BeatTheHeat: A federated database system for urban planning and ESG

Understanding the limits of current thermal mapping for urban planning

Monitoring land surface temperature (LST) is a cornerstone of sustainable urban planning. Unlike air temperature, which reflects conditions at fixed measurement points, LST offers spatially explicit insights into how different materials such as soil, vegetation, and built infrastructure absorb and release heat. This variable is crucial to identifying and managing urban heat islands, areas where surface temperatures can rise several degrees higher than surrounding zones due to dense construction, reduced vegetation, and human activity.

However, despite the growing availability of satellite data, current thermal mapping systems face significant limitations that constrain their operational value. The first challenge lies in data fragmentation: temperature measurements are dispersed across multiple datasets, often characterized by varying formats, calibration standards, and temporal resolutions. This heterogeneity makes it difficult to integrate information and extract reliable, continuous insights at the city scale.

Another structural constraint is the long-standing trade-off between spatial resolution and revisit frequency. High-resolution missions such as Landsat provide valuable local detail but only every 16 days, missing short-term variations that shape urban heat dynamics. Conversely, sensors like MODIS capture thermal data multiple times per day but at coarse, kilometer-scale resolution, insufficient to distinguish microclimatic differences across neighborhoods or urban infrastructures.

These gaps translate into practical challenges for planners and policymakers. Without frequent, high-resolution datasets, it becomes difficult to detect emerging hotspots, assess the impact of mitigation strategies, or model future heat exposure scenarios. The result is a disconnect between scientific observation and actionable urban intelligence.

Overcoming these constraints requires new approaches that improve interoperability and continuity across datasets for sustainability: approaches capable of combining spatial precision with temporal depth to deliver truly operational Earth Observation solutions for climate-resilient cities.

How federated database systems enhance satellite-based environmental data

As cities face increasing climate pressure, the need for comprehensive and interoperable satellite data becomes essential to support sustainable urban planning. Yet, conventional approaches to managing environmental datasets often rely on centralized systems, which struggle to handle the scale, diversity, and fragmentation of information generated by different Earth Observation missions. This is where federated database systems emerge as a strategic innovation.

A federated database is a virtual data architecture that integrates multiple, autonomous databases into a single logical system. Unlike traditional repositories, data in a federated environment remains stored at its original source but becomes accessible through a unified interface. This model allows researchers, public administrations, and private organizations to query and combine datasets for sustainability, such as land surface temperature, air quality, or land cover indicators, without the need for physical data consolidation.

In the context of Earth Observation (EO), this approach offers a significant operational advantage. Satellite missions differ widely in spatial and temporal resolution, spectral characteristics, and acquisition frequency. By interconnecting these heterogeneous data streams, federated systems enable the creation of integrated environmental datasets that overcome the intrinsic limitations of individual missions.

From a technical standpoint, federated databases rely on standardized metadata, ontologies, and data models to ensure semantic alignment across sources. They support distributed query processing and automated harmonization protocols, making it possible to retrieve near real-time information while preserving the integrity and ownership of the original data. This architecture not only reduces redundancy but also promotes transparency and reproducibility, two essential principles for environmental governance and ESG reporting.

By breaking data silos and enabling cross-domain analytics, federated database systems transform satellite data into actionable knowledge. Their capacity to integrate thermal, atmospheric, and socio-economic indicators opens new pathways for urban heat island monitoring, risk assessment, and sustainable infrastructure design. In essence, they redefine how environmental information can be shared and applied, building a data ecosystem where scientific precision directly supports climate adaptation and sustainable urban decision-making.

Integrating multi-source datasets for ESG-driven urban intelligence

Building a federated database for environmental monitoring requires more than just technological infrastructure, it demands a shared data ecosystem that brings together multiple actors, sources, and methodologies. The ultimate goal is to create datasets for sustainability that are not only scientifically accurate but also directly usable for ESG-oriented decision-making and urban planning.

A federated approach enables the integration of multi-source data streams coming from satellites, in-situ sensors, municipal databases, and citizen observatories. Each of these datasets contributes a specific layer of environmental intelligence. Satellite data deliver wide-area, consistent measurements of variables such as land surface temperature (LST) or vegetation cover. Ground-based monitoring networks add local precision on air pollutants, humidity, or albedo changes. Urban administrations contribute socioeconomic and infrastructural datasets on population density, land use, or building typologies that contextualize the physical environment. When harmonized under a federated architecture, these heterogeneous sources form a single operational framework for urban intelligence.

From a methodological perspective, data federation is achieved through semantic interoperability and standardized metadata models. Ontology-driven systems allow the mapping of diverse data schemas into a shared semantic structure, enabling coherent interpretation across domains. This is particularly relevant for LST, where values can vary depending on acquisition geometry, atmospheric correction algorithms, or sensor calibration. The federated system abstracts these differences, ensuring that temperature maps from various missions such as Landsat, MODIS, or Sentinel can be seamlessly compared and jointly analyzed.

Such integrated datasets are crucial for monitoring and mitigating urban heat islands, one of the most pressing sustainability challenges for modern cities. By cross-referencing LST patterns with indicators of vegetation health, surface materials, and social vulnerability, decision-makers can identify priority areas for cooling interventions such as green roofs, reflective pavements, or urban reforestation. Moreover, federated environmental data can support financial and ESG reporting frameworks, where metrics related to climate resilience, energy efficiency, and environmental impact are increasingly required by investors and regulators.

The key players in this federated ecosystem include:

• Big Data providers and research institutions, responsible for managing and preprocessing satellite and sensor datasets.
• Service providers and technology firms, such as Latitudo40, that develop AI-based tools for data harmonization, visualization, and analysis.
• Local communities and municipalities, which contribute ground truth data and contextual knowledge essential to validate and interpret satellite observations.

The collaboration among these actors allows the creation of dynamic, multi-scale datasets capable of bridging the resolution-frequency gap in current EO systems. The resulting framework delivers actionable intelligence for both operational and strategic applications—from climate adaptation planning to compliance with ESG indicators.

In essence, federated data integration transforms fragmented satellite data into a unified decision-support infrastructure. It translates complex environmental signals into measurable, comparable, and traceable insights, empowering cities, companies, and institutions to design data-driven policies and sustainable development strategies grounded in real, spatially explicit evidence.

BeatTheHeat: a federated approach to monitor urban heat islands

BeatTheHeat is an EU-funded project developed under the broader Data Space for Smart and Sustainable Cities and Communities (DS4SSCC) initiative. Its purpose is to strengthen urban resilience and sustainability by improving how cities access and use environmental information derived from Earth Observation and satellite data.

The initiative focuses on one of the most critical environmental indicators for urban planning — the land surface temperature (LST). The project aims to build a dedicated LST layer through the creation of a federated database, a virtual system that integrates heterogeneous data sources while keeping them interoperable and accessible through a single platform. This approach enhances both the spatial accuracy and temporal continuity of temperature data, overcoming the traditional fragmentation of existing datasets.

The federated layer developed within BeatTheHeat is being tested and validated across a set of pilot cities—Helsinki, Flanders, and Murcia—each chosen for its distinct climatic and urban characteristics. These testbeds allow researchers and local authorities to evaluate how integrated thermal information can support decision-making in areas such as urban heat island monitoring, risk prevention, and sustainable land-use design.

Latitudo 40 contributes to the project by providing its expertise in geospatial data processing and AI-based analytics. The company develops the data integration and visualization components that make complex satellite-derived information accessible to non-expert users. Through its proprietary algorithms and platform capabilities, Latitudo 40 helps transform raw LST measurements into operational tools that cities can use to plan cooling strategies, assess heat vulnerability, and track the effectiveness of climate adaptation policies.

BeatTheHeat demonstrates how federated data ecosystems can translate advanced satellite data into actionable urban intelligence, setting a replicable model for sustainable and data-driven city management across Europe._‍

From thermal mapping to virtual constellations: the future of Earth Observation

The experience gained from BeatTheHeat highlights a new paradigm for Earth Observation (EO): one where innovation no longer depends solely on launching new satellites but on reimagining how existing satellite data are processed, integrated, and interpreted. As urban environments become more complex and climate impacts more dynamic, the ability to generate accurate, continuous, and scalable information layers will be decisive for the next generation of urban planning and datasets for sustainability.

The federated approach adopted for land surface temperature (LST) mapping demonstrates that high-quality environmental intelligence can emerge from the integration of multiple heterogeneous data streams. This principle can now be extended far beyond thermal monitoring, to domains such as carbon storage estimation, vegetation health assessment, and risk prediction models for flooding, drought, or air quality degradation. By combining data from optical, thermal, radar, and LiDAR sensors, federated systems enable a multidimensional view of environmental processes that supports cross-sector ESG strategies and evidence-based policymaking.

A particularly promising evolution of this approach is the concept of the virtual constellation. Instead of relying on a single mission, virtual constellations use AI algorithms and data fusion techniques to synthetically combine outputs from multiple satellite platforms. The result is a new generation of virtual datasets—continuous, harmonized, and statistically consistent—that replicate the performance of a dedicated physical constellation without the cost or delay of deployment. These synthetic layers can capture daily LST variations or other key indicators at unprecedented spatial detail, providing the temporal density and precision required for operational applications in urban resilience and sustainability.

As cities and industries move toward data-driven climate adaptation, virtual constellations represent the next logical step in the evolution of Earth Observation, transforming fragmented information into cohesive, predictive, and actionable insight for a more sustainable planet.

‍