Authentic AirQuality GIS home screen used as the project hero.
ENVIRONMENTAL GIS / DECISION SUPPORT

Belgrade Air Quality GIS

Environmental measurements, spatial predictors and validated NO2 modelling in one operational GIS workflow.

Belgrade Air Quality GIS integrates city-scale environmental data, validated NO2 prediction models and desktop/Web GIS workflows for scenario analysis and planning-related reporting. Its operational status was reconfirmed with the client in September 2026.

LocationSerbia / City of Belgrade
ClientCity Administration of the City of Belgrade - Secretariat for Environmental Protection
MapSoft roleProject implementation lead and supplier; responsible for all project activities except the prediction-model activity group, which was implemented jointly with the University of Belgrade Faculty of Mechanical Engineering.
Project context

From sparse measurements to city-scale environmental understanding

Belgrade needed one operational environment capable of combining air-quality measurements with terrain, land use, traffic, population, heating sources and urban geometry. The challenge was not to display those datasets separately, but to turn them into a repeatable evidence base for planning-related environmental decisions.

Desktop GIS view showing the spatial distribution of air-quality measurement sites across Belgrade.
Air-quality measurement locations
Shaded terrain visualization representing the city-scale DTM used as a spatial predictor.
Digital Terrain Model shaded relief
6,328NO₂ measurements / 33 locations
10.95 µg/m³Winter NO2 model RMSE
48,302Geocoded fireplace addresses
208Traffic intersections integrated
Spatial predictor foundation

Build the model from the city around the measurement points

The modelling foundation integrated heterogeneous spatial variables that could explain pollutant distribution: traffic load, individual-heating sources, population density, land use, building use, special buildings, terrain and 3D urban geometry. MapSoft normalized these inputs into one controlled GIS/database environment.

GIS layer showing current land-use categories used as environmental modelling context.
Current land use
Road-network visualization of normalized traffic-flow values used in modelling.
Traffic flow
GIS layer classifying building use across the Belgrade area.
Building use
GIS layer showing population-density classes used in the environmental data foundation.
Population density
GIS layer showing schools, hospitals and other special-building categories.
Special buildings
Grid-based visualization of chimney counts used as a proxy for individual-heating emissions.
Chimney density grid
GIS layer showing geocoded chimney/fireplace-related locations across Belgrade.
Existing chimney locations
GIS layer showing road categories across the Belgrade network.
Traffic categories
Desktop GIS view of the terrain model integrated into the operational project environment.
DTM in the desktop GIS environment
3D urban context

Terrain and buildings were part of the analytical environment

The project delivered terrain and prismatic 3D building geometry for desktop and Web visualization, providing spatial context and analytical support rather than a full Digital Twin environment.

Detailed perspective view of the delivered 3D building model.
3D building model detail
Hillshade overview of the Belgrade terrain model used for spatial analysis.
DTM hillshade overview
NO₂ modelling & air-quality intelligence

Turn measured values and spatial predictors into comparable city-wide surfaces

MapSoft and the University of Belgrade Faculty of Mechanical Engineering jointly developed and evaluated the prediction-model activity group. Validation supported reliable winter and summer NO₂ models, while the available data for other pollutants was insufficient for equally dependable final prediction.

Final prediction focused on NO₂, where the available monitoring data supported reliable seasonal modelling.

Desktop GIS visualization of a city-wide NO₂ prediction surface for existing conditions.
NO₂ prediction surface for existing conditions
Summer-period NO₂ prediction visualization covering the Belgrade study area.
NO₂ prediction — June to August
Winter NO₂ modelRMSE 10.95 µg/m³29 stations; adjusted R² 0.50.
Desktop GIS view showing air-quality index classes derived from the NO₂ prediction surface.
Air-quality index estimated from prediction
Operational planning workflow

The result was a working case process — not a standalone model

For an active planning case, users could define the spatial extent, update planned-state layers, run the air-quality calculation, compare existing and planned conditions, generate cartographic and tabular reports and retain the case documentation for later reference.

  1. 01Define case extent and metadata
  2. 02Update planned-state layers
  3. 03Run air-quality model/index calculation
  4. 04Compare existing and planned conditions
  5. 05Generate cartographic and tabular reports
  6. 06Archive completed case documentation
Desktop GIS view combining planned-condition air-quality index analysis with the generated case report.
Planned-condition analysis and report
Web GIS screen showing an active case, thematic layers and object information over the map.
Web GIS case interface
Desktop + Web GIS

One environmental workflow across data maintenance, analysis, 2D/3D Web access and reporting

MapSoft implemented the central data model, customized the desktop GIS workflow, developed the Web GIS/Geoportal and server-side modelling/reporting functions, and connected the environment through documented spatial and application interfaces.

Implemented environmentMicrosoft SQL Server 2016 Express · CadCorp SIS MapEditor 8 · CadCorp GeognoSIS 8 · MapSoft Geoportal · IIS · WMS/WFS · SOAP/WCF

A documented API plus SOAP/WCF and WMS/WFS interfaces support connection with external information systems and spatial services; heterogeneous vector/raster/source datasets were normalized into the central environment.

Authentic 3D Web GIS scene showing thematic air-quality values over Belgrade building geometry.
3D Web GIS view with NO₂ thematic visualization
Quality & validation

Validate each model against the available monitoring data

Validation covered the NO₂ models and DTM quantitatively, together with geometric, topological, thematic and database-integrity controls. The same process identified where monitoring data was insufficient for dependable prediction.

Winter NO₂ modelRMSE 10.95 µg/m³Quantitative validation documented in the final realization sources.
PM10, SO₂ and Pb datasets were assessed as insufficient for reliable final prediction and were therefore excluded from the final dependable model set.
Delivered outputs

Data, models and software delivered as one decision-support environment

The project combined city-scale spatial datasets, validated analytical surfaces, a central database, desktop/Web GIS applications, reporting services, deployment, training and support.

Shaded terrain visualization representing the city-scale DTM used as a spatial predictor.
GeoTIFF

Digital Terrain Model

10 m DTM covering 3,200 km².

10 m; mean height error 1 m · 3,200 km²
Detailed perspective view of the delivered 3D building model.
DWG

3D building model

Prismatic building volumes for desktop/Web 3D context.

GIS layer showing current land-use categories used as environmental modelling context.
GeoTIFF

Land-use raster

5 m General Plan-area land-use dataset classified to CLC categories.

5 m
Desktop GIS visualization of a city-wide NO₂ prediction surface for existing conditions.
Raster

Validated NO2 prediction surfaces

Winter and summer 100 m prediction rasters with documented validation statistics.

100 m · 2 surfaces
Desktop GIS view showing the spatial distribution of air-quality measurement sites across Belgrade.
Microsoft SQL Server

Integrated environmental spatial database

Centralized multi-source spatial and measurement database.

Authentic 3D Web GIS scene showing thematic air-quality values over Belgrade building geometry.
Software

Desktop and Web GIS applications

Operational existing/planned-state workflow, 2D/3D visualization, queries, editing, reports and controlled access.

Desktop GIS visualization of a city-wide NO₂ prediction surface for existing conditions.
What this project proves

Environmental modelling becomes useful when it is embedded in an operational GIS process

Belgrade Air Quality GIS demonstrates MapSoft's ability to combine heterogeneous geospatial/domain data, statistical environmental modelling and custom GIS software into an operational public-administration decision-support system, with a clear academic partnership boundary for specialized model interpretation.

6,328 NO₂ measurements across 33 locationsWinter NO₂ model RMSE 10.95 µg/m³48,302 geocoded fireplace addresses208 traffic-count intersections integrated
Related projectsExplore projects →
Authentic Web GIS screen showing Belgrade environmental layers and LRIZ pollution-source information.software

Belgrade Local Register of Pollution Sources (LRIZ) — Web & GIS Platform

Modernization of Belgrade's Local Register of Pollution Sources through migration from a legacy Microsoft Access database to a server database, development of an authenticated web application, reporting, and integration with the existing desktop and Web GIS environment.

5 domainsAnnual reporting domains

View project
Presentation hero composed from the supplied EIONET logo and authentic EIS map context showing Montenegro; it is not an application screenshot.software

Montenegro Environmental Information and EIONET Reporting System

A nationwide environmental reporting platform modernizing Montenegro's Environmental Information System with role-based reporting workflows, data validation, dashboards, GIS/geoportal functions and EIONET-aligned metadata and exchange procedures.

10 modelsPriority ROD models

View project
Wetland landscape with a composited view of the NWI portal used as contextual project imagery.integrated

Serbia National Wetland Inventory — National Wetland Data & Geoportal Platform

A national-scale wetland inventory combining Earth Observation and geospatial processing, a standards-aware environmental data model and an operational Web GIS for visualization, editing, analysis, reporting and long-term change tracking.

National Republic of SerbiaNational coverage

View project

Need to connect environmental data, spatial modelling and operational decision support?

Talk to MapSoft about environmental information systems, spatial data integration, modelling workflows, desktop/Web GIS and reporting for public administration and planning.

Discuss your project