A new global dataset combines satellite observations, building-height data, and regional construction archetypes to show how researchers can track eight key materials across the world’s building stock.

Paper: MaterialCities: mapping building material stocks and floor area of the world’s 1.8 million settlements by structures and functions. Image Credit: Joyfull / Shutterstock
*Important notice: This news reports on an unedited version of an accepted paper and is awaiting final editing. Therefore, the paper should not be regarded as conclusive or treated as established information.
A new study published in Scientific Data introduces MaterialCities, a detailed global dataset that estimates and maps the floor area and construction material stocks across 1.79 million data-defined spatial units representing human settlements.
The dataset separates estimated floor area and material stocks by building function, structural system, and material composition. The data can be used to study construction material demand, resource management, and circular economy strategies.
Mapping the material resources embedded in global buildings
Buildings hold a substantial proportion of the materials used in modern societies. Concrete, steel, bricks, timber, and other construction materials are used in housing, commercial buildings, industrial buildings, and other non-residential uses. Producing these materials consumes large amounts of natural resources and energy and contributes to environmental degradation and greenhouse gas emissions.
Material stock assessments quantify the resources stored in buildings and identify opportunities for reuse and recycling. They can also help researchers estimate future construction demand, demolition waste, and the supply of secondary raw materials. Previous studies have faced several limitations. Some provide detailed geographical information but do not distinguish between building functions or structural systems.
These limitations make it difficult to compare material stocks across regions and assess how construction practices affect resource use. Building function and structural type are important because they influence material requirements, building lifespans, and opportunities for material recovery.
To address these gaps, the researchers developed MaterialCities, a global dataset that links building floor area with construction material stocks. These estimates allow users to compare building material stocks across individual settlements, countries, and larger geographical regions.
Combining satellite data and material intensity estimates
The researchers developed MaterialCities by combining nighttime-light observations, global built-up-surface and building-height datasets, regional construction classifications, and material intensity estimates. Local building-footprint inventories were used to calibrate and validate the global floor-area estimates. The dataset represents building stocks around 2020, based on the coverage of the underlying data sources.
They first used nighttime light observations from the Visible Infrared Imaging Radiometer Suite (VIIRS) and a watershed segmentation algorithm to delineate settlement units. This process produced 1,790,264 Nighttime Light Cells (NLCs), covering more than 11 million km². These are data-defined spatial units: a large city can contain several NLCs, while adjoining municipalities can fall within one NLC.
The researchers then estimated the gross floor area (GFA) by combining a global built-up surface dataset with building-height estimates from World Settlement Footprint 3D (WSF3D) and the Global Human Settlement Layer (GHS-H). Validation showed that WSF3D performed better overall, so it supplied the height estimate for 98.4% of NLCs; GHS-H was used for the remaining 1.6%.
The team classified the estimated floor area into three building functions: residential single-family, residential multi-family, and non-residential. Each group was further divided into five structural categories: reinforced concrete, masonry, steel frame, timber, and unspecified construction, producing 15 combinations.
To calculate material stocks, the researchers used regional material intensity data from the RASMI dataset. The assessment covered eight materials: concrete, bricks, wood, steel, glass, plastics, aluminum, and copper.
They multiplied each building category's floor area by its corresponding material intensity and generated five estimates based on the 5th, 25th, 50th, 75th, and 95th percentiles of the RASMI material-intensity distributions. These variants reflect variability and uncertainty in the material-intensity coefficients; they are not confidence intervals or formal descriptive statistics of the actual global building stock.
Global building stocks and regional construction patterns
MaterialCities estimates the world's total gross floor area (GFA) at 1.127 trillion m² across 1.79 million NLCs representing settlements. By allocating estimated floor area across function and structural categories, the dataset shows differences in construction patterns across regions.
Residential single-family buildings account for about 67.1% of total global floor area, based on the rounded category values reported in the study.
Of total global GFA, residential single-family buildings in reinforced concrete account for 16.5%, masonry for 28.8%, and timber for 17.3%. Residential multi-family buildings in reinforced concrete account for 13.7% and masonry for 8.5%. These variations show why both building function and structural type matter when estimating material requirements.
The dataset also captures regional differences in construction practices. In an eastern Mediterranean example covering 16 countries and territories, MaterialCities indicates that Cyprus and several Levantine countries use steel-frame structures in non-residential buildings but have no recorded timber structures in this category. Turkey showed greater use of timber and less reliance on steel frames.
The material-stock estimates also depend on assumptions about material intensities, so the authors report several variants to convey their variability and uncertainty. The paper treats p50 as the central estimate and recommends using p25 and p75 with it to communicate variability and uncertainty.
The p05 and p95 variants provide a wider range but can be more influenced by outlying material-intensity values. They compared these estimates with 30 previous studies, covering 122 material stock estimates. The comparison revealed no systematic overestimation or underestimation. Sixty-nine of the 122 comparison values fell within the MaterialCities p25-p75 range. The authors describe this comparison as a reasonableness check and state that accuracy and precision can only be established through ground-truthing.
Around 4.1% of global floor area belonged to unspecified construction categories and could not be converted into material-stock estimates because suitable material intensity data are lacking. This category includes informal, indigenous, and other construction styles for which RASMI does not yet provide material-intensity coefficients.
MaterialCities uses a consistent method to estimate the location, quantity, and composition of materials in the global building stock.
Supporting circular construction and global resource planning
MaterialCities provides modeled estimates of construction materials in buildings worldwide. By linking gross floor area (GFA) with building functions, structural systems, and regional material intensities, it offers a consistent basis for assessing global building material stocks.
Several applications in sustainable resource management can draw on these estimates. Their detailed information on material quantities and locations can help identify urban mining opportunities. Urban mining involves recovering materials from existing buildings through renovation, demolition, and recycling. This information can also guide construction waste management and improve the recovery of reusable materials.
For studies of future construction demand and environmental impacts, MaterialCities can provide a baseline. Researchers can combine MaterialCities with scenarios and other models to examine resource consumption, energy requirements, and emissions associated with different construction systems.
In addition, building material data could support disaster planning by helping estimate construction rubble from natural disasters and human conflicts.
Users can also examine material stocks at different spatial scales, from individual settlements to countries and global regions. Its accompanying data formats and interactive platform also make it easier to filter and analyze specific material categories. The full browser platform can be run locally from the Zenodo record, while the online demo has limited functionality.
Further improvements depend on better building footprint, geometry, height, and floor-area data, along with more calibration and ground-truth validation. Expanding material intensity data for informal and less-documented buildings would also improve coverage.
MaterialCities gives researchers a shared global reference point for comparing building material stocks and testing future scenarios. Its spatial estimates can help researchers and decision-makers plan material recovery, compare construction practices, and assess resource use across places.
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.
Source:
- Fishman, T., Peled, Y., et al. (2026). MaterialCities: mapping building material stocks and floor area of the world’s 1.8 million settlements by structures and functions. Scientific Data. DOI: 10.1038/s41597-026-08719-6, https://www.nature.com/articles/s41597-026-08719-6