The ESA Global Development Assistance Urban Sustainability (GDA Urban) activity is working with the European Bank for Reconstruction and Development (EBRD) to strengthen climate resilience in Novi Sad, Serbia. The collaboration uses satellite-based Earth Observation (EO) to identify extreme urban heat, understand where the most exposed population is, and guide targeted, investible cooling solutions for the city.
This case study forms part of the GDA Urban consortium’s wider effort to demonstrate International Financial Institutions (IFIs) with decision-ready, user-friendly EO services. In Novi Sad, the focus is on finding heat hotspots, mapping population vulnerability, and pinpointing where green infrastructure and other cooling actions can have the biggest impact.
This story explains how the GDA Urban consortium is working with the EBRD climate strategy team to translate EO insights into practical evidence that can support early planning and future investment decisions.
Why this matters: Europe’s cities are heating up
Temperatures are rising and heat waves are becoming more frequent and intense across the EBRD’s regions. Europe is warming twice as fast as the global average. Extreme heat harms the economy (lost work time, lower productivity) and public health. Cities are hit harder because of the Urban Heat Island (UHI) effect: built-up areas stay hotter than their surroundings.
Novi Sad is vulnerable to these trends. By 2050, according to the UN IPCC report, summer peak temperatures are projected to rise by ~2.2–5.7 °C. The city also has features that trap heat—few green spaces, fragmented land use, and many hard, non-permeable surfaces. Through the EBRD Green Cities initiative (which Novi Sad joined in 2019), the city aims to plan and finance practical investments that reduce heat risk, and EO analysis can help direct those investments to the right places.
UMAP, a live platform for smarter heat planning
The GDA Urban consortium is supporting the EBRD’s work in Novi Sad. At the heart of this activity is UMAP (Urban Microclimate Analysis Platform), a live, city-by-city heat intelligence platform that turns complex EO and street-level measurements into clear, actionable maps. It’s built for planners, health teams, utilities, and analysts who need fast answers to simple questions: where is heat building, when does it linger, and who is most at risk?
UMAP integrates satellite heat maps, street-level measurements and vulnerability/exposure indicators into a single decision-making dashboard. The data is produced by the consortium (formed by EO European-based companies, such as GAF, as the prime, DLR, MindEarth and Constellr) and its local partners and then fed into the platform:
- Satellite heat mapping. A decade of Landsat and Constellr observations is used to map Surface Urban Heat Island Intensity (SUHII). See some images below.
Figure 1:This image shows daytime Land Surface Temperature derived from LST Precision (Skybee-1) in September. Warmer tones highlight sun-exposed built surfaces, while cooler areas correspond to the river and vegetated zones. Credit:Contellr
Figure 2: This image shows night-time Land Surface Temperature derived from LST Precision (Skybee-1) in November. Warmer areas indicate surfaces retaining heat after sunset, while cooler tones reflect vegetation and open land cooling more rapidly. @Constellr
- Vulnerability and exposure. EO-derived factors (e.g., albedo, tree canopy, shade) are combined with population layers from the World Settlement Footprint (WSF) Tracker, a suite of open, high-resolution global datasets from DLR, and WSF Population (10 m) to find places where high heat overlaps with limited ability to cope, useful for targeting action.
- Local measurements. MindEarth carried out a two-week field campaign using Atmotube PRO sensors on predefined walking routes to record air temperature, adding on-the-ground context to satellite-derived land surface temperature (see below).
Surface Urban Heat Island intensity can be seen at the block/parcel level (how much hotter a block is than nearby natural areas), flip on land cover, imperviousness, and population layers, and even see 3D buildings to understand “canyon” effects. A timeline slider lets you scan back through years of data to spot trends and change (e.g., new roofs, new asphalt), while parcel pop-ups show quick stats without leaving the map.
For planning “what-ifs,” the Heat Event simulator lets you “dial up” a hot day and preview which parts of the city would spike first, revealing pockets where cooling measures will matter most. Vulnerability views combine exposure, adaptive capacity, and access to green space, with adjustable weightings so local teams can align the index to their context. An hourly/daily module shows how parcels warm and cool through a typical day—useful for identifying places that stay hot overnight, when health risks rise.
UMAP also offers a Greening Scenario Simulator. The tool lets users test how replacing impervious surfaces with vegetation could influence local temperatures. By adjusting a simple greening slider, planners can see estimated parcel-level changes in land surface temperature, based on a model linking surface characteristics to heat retention.
Initial results underline an important finding: greening is most effective when implemented in clusters, where nearby green areas reinforce each other. For Novi Sad, this can help pinpoint priority zones where coordinated investments in trees, parks or green roofs could deliver meaningful cooling benefits.
UMAP is designed to be intuitive and visually engaging, thereby lowering the barrier for non-technical stakeholders who wish to engage directly with heat evidence. Its interactive maps, parcel-level summaries and scenario tools enable users to explore patterns, test assumptions and immediately visualise the spatial implications of different choices. This makes it particularly valuable in policy and planning contexts, where decisions are often shaped through discussion rather than technical modelling. By presenting heat risk, vulnerability and potential solutions in a clear, comparable format, UMAP helps city officials and IFI teams translate evidence into concrete planning measures, prioritised investments and more informed urban development strategies.
Walking the heat in Novi Sad
To make the analysis locally relevant, MindEarth led a dedicated field campaign in late July/early August 2025 to measure air temperature at the street level. The approach uses a crowdsourced mapping setup to capture intra-urban temperature variations, which is fundamental for understanding urban heat dynamics and for contributing to the reliability of EO products.
Five walking circuits were designed to cover distinct parts of Novi Sad: compact streets in the city centre, residential areas with urban green spaces, wide avenues, riverside paths, and industrial zones. Route design combined OpenStreetMap, UMAP (SUHII parcels and land-use layers), recent
satellite imagery, and local stakeholder input to ensure the paths are both walkable and representative, and to reflect the city’s thermal heterogeneity.
The campaign was implemented in close collaboration with the Faculty of Technical Sciences, University of Novi Sad, involving Environmental Engineering students under the guidance of Assoc. Prof. Dr. Aleksandar Anđelković, Director of the Energy Innovation Center, and Prof. Dr. Nemanja Stanisavljević, Head of the Department of Environmental Engineering. The students’ active role in data collection was supported by the professors’ expertise, local knowledge, and logistical support, all of which were essential for the successful delivery of the campaign.
Local student systematically walked assigned routes twice per day (morning and early afternoon) for two weeks. Time windows were chosen to capture diurnal variation and, when feasible, to coincide with Landsat-9 constellr SkyBee overpasses.
Although the field data cannot be used to directly validate satellite-derived LST—as LST and air temperature measure different physical properties—it provides a valuable complementary dataset. In particular, the in-situ measurements offer insight into the human experience of heat at street level, which is not always captured by remote sensing alone. The field data also supports potential calibration and cross-comparison with new thermal EO products, such as those from the SkyBee satellite, especially in areas where spatial resolution is high enough to enable meaningful correlation.
In addition to the technical value of the data collected, the campaign also fostered local engagement. The student volunteers received hands-on training, not only in using the sensors and app, but also in understanding the broader context of the project, why urban heat matters and how data can shape climate resilience planning.
EBRD reflections on UMAP’s strengths and gaps
Feedback from the EBRD team highlighted the value of UMAP as a visually engaging and intuitive way to explore the urban heat island effect. They appreciated the clarity of the data presentation and the potential for the platform to inform policy dialogue and early-stage planning discussions. At the same time, the reviewers noted several areas for improvement before the tool could be scaled or mainstreamed. These include adding clearer guidance for new users, refining certain visual layers to better match real-world conditions, and improving how ground and satellite data are integrated to strengthen analytical reliability.
They also emphasised the importance of modularity and practical scalability, for example, understanding how much local data or engagement is required to apply the tool in new cities.Together, this feedback provided a clear direction for refining UMAP into a more robust and operational tool for urban heat planning.
