Researchers from the Norwegian University of Life Sciences (NMBU), in collaboration with the University of Coimbra in Portugal, have released an open dataset of climate files for building design and energy simulation across Europe.

The core problem is that standard weather files used by engineers (Typical Meteorological Years, or TMYs) are assembled from historical observations, often based on data that can be decades old. Buildings designed and built today are expected to perform well into the 2050s, 2080s, and beyond under a warmer climate.

This creates a mismatch that can exceed half a century between the climate used in calculations and the conditions the building will actually experience.

The platform ClimateDataForBuildings (climatedataforbuildings.eu) provides:

Typical Meteorological Year (TMY) files: Updated, location-specific files based on high-resolution CERRA reanalysis data for thousands of European locations (reports cite more than 3,600- 5,000 and sites, with around 7,200 TMY files in some descriptions). These include both a longer-term baseline (e.g., 1991–2020) and a more recent one (e.g., 15-year period) to better capture recent warming, especially rising cooling loads.

Future Meteorological Year (FMY) files: Newly released projections for mid-century (roughly 2036–2065) and late-century (2066–2095) periods. These apply climate- change signals from selected reliable global and regional models under multiple emissions pathways (CMIP5 RCP4.5/RCP8.5 and CMIP6 SSP scenarios ranging from low to high emissions).

The FMYs are built on the same foundation, resolution, reference period, and methodology as the TMYs, so historical and future files are internally consistent. Files are provided in standard EPW format, compatible with common building performance tools such as EnergyPlus, IDA ICE, TRNSYS, and others.

It enables more accurate modeling of heating and cooling loads, indoor comfort, HVAC design, solar energy, and climate- robust renovation or new construction.

Temperature rises are uneven (often stronger in winter in northern areas), which can significantly affect energy demand where heating dominates.

Design choices (e.g., shading or cooling capacity) can differ depending on the emissions scenario chosen, turning it partly into a risk- management decision.

The data is free, open, documented, and scientifically validated, addressing previous limitations in transparency of commercial or other datasets. Researchers in countries including Denmark, Poland, and Spain are already using it.

Professor Thomas K. Thiis (NMBU) and colleagues emphasize that the building sector needs data reflecting present and future conditions rather than the past, to support energy- efficient, resilient, and locally adapted buildings.

The work builds on earlier Norwegian national datasets (for all municipalities) and peer- reviewed publications describing the methodology.

In short, the open platform supplies simulation-ready climate files so Europe’s building stock can be designed and upgraded for the climate that is arriving, not the one that already no longer matches historical norms.

The full dataset is available at climatedataforbuildings.eu.

_____________________________________________________________________________________

Researchers from the Norwegian University of Life Sciences (NMBU) and the University of Coimbra (CURA Lab, led by Eugénio Rodrigues) generated the FMYs by taking the project’s historical Typical Meteorological Year (TMY) files, derived from high- resolution CERRA reanalysis data and morphing them with the open- source Future Weather Generator (FWG).

Baseline: The TMYs provide a statistically representative historical year of hourly weather data (temperature, humidity, solar radiation, wind, etc.) that preserves realistic daily and seasonal patterns and the relationships among variables.

Climate signal: Monthly climate- change deltas (additive shifts or multiplicative factors, depending on the variable) are derived from carefully selected ensembles of regional and global climate models. These include EURO-CORDEX simulations driven by CMIP5 (RCP pathways) and CMIP6 global models (SSP pathways).

Morphing: FWG applies these monthly signals to the baseline EPW files while retaining the original hourly sequence and temporal structure. This produces future weather files for mid-century (≈2036–2065) and late-century (≈2066–2095) periods under multiple emissions scenarios.

Consistency and selection: Only models that adequately reproduce European summer and winter conditions were retained. Because the same reference period, spatial foundation, and methodology underlie both the TMYs and FMYs, the climate-change signal is not distorted by methodological mismatch.

FWG itself is a free, open-source, cross- platform Java tool (developed at the University of Coimbra’s CURA Lab) that implements the established morphing approach of Belcher et al., extended with modern climate resources, spatial interpolation options, quality controls, ensemble support, and full provenance tracking.

It supports both individual models and multi-model ensembles and outputs standard EPW files ready for tools such as EnergyPlus, IDA ICE, and others.

All resulting TMY and FMY datasets (covering more than 5,000 European locations) are openly accessible at climatedataforbuildings.eu, creating a transparent, reproducible link between climate-model projections and practical building-energy and resilience simulations.

This allows designers and policymakers to evaluate heating and cooling demand shifts, overheating risk, renovation strategies, and adaptation measures under consistent present- day and future climate conditions.

_____________________________________________________________________________________

Evaluation and selection of future meteorological years for building performance simulations in Europe

The generation and selection of Future Meteorological Years (FMYs) for building simulation involves methodological challenges related to climate-model selection, scenario interpretation, spatial consistency, and engineering applicability.

This study presents a framework for generating, evaluating, and selecting FMYs and demonstrates its application across 5,028 European locations using engineering-oriented selection criteria.

Three principal contributions:

1. A scalable and transparent methodology for generating FMYs from RCM-based RCP (CORDEX-CMIP5) and GCM-based SSP (CMIP6) climate projections.
2. A performance-driven framework for selecting climate-model combinations suitable for engineering applications.
3. A scenario-specific guidance scheme that directly supports simulating heating and cooling demand in buildings under climate change.

Method: Using the open-source Future Weather Generator (FWG), historical Typical Meteorological Years (TMYs) were morphed with monthly climate deltas derived from ensembles of CORDEX-CMIP5 and CMIP6 models.

All datasets are openly accessible, providing a reproducible, climate- consistent bridge between climate science and building engineering practice.

This is the peer-reviewed paper that underpins the Future Meteorological Year (FMY) component of the open ClimateDataForBuildings platform (climatedataforbuildings.eu).

It describes how the team systematically evaluated a large pool of climate models, retained only those that adequately reproduce European summer and winter conditions, and produced consistent FMYs for mid-century (≈2036–2065) and late-century (≈2066–2095) periods under multiple emissions pathways (RCP4.5/8.5 and SSP1-2.6 through SSP5-8.5).

The resulting files maintain internal consistency with the project’s historical TMY dataset (based on CERRA reanalysis), so the climate-change signal is not distorted by methodological differences.

The work explicitly supports practical building-energy and resilience simulations while noting that morphing-based FMYs represent typical rather than extreme conditions.

Published:  Journal of Building Performance Simulation (2026)

DOI: 10.1080/19401493.2026.2697216

Provided: Norwegian University of Life Sciences

Authors: Thomas K. Thiis, Arnkell Jonas Petersen (Norwegian University of Life Sciences – NMBU), Eugénio Rodrigues (University of Coimbra / CURA Lab)

Abstract

The generation and selection of future meteorological years (FMYs) for building simulation involves methodological challenges related to climate-model selection, scenario interpretation, spatial consistency and engineering applicability. This study presents a framework for generating, evaluating and selecting FMYs for building simulation and demonstrates its application across 5,028 European locations using engineering-oriented selection criteria. The work makes three principal contributions: (i) a scalable and transparent methodology for generating FMYs from RCM-based RCP and GCM-based SSP climate projections, (ii) a performance-driven framework for selecting climate model combinations suitable for engineering applications and (iii) a scenario-specific guidance scheme that directly supports simulating heating and cooling demand in buildings under climate change. Using the open-source Future Weather Generator (FWG), historical Typical Meteorological Years (TMYs) were morphed with monthly climate deltas derived from ensembles of CORDEX-CMIP5 and CMIP6 models. All datasets are openly accessible, providing a reproducible, climate-consistent bridge between climate science and building engineering practice.