitable for a whole local authority may be more or less complex based on how varied the housing stock, household density and existing infrastructure are. Our data can be used to look at which areas have the most and least variability in the scores, highlighting certain areas where decision making might be more simple and some which require a much more careful, localised approach.

4. Areas with energy inefficient social housing could be targeted for heat network zones

Heat networks can decarbonise a large number of homes and could be particularly viable in areas with social housing. We see some local authorities with high proportions of social housing and generally high suitability for heat networks (although as we’ve seen in the previous section there can sometimes be high variability within a local authority). For example, on average 47% of properties in LSOAs in West Dunbartonshire near Glasgow are social housing - the highest proportion in Great Britain. This area also has relatively high suitability for heat networks, making it potentially a great candidate for local authority-wide heat network plans. We can also identify regions with high proportions of social housing but also generally low EPC ratings. For example, social housing accounts for around 30% of homes in both South Tyneside and Sunderland and the majority of properties also have an EPC rating of D or below.

Insights like these could help planners decide on collective switching strategies for the neighbourhoods most in need. If you are part of a local or planning authority and would like support producing insights like these for your area please get in touch.

Taking Norwich as an example of one of the local authorities with both a high proportion of social housing and high suitability for heat networks, we can dive further into the neighbourhoods which are particularly suitable for heat networks and which have low EPC scores. Five Norwich LSOAs which most fit these conditions are labelled and shown on the map below.

In Norwich (and generally), we see that whether or not an anchor property is present, and the proportion of flats in a neighbourhood both strongly affect the heat network suitability score. We can find particular neighbourhoods which might have most to benefit from a heat network zone - for example, a neighbourhood just south west of the centre of Norwich (in light blue on the map, labelled 011H) has a large proportion of flats, 62% of properties are social housing, and almost half the properties have an EPC rating of D or below.

By targeting specific neighbourhoods using our dataset in this way, local energy planners can kickstart a conversation to support collective neighbourhood switching in areas that might most benefit from it. If you are part of a local or planning authority and would like support producing insights like these for your area please get in touch.

5. Our heat network suitability scores compare well with the DESNZ heat network pilot zones

The Department for Energy Security and Net Zero’s (DESNZ) heat network zone identification pilot developed and implemented a methodology to identify areas suitable for heat networks within 28 pilot locations across England. We extracted geographic data from the reports released from these pilots, which allowed us to assess how our heat network suitability scores compared to DESNZ’s recommendations for each local authority included in the pilot areas.

For example, below we have plotted the local authority of Plymouth, and have provided a toggle of the areas identified as heat network zones by DESNZ and those not.

When we compare across all the local authorities included in DESNZ’s work, we find that our heat network suitability scores are higher in areas included in the DESNZ heat network zones and lower in areas not included. This provides some validation of our methodology and therefore allows us to make some suggestions for potential future heat network pilot zones. The areas we have identified as being highly suitable for a heat network, but aren’t included in the DESNZ report, include for example: Cambridge, Oxford, Brighton and Hove, Norwich, Portsmouth, Hastings and Lincoln.