Favourability index: an early-stage geospatial screening tool for low-enthalpy geothermal assessment mapping in superficial aquifers Open Access

For our input datasets we have taken a Findability, Accessibility, Interoperability and Reuse (FAIR) (Wilkinson et al. 2016) and Transparency, Responsibility, User focus, Sustainability and Technology (TRUST) (Lin et al. 2020) approach to create a data repository for visualization and decision-making. This was an important aspect in the creation of the tool as we value justice and a just transition (Wang and Lo 2021) as we move towards a sustainable future and the rollout of net-zero technologies. These approaches and concepts are crucial to ensure parity of power and availability of data and information between all those involved in decision-making to create this justice. Six key criteria for geothermal favourability visualization in superficial deposits are identified, comprising hydrogeological and societal information differentiated into heat resource factors and heat demand factors (Table 1). The following section investigates these criteria and identifies representative datasets.

The study boundary is formed by the c. 50 000 km² of superficial deposits covering Scotland (c. 64% of the total land area; Fig. 1). Lithology is a primary factor in determining a deposit's hydrogeological and thermal properties (O Dochartaigh et al. 2015b; Dalla Santa et al. 2020), so lithological variation, in addition to deposit presence, is considered key to geothermal favourability. This criterion is represented by the British Geological Survey's (BGS) ‘BGS Geology 625K’, superficial theme (BGS 2023b). This is the largest-scale geological map series made available by the BGS and provides an appropriate scale (1:625 000) for this resource. Data are available under the Open Government Licence via the BGS website, and comprises indexed and georeferenced superficial deposit vector polygons. Digitized from original Quaternary maps produced in 1977, data are accessible and interoperable with other georeferenced data.

Deposit thickness has significant impacts on the type and configuration of GSHP used and is key due to its influence on possible aquifer volumes and groundwater temperatures (Busby et al. 2009). The BGS Superficial Deposit Thickness Model (Basic) (SDTM) (Lawley and Garcia-Bajo 2009) dataset displays deposit thickness using three distinct raster grids. Raster grid products are derived from historical borehole, geological map and digital terrain model data. The Basic Superficial Thickness Model (BSTM) (Lawley 2016) was selected for integration into the favourability map because thickness is modelled directly from borehole records, reducing through hill modelling instances, where elevated topography is misidentified as unconsolidated deposit due to interpolation across rapidly varying elevations using insufficient data (Lawley 2016). This typically affects mountainous regions that host few historical borehole records. This dataset is a 1 km vector hex-grid at a 2.6 km² aerial resolution, and coverage is restricted by observation sites (borehole locations), and thus is extensive only in lowland regions of Scotland. Generally, deposit extents are relatively shallow (average 1.8 m); however, limited areas exceed 80 m, usually palaeovalley expressions (also known as buried valleys) at specific locations (e.g. the Forth, Clyde, Tay and Moray firths). Palaeovalleys comprise identifiable geomorphological features in the bedrock topography and are considered independently from other superficial deposits. The BGS Buried Valleys (Kearsey et al. 2017, 2019a, b) dataset displays the palaeovalley modelled thickness in Britain, derived from borehole records and geological maps, alongside their historically interpreted extent, produced at a 1:250 000 scale (vector format).

The BGS Buried Valleys dataset (Kearsey et al. 2017, 2019a, b) is also used to model depth (Kearsey et al. 2019a, b). The dataset represents a valuable addition to this study due to: (1) the relatively thick successions that occur in palaeovalleys often achieving the thickest values indicated by the resource (>40 m); (2) their demonstrable application to low-temperature geothermal projects (Allen and Milenic 2003; Allen et al. 2003); and (3) locations that often underlie densely populated regions such as Glasgow, Perth, Aberdeen and Inverness. The palaeovalley distribution is in broad agreement with depth figures derived from the SDTM, with 18 identified throughout the Midland Valley (Kearsey et al. 2019a). Significant features exist outwith this region in small numbers, notably around the Moray Firth, and the lands surrounding Elgin, in Moray Local Authority, and Stranraer, in Dumfries and Galloway Local Authority. Generally, buried valleys become more numerous and extensive to the south, observable by their increased prevalence in southern Britain.

Groundwater is the thermal transport medium in open-loop GSHP systems, so sustainable aquifer productivity or yield is key to enable function (Banks 2009a; Busby et al. 2009; Abesser et al. 2014). Individual groundwater factors (e.g. hydraulic conductivity, porosity, permeability, transmissivity, storage coefficient and lithology) influence well yield (Busby et al. 2009). Yields are therefore used as a proxy to identify deposits where these factors interact to provide sufficient abstraction to host an open-loop GSHP system. This effectively means that aquifer favourability is partially surmised by its ability to abstract and reinject water to the subsurface.

Sediment saturation has a strong influence on superficial sediment thermal properties, such that highly saturated sediments transfer thermal energy more efficiently than dry sediments (Busby et al. 2009; Dalla Santa et al. 2020). Significant hydraulic gradients may reduce thermal interference likelihood between abstraction and injection wells (from which a cold ‘thermal plume’ will radiate, under system heating conditions) by dispersing the thermal plume resultant from cool water reinjection (Banks 2009a). Furthermore, hydraulic gradient can determine eventual system end use: in aquifer thermal storage applications, significant groundwater presence flow is avoided to minimize loss from the thermal store. Considering groundwater flow is unfeasible at this resource scale, and any interpretations made should take care to judge the impact of this variable on potential GSHP geothermal or thermal storage applications (Snijders and Drijver 2016).

The BGS datasets for Hydrogeological Maps of Scotland (data and user guides can be assessed from BGS at https://www.bgs.ac.uk/datasets/hydrogeological-maps-of-scotland/; O Dochartaigh et al. 2015a) and Susceptibility to Groundwater Flooding (BGS 2023a) are used to represent productivity. The BGS Hydrogeological Maps of Scotland (O Dochartaigh et al. 2015a) display estimated sustainable aquifer borehole yields throughout Scotland, delineating between bedrock and superficial resources, alongside defining the dominant groundwater flow type in each feature (Fig. 1). This resource characterizes and locates productive aquifers and is designed for use at a 1:100 000 scale with a minimum 100 m buffer for any spatial application. Inherent superficial deposit heterogeneity, and inferior physical mapping detail represent uncertainty sources within this dataset, and numerous local factors will influence yield at the site level. Nevertheless, this resource ideally surmises likely productivity in a uniform and well-considered manner across the study area. Most superficial deposits do not comprise a significant aquifer; however, moderate–high and high productivity aquifers do follow similar distributions to deposit thicknesses, following major river valleys, particularly the Spey, Dee and Tay. Further high productivity areas include the coast surrounding the Moray Firth, Dumfries and Stranraer.

The BGS Susceptibility to Groundwater Flooding (BGS 2023a) dataset identifies regions at risk from groundwater flooding and is utilized by this resource to help spatially constrain groundwaters close to the surface, and therefore is accessible for geothermal exploitation. Data were produced using predicted deposit permeability in conjunction with local hydrogeological data and a digital terrain model to divide land areas into four susceptibility categories. Resource resolution is limited to ‘greater than a few hundred metres’ and is suited to regional assessment. Susceptibility in Scotland is based off river levels exclusively, and therefore introduces an uncertainty into this map resource away from river basins. High-susceptibility area distribution is limited to major river valleys and coastal floodplains; this is also in part due to observation bias in these areas.

Differential input/output fluid temperature is a principal efficiency governor in all HPs (Banks 2009a; Wu 2009; Rees 2016). This makes groundwater temperature at abstraction a crucial determinant in the eventual capacity any system could economically achieve, and thus defines its ability to affordably displace conventional heating practices. The near-surface temperature profile is decreasingly influenced by seasonal climate with depth, rendering it ineffective for GSHP applications (Busby et al. 2009; Sani et al. 2019). Seasonal temperature variation limit is constrained from c. 10 m bgl (Sani et al. 2019; Dalla Santa et al. 2020) to c. 15 m bgl (Ouzzane et al. 2015), and is coupled to average air temperatures (Rybach and Sanner 2000), allowing this resource to assume aquifer temperature at c. 15 m depth to approximate mean annual air temperature (Busby et al. 2009; Sani et al. 2019). Below c. 15 m bgl, temperature in the subsurface generally increases according to the local geothermal gradient (2.8°C/100 m in northern Britain; Busby et al. 2011). Note that thermal energy carried by advective and vertical groundwater migration means that site-specific geothermal gradients are variable.

Temperature is represented by a 20 year temperature average from January 1981 to December 2000 from ‘HADUK-Grid Climate Observations by UK countries’ dataset (Met Office et al. 2022). This is a climate variable collection taken from surface meteorological stations and interpolated across a 1 km grid. Methods are described by Hollis et al. (2019). Surface temperature is strongly dependent on altitude, such that the Scotland temperature map resembles an elevation model. Extreme lows (1–3°C) are restricted to the central Cairngorms Plateau and a few limited areas outwith. Average annual temperature highs do not exceed 10°C and are exclusive to lowland regions, with extensive coverage in southern and central Scotland, and lowland Aberdeenshire. Because elevation has such a strong influence on temperature, suitable subsurface temperatures are potentially achieved in proximity to extreme cold areas, where coastal glens ascend to mountainous terrain.

All GSHPs involve subsurface infrastructure: from shallow horizontal coil loops to multiple boreholes more than 100 m in depth. Configuration type, life span and additional measures or specialized materials are partially determined by ground physical and chemical properties. This informs appropriate scaling and installation methods for the geology (not just the consumer), reducing the overscaling/underscaling risk and the associated costs (Busby et al. 2009). Water quality is significant because corrosive conditions prematurely degrade casing, and scaling and clogging issues may indicate problematic dissolved gases and ions. Busby et al. (2009) recommend isolating reducing groundwaters from oxygenating conditions, and potentially pressurization to prevent degassing.

The BGS datasets Civils: Corrosivity (Ferrous) (Tye and Entwisle 2011) and Civils: Sulfate and Sulfide Potential (Entwisle et al. 2015) are used to evaluate the very shallow subsurface. The BGS Civils: Corrosivity (Ferrous) database (Tye and Entwisle 2011) is based on 1:50 000 scale bedrock geology maps and comprise a resource suite aimed at providing ground condition information to professional users. The Corrosivity dataset (1:50 000, ground resolution 50 m) identifies where conditions below the topsoil are considered aggressive/corrosive, or otherwise detrimental to subsurface infrastructures. This dataset is aimed at iron structure degradation, which could particularly influence GSHP deployment for integration into building piles/foundations or other such energy geostructures (Sani et al. 2019). Data are provided in georeferenced vector shapefiles, with land area categorized using a 1–5 corrosion potential based on the following criteria: moisture, redox status, pH, sulfates/sulfides and resistivity (Tye and Entwisle 2011). Most land is designated ‘unlikely’ to cause iron corrosion and special measures are likely to not be necessary; however, significant regions exist where this is not so. These areas are dispersed throughout the wider Grampian and Southern Upland regions, with concentrations in Sutherland, Caithness, Shetland, Lewis and Skye. The poorest quality regions surround estuarine environments near the Solway, Forth, Tay and South Esk rivers.

The BGS Civils: Sulfate and Sulfide Potential database (Entwisle et al. 2015) identifies where geohazards may occur due to sulfate-related subsurface chemistry at a 1:50 000 scale. This focuses on concrete damage from gypsum, but also general acid conditions and volume disruptions that could prove damaging to subsurface infrastructure. Although this dataset is chiefly concerned with the very shallow subsurface (c. 62 m bgl), significant impacts on GSHP maintenance costs are represented that potentially limit uptake when competing against conventional heating systems. Dataset coverage is complete across Scotland, with distribution very similar to those described for corrosivity. Major areas where acidic conditions could occur are centred in the far north and the northern and western isles. Besides estuarine environments, most land area does not have a significant sulfate/sulfide presence.

The dataset used to represent this is the Scotland Heat Map (Scottish Government 2022a), a virtual map tool that allows users to produce reports and visualize energy consumption for geographical areas in Scotland. Data are available at 50 m grid resolution in urban areas, alongside georeferenced heat network locations and a data confidence layer. Last updated in February 2022, this resource provides accurate and precise heat demand coverage for Scotland. Data distribution does not cover all land area due to low population densities in the central Grampians, NW Highlands and inland island regions. Demand is greatest over large urban areas (e.g. Glasgow, Edinburgh, Aberdeen and Dundee), reaching more than 180 GWh a⁻¹ km⁻² in some inner-city regions. Demand rapidly declines beyond major cities.

To observe open-loop GSHP favourability using the datasets identified in the ‘Input datasets’ section, data therein are recategorized into a common reference frame using methods adapted from Banks et al. (2020) to produce the FI. The data ranges from each nine dataset inputs are divided into categories ranging from 1: unfavourable to 5: favourable (Table 2) and these categories are assigned to features georeferenced within each dataset. A feature's FI score effectively quantifies the geothermal favourability magnitude determined by data recategorization within the FI. Data recategorization is allocated by qualitative geoscientist assessment within the unique heat resource ((hydro)geological) and heat demand (societal) context of Scotland. These FI valued georeferenced features are allocated a traffic-light colour code synonymous with their FI category (red: unfavourable to green: favourable), allowing the user to visualize and consequently spatially constrain varying favourability magnitudes within each dataset using the same reference frame (Fig. 2). This provides the ground piece for a quantitative spatial favourability assessment across simultaneous datasets and preludes an integrated favourability map resource.

Scored and colour-coded maps are converted into a raster format, using a 1 km grid and British National Grid projection, allowing comparison within the same 1 km² cell across all variables. Cells at the same location are summed across inputs to achieve a cumulative FI score for each georeferenced cell. Cumulative possible scores thus range from a minimum of 9 (1 × 9) to a maximum of 45 (5 × 9). This data range is divided equally into five bins (their boundaries rounded to the nearest integer, corresponding to the five favourable–unfavourable FI categories; Table 2) and colour coded accordingly (Table 3). Note that the actual output range depends on data inputs. The resulting maps therefore visualize cumulative favourability over all input datasets as an FI score or category (Fig. 3). Because there is no relative importance assessment between inputs to inform weighting, inputs hold equal weights. This method is repeated nine times, each one excluding a different input dataset, producing a map suite that compares varying input dataset spatial influence under cross-examination (Fig. 4).

The map resource is developed at a 1 km² resolution with an intended 1:200 000 working scale. No additional detail is gained by study at higher resolutions. Input datasets are not all geographically extensive over the same grid cell area because the geographical features they represent are not always coincident. The FI and uncertainty factor index (UFI) (see ‘Data uncertainty I’ in this section), however, require all cells within the study boundary to hold data to receive a score and enable fair comparison. This is resolved by assigning a value to these ‘blank’ cells that reflects feature absence. This value is 1 (unfavourable) for all datasets, excluding the SDTM (Lawley and Garcia-Bajo 2009) where 3 is assigned because this dataset is compromised by observation bias, and a wide superficial deposit thickness deviation is likely to exist in cells within the study boundary but outwith this dataset.

The necessity to balance heat resource, heat demand and socio-demographic factors when conducting an integrated data approach to geothermal favourability geospatial analysis has been introduced. The above methodology outlines an FI resource tool that provides key heat resource and heat demand factor consideration but excludes external factors. This ‘core’ resource is intended to provide a base level that is adapted to the end user who may or may not have a geological background. External considerations are left free to allow the end user to tailor the map resource to their own context. ‘External considerations’ is therefore any spatial information that could impact potential installation success, namely socio-demographic influences.

Two external data integration methods are available to the FI map resource. The most simplistic involves overlaying an external dataset on the FI resource (Fig. 3) in a separate layer in a GIS software system. This is a suitable option to quickly spatially visualize the data-rich FI score against one or two distinct and separate parameters. For example, a user may be interested in the role that LE geothermal could play in providing low-carbon heating to outdated building stock that has poor thermal efficiency. In this situation, overlaying thermal efficiency ratings by postcode is an effective way to identify interest areas. This is a simple step that is achievable with basic data manipulation skills in GIS software.

Full data integration is advisable if a complex dataset array requires simultaneous assimilation so their combined impact on the geothermal favourability for one or more large regions is visualized. External datasets are thus additional inputs in the FI (Table 2) and follow the same integration methodology. Because the dataset number in the FI increases, so does the total FI value range (Table 3). This means that the FI category boundaries will change because they must remain in equal proportion to enable an unbiased comparison across the map suite. This method is thus more precise than simply overlaying independent datasets but requires deliberate action in the upstream project process to identify what parameters are important to the end user.

Data uncertainty is observed by creating the UFI (Table 4) and accompanying map similar to the FI and ‘Combined Factors Map’. Uncertainty in this study assesses how well each input dataset represents the key criteria, and so how well FI values represent the actual ground conditions. A variety of six epistemic and aleatory biases are considered, and a colour-coded 1–5 uncertainty magnitude category sequence (1: low confidence to 5: high confidence) is produced. Each input dataset is scored by each bias, and a cumulative score for each individual dataset produced. The theoretical values range from a minimum of 6 (1 × 6) to a maximum of 30 (5 × 6). This 24-point range is subsequently divided equally into five categories, equating to a 1–5 low to high confidence sequence, and the dataset cumulative scores are categorized accordingly. This is the UFI score for the data extent of each of the nine input datasets. The UFI score is summed and visualized during the dataset integration process (see ‘Data uncertainty II’ in the following section). This data uncertainty layer is therefore a nine-factors map, and summed UFI scores for each grid cell are visualized by utilizing Table 3. Likewise, the FI ‘blank’ cells in each dataset that are within the study boundary but do not have associated data are assigned an appropriate UFI value. Excluding heat demand, all such ‘blank’ cells are assigned a UFI value of 1, reflecting the inability to determine whether features are present in these cells or not. Blank cells in ‘Heat demand’ receive a high confidence score (4) because heat demand is measured, not modelled by input datasets.

An 18 map suite was produced, excluding uncertainty layers (Figs 3 and 4). Observable in all maps is the study boundary inherited from superficial coverage data. A ‘Combined Factors Map’ was also produced where FI values are derived from all nine input datasets (Fig. 3), and so exhibits the maximum possible FI score range (9–45; Table 3). This map therefore visualizes geothermal favourability in superficial aquifers across the study area, according to both heat resource and heat demand factors. Continuous or categorical data viewing mode is optional and determined by use requirements. Broadly, categorical viewing is useful for generalized LE geothermal potential assessments over large regions, whereas continuous viewing provides maximum detail (Fig. 3).

The remaining 16 FI maps each exclude a criterion (Fig. 4 and Table 2). FI category-consistent proportionality between nine inputs maps, and those using eight (Table 3), enables a comparison of the maps across the suite when visualizing them using the FI categories (Fig. 4). By taking this step, the map resource can cross-reference itself to compare data and identify how each criterion spatially interacts with the others. In this way, and through external dataset addition pertinent to the end user, the resource is used dynamically to suit the end-user needs. Geospatial data are extracted from map layers using standard GIS analysis methods; for example, when identifying area metrics and FI scores in a particular category.

The methodology describing how external data are integrated with the resource is outlined in ‘Integrating external data I’ in the previous section. By producing a screening tool that is interoperable with various data types pertinent to the end user, a resource is provided that can respond to specific geospatial queries that are immediately answerable. Intrinsic data (Table 2) are specifically (hydro)geological and heat demand-based and provided to the end user ‘pre-calibrated’ to LE geothermal technologies. An end user who does not have a (hydro)geological background may then integrate external data that are relevant to them, producing a dynamic end resource that can interoperate both geological and non-geological expertise. The tool can be used to inform non-geological experts how geological factors impact their current shallow subsurface resource before contact with a geological expert in person. Two methods are given that utilize varying degrees of data integration. Both methods exemplify external data interoperability with the intrinsic FI resource data (i.e. heat resource and heat demand factors). This section briefly demonstrates these methods and their utilization.

Any georeferenced data visualization and manipulation undertaken by the end user, provided the display uses compatible projection systems, is supported by the FI map resource using this method. This method is a time-effective way to integrate external data because data do not require standardization and raster format conversion; for example, simultaneous visualization using FI categories in Glasgow alongside percentage populations in social housing by postcode (vector data 5 in Fig. 5a). The integrated map resource is used to identify postcodes where both a high FI category and a high social housing proportion are located, demonstrating how basic geoinfographic data are extracted quickly from analysis (5 in Fig. 5b). This data overlay may provide sufficient information to an end user without more complex integration; for example, a government team may wish to identify urban areas that warrant consideration for a low-carbon heat network installation on a local scale.

A more thorough data integration method is to expand Table 2 to add data as an integral part of the FI. This affects the favourability mapping process end product. Data require standardization and conversion to 1 km² raster format. This new raster is summed with the existing map resource, and the numerical boundaries between FI categories recalibrated (Table 3). This takes more time than simply overlaying data and necessitates the end user to independently assign a favourability range to added data, but it also allows the simultaneous visualization of multiple external factors.

The data uncertainty methodology is outlined in ‘Data uncertainty I’ in the previous section. Uncertainty is visualized in a layer that is compared to the screening tool outputs: high values (darker colours) indicate regions where there is more confidence that input datasets accurately represent the conditions in the shallow subsurface (Fig. 6a). Note that cells within input datasets with no associated data that received a neutral/unknown FI value also receive a low confidence score (low number) in the data uncertainty process. This generally results in higher confidence in regions where there is consistent data coverage across input datasets. When visualized using 1–5 favourability categories, all UFI values comprise neutral or encouraging categorizations (Fig. 6b). This effectively produces a data uncertainty mask that can be superimposed on the FI resource to communicate to the end user the spatial ramifications that the data quality introduces to the map resource (Fig. 6c). Note that UFI maps must be used with FI maps produced from the same data inputs to provide valid results.

Approximately 50 000 km² (c. 64% of the Scottish land area) are assessed; land outwith this is discounted from the study due to a lack of superficial deposits (Fig. 1). The distribution of superficial deposits is therefore the primary geothermal resource extent driver. Upland areas (NW Highlands, Grampian Mountains and the Southern Uplands) have few extensive deposits, with accumulations preserved only in deep glens. Generally, the eastern lowlands and the Midland Valley host greater deposit extents that allow assessment of the superficial aquifer favourability. High favourability areas (FI category 5, Favourable to category 4, Encouraging; or FI values 31–45; Table 3) tend to occupy low-altitude regions within this extent, notably expressed in the Midland Valley and Moray coast (Fig. 3). These areas host several features that contribute to favourable GSHP conditions: relatively thick superficial successions often in the form of palaeovalleys, major river systems that provide a large hydrological input (e.g. Clyde, Forth and Ness) and major urban populations. The greatest confidence in the FI values exist in the Midland Valley, the eastern coastal regions and the land surrounding the Moray Firth (Fig. 6)

It is useful at this stage to reiterate the study aims. Namely, to provide a science communication tool that allows dynamic integration and visualization of disparate data to aid geothermal resource understanding amongst policy-makers, implementers and community groups. High capital and running costs, seasonal efficiency, and lack of benefit awareness in policy-makers and consumers of heat-pump technology are some of the contributing factors to the low uptake of heat pumps in the UK when compared with similar nations (Singh et al. 2010; Kokoni and Leach 2021; Ahmad 2023). In Canada, Corbett et al. (2023) identified targeted policy based on associated public perception of policy as a potential way to increase the uptake of heat pumps in specific regions of the country. Studies of such geospatial societal relationships are ideally suited to the FI geothermal methodology proposed. As a tool to inform and therefore improve policy effectiveness in the UK, we focus on the need to enhance top-down approach effectiveness in current and future policy implementation aimed at enabling the uptake of GSHPs in Scotland and the UK, whilst simultaneously encouraging bottom-up community-led initiatives and empowerment (Singh et al. 2010; Ahmad 2023) by providing an accessible infographic tool, underpinned by the FAIR and TRUST principles. This is approached by encouraging interdisciplinary communication, enhancing awareness of fundamental geological considerations and associated engineering requirements pertinent to geothermal utilization, and simultaneously encouraging the end user (local government or community-based initiatives) to consider how societal parameters important to them interact with these to influence geothermal favourability. In this way we address the limited geoscientific and socio-demographic information transfer in a (local) government context for the delivery of geothermal district heating schemes within the LZEHS strategy deployment across Scotland (Acheilas et al. 2020).

The core map resource comprises a suite of raster maps that fundamentally visualize the superficial aquifer geothermal favourability distribution according to the sum of heat demand and heat resource factors (Figs 3 and 6). Low-altitude regions, particularly palaeovalleys exhibiting thick superficial sequences, represent high favourability areas; Allen and Milenic (2003) and Allen et al. (2003) considered the palaeovalley geothermal potential in Cork, Ireland, which has similar geological and socio-demographic settings to those found across Scotland. Underlain by the Lee palaeovalley, results indicate that palaeovalley deposits have the potential to supply a high heat demand urban area located above them. Up to 18 palaeovalley features have been identified in the Midland Valley, Scotland's most densely populated region (Kearsey et al. 2019a), suggesting that these features represent high potential targets for the provision of LE geothermal heating.

The end-user ability to observe variations in the FI values and categories resulting from external socio-demographic factors at specific sites and excluding chosen input factors (Fig. 4) produces a dynamic resource that responds to socially relevant geospatial queries. The resultant map resource therefore deviates from similar-scale geothermal favourability resources in Europe (England and Wales, Abesser et al. 2014; Finland, Arola et al. 2014, 2019: France, Bezelgues et al. 2010; Italy, Casasso and Sethi 2016; and Slovakia, Krcmar et al. 2020) in two ways: direct user interaction in the map data source and production, and encouraging wider non-geological data integration beyond the heat demand.

First, the map resource is designed for unique adaptation to end-user purposes, requiring users to select and integrate data themselves. This may include constraining the study area, adding societal heat demand data or even adding auxiliary/unconventional heat sources (Meibodi and Loveridge 2022) that may contribute to a well-balanced and versatile heat network where geothermal serves base-load demand (Acheilas et al. 2020). By doing this, a streamlined tool is produced that is more likely to give relevant answers when considering spatial relationships between the subsurface resource and the above-ground demand. Such adaptability is particularly useful in urban contexts, where geospatial understanding of domestic and industrial demand centres, subsurface heat resource, socio-demographic trends and the urban environment spatial nature is crucial to geothermal district heating scheme decision-making (Acheilas et al. 2020). For example, a high favourability area may warrant exclusion from consideration if it already utilizes an alternative green-heating solution. Adaptation and addition to/of the core generalized resource by multiple end users encourages a tailoring process whereby a specialized geospatial assessment that meets specific research goals is produced. Tool design thus targets not only a range of purposes, but also a range of people. This quality is utilized effectively when considering rural heating decarbonization.

The ‘just transition’ concept has clear spatial applications (Walker 2009; Wang and Lo 2021); however, small, isolated communities are often the last to benefit from national-scale amenity rollout (Supapo et al. 2021). The map resource seeks to encourage a just transition in several ways. First, by providing a tool for regionalized knowledge transfer and resource/demand/socio-demographic spatial visualization; this improves knowledge capacity and understanding of energy materiality capacity across local government and community indicatives, an important step in actuating the Scottish Government's local heat and energy efficiency strategies (Wade et al. 2022). Furthermore, as a science communication tool, the map resource can help local government to actively engage and empower rural communities by increasing involvement in the renewable energy solution processes (Kallis et al. 2021), thereby advancing bottom-up decarbonization strategy development alongside a top-down approach (Lennon et al. 2019). In addition, this approach aims to encourage community-owned/led heating scheme developments, and so promote achievement of net-zero ambitions in a socially equitable way (Bush and Bale 2014).

This resource explores the spatial relationship between geological resource and societal demand, synonymous with below-ground and above-ground factors. A basic, yet far reaching, observation of all direct-use geothermal systems is that thermal energy cannot be transported any significant distance. Consequently, the geothermal resource demands that heat consumption occurs near to production. By extension, to utilize geothermal heat at a location, both a geothermal resource and an appropriate heat demand must exist. Future direct-use/GSHP geothermal project distribution will thus be likely to reflect the population density over geological favourability. Heat demand distribution is therefore a critical consideration for any geographical favourability tool, and to fully understand and visualize the direct-use geothermal potential, an interdisciplinary knowledge transfer is needed (Acheilas et al. 2020). In addition, understanding the implications of heat demand for LE, direct-use geothermal in the present and future requires more than simple visualization. Demographic predictors for where demand will fluctuate according to the population density, poverty, occupation, age, etc.; current energy supply/fuel type and property insulation; regional development funds and political landscape; and industrial site distribution, including farms, will all have a significant impact on how much thermal energy (including cooling) is required in a locality, and how that demand will change spatially and temporally (Flower et al. 2020).

To this end, visualizing FI values with some of the dataset inputs deactivated aids understanding of the variability in the geothermal favourability. For example, comparing the ‘Combined Factors Map’ (Fig. 3) with the ‘Minus Heat Demand Map’ (Fig. 4d) provides an important fail-safe process whereby high favourability areas identified in the ‘Combined Factors Map’ have verified (hydro)geological derivation, and are not heat demand artefacts. Interestingly, average FI values increase when heat demand is discounted, suggesting the reverse: map layer comparison is crucial to ensuring that high FI values are not representative of hydrogeological sources alone, and are still consistent with spatial heat demand distributions. Because the limiting factor in Scotland appears to be the heat demand distribution, rather than the land area where geological factors are ideal, indicates that the spatial capacity to host LE geothermal in Scotland has not been reached. Note that this ‘spatial capacity’ does not indicate subsurface energy potential at individual sites, rather that new build sites will be likely to coincide with high FI values. This reiterates the crucial importance that heat demand and, by extension, socio-demographic factors have on direct-use geothermal favourability: the subsurface resource must meet the needs of people and the infrastructure existing above it. In this way, spatial resource screening for direct-use geothermal fundamentally differs from transportable energy sources, such as petroleum-based fuels and, to a lesser extent, electricity.

The map resource is developed at a 1 km² resolution with an advised 1:200 000 working scale. To fulfil map resource potential, end users are encouraged to select additional non-geological data for integration. In this way, the resource is a dynamic geospatial analysis method that supports adaption for use at any location, utilizing different datasets and considering different, or multiple, technologies. Whilst this step tailors the resource to a specific area and needs therein, it simultaneously introduces unknown data, making an accurate working scale impossible to prescribe prior to data integration. The working scale given above is based on intrinsic data that will remain common to all subsequent resource versions that utilize the core resource, and so is a practical solution that represents a lower-scale limit. Resolution limitation introduced by integrated datasets necessitates consideration by the end user, who is referred to Table 1. The map resource cannot reliably constrain geothermal favourability below this scale, highlighting the regional nature of this tool and the continued importance of site-specific assessment and due diligence when assessing and utilizing the subsurface thermal budget.

The lithological variations of superficial deposits in the depth dimension are not captured because this is not represented by data sources. The resource therefore assumes that deeper superficial sequences are more likely to include suitable aquifer layers at appropriate depths: superficial aquifers exceeding 80 m in thickness exist in palaeovalley features in Scotland (O Dochartaigh et al. 2015b). Particularly in these deeper deposits favoured by this tool, however, lithological properties are likely to vary considerably both laterally and along a borehole length, so this tool cannot replace the role of site investigations. Concerning palaeovalleys, for example, the glaciogenic fill is notably heterolithic, and the interconnectivity between ‘clean’ sands and gravels, and the potentially silt- and clay-rich glacial till/diamict is difficult to ascertain (Kearsey et al. 2019b). Similarly, advective heat transfer in the subsurface is not considered due to resource scale; however, it has clear impacts on GSHP system effectiveness (Banks 2009a, b). Temperatures encountered are likely to deviate locally from those predicted by this resource.

Varying dataset and data source interoperability is a key method concept, encouraging interdisciplinary understanding; however, it also introduces a wide range of limitations into the resource. The uncertainties and limitations of each dataset are inherited by the map resource. The reader is directed to Table 1 and the user guides for each input dataset for further details. In addition, external data added by the end user will itself bring uncertainty into the resource. The UFI goes some way to spatially define the magnitude of uncertainty introduced by each input dataset by cumulatively quantifying how well each dataset represents the key criteria considered. The resultant uncertainty mask is intended to highlight areas of greater data uncertainty to the end user, and to prompt them to interrogate their own data for uncertainty, thereby increasing the resource effectiveness (Fig. 6). Notably, high uncertainty areas simply identify areas with a greater chance of encountered value deviation from those predicted by the map resource, but this does not always mean this deviation will negatively impact a project. The fact that higher UFI values generally coincide with populated regions (the Midland Valley and east coast) is indicative of the greater data quantity and quality in these regions.

To achieve ambitious 2045 net-zero emissions targets, Scotland must decarbonize space heating and embrace low and zero emissions heating systems (LZEHS), including heat pumps (HPs) and district heat networks (DHNs). Government documents such as the ‘Heat in Buildings Strategy’ and similar outline the policy framework for advancing this facet of the energy transition; however, the role of geoscience and the subsurface in relation to energy takes a minimal role.

Low-enthalpy/low-temperature geothermal and ground-source heat from shallow groundwaters have noteworthy potential to provide for local heating demand when used in conjunction with HP and DHN technologies, already widely used globally. The ability to utilize the low-enthalpy thermal budget of superficial aquifers relies on the geospatial combination of heat resource ((hydro)geological) factors, heat demand factors and socio-demographic factors. The map-based methodology and core resource proposed here aims to facilitate visualization and understanding of these geospatial considerations by local government and community organizations. By doing so, this method can help to inform future policy and community energy transition decision-making. Particularly relevant are the policy implications of recognizing that low-grade thermal resources cannot be transported far from where they are produced. To effectively utilize GSHPs and their carbon-saving benefits, this fundamental resource localization demands that related energy policy be addressed according to local considerations. This shift to accommodate decentralized energy system planning alongside conventional centralized energy source utilization will place new demands on local and regional government, including an understanding of the subsurface resource they hope to utilize.

While this method is intended to be applied in a variety of contexts and locations (determined by the end user), initial findings in Scotland indicate that (hydro)geologically suitable superficial aquifers are often coincident with large urban conurbations, resulting in large extents where heat resource and heat demand are in proximity. With the addition of Scotland's generally encouraging attitude towards climate and energy transition policy, it is suggested that Scotland exhibits the key prerequisites to meet a significant portion of its urban heat demand by utilizing superficial aquifer thermal resources through HP and DHN technologies. A potential avenue for the application of this decision tool in town planning could involve considering land use based on above- and below-ground geothermal favourability in new-build zones. In particular, palaeovalley/buried valley geomorphological features provide deep and generally productive superficial aquifers that are often overlain by cities.

The proposed resource pertains to a regional and indicative assessment of geothermal favourability in superficial aquifers. It is not designed for the planning of individual projects/properties and should not replace robust hydrogeological and thermogeological assessments of specific systems. Nevertheless, it can provide a valuable early-stage geographical screening tool by which organizations can assess the validity of ground-source/geothermal heating against alternative heating solutions.

The authors acknowledge the British Geological Survey for the provision of academic data licences. The authors are grateful to the reviewers for their time and useful comments.

TAR: conceptualization (equal), methodology (lead), visualization (lead), writing – original draft (lead), writing – review & editing (lead); AJH: conceptualization (equal), supervision (equal), writing – review & editing (supporting); CEB: conceptualization (equal), supervision (equal), writing – review & editing (supporting).

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The data that support the findings of this study are available from the British Geological Survey (BGS); restrictions apply to the availability of these data, which are used under licence for the current study and so are not publicly available. Data are, however, available from the authors upon reasonable request and with permission of the BGS.