Map data, elevations and shadow model
Where our maps and elevations come from, which details the shadow calculation can resolve and where its limits lie.
Current: terrain and canopy heights
Our best current shadow model combines Mapterhorn ground terrain with canopy heights from Meta / WRI CHMv2. These additional tree heights are already active wherever suitable ground data and valid canopy data are both available.
Canopy heights in verified areas: Switzerland, Liechtenstein, Germany, France, Italy, Austria, Netherlands, Norway, Portugal, Spain, Czechia, United Kingdom, USA.
Verified on: Sept. 8, 2026. The countries listed have canopy data in some areas, not nationwide coverage. Further suitable areas are being added gradually.
Meta / WRI CHMv2 tree heights are estimates from satellite imagery on a 1 m source grid. We retain the tallest crowns per grid cell and add them only above verified bare-earth terrain to avoid counting trees twice. Tree heights may differ by several metres; no error valid for every location is known. Recent felling, individual branches and crown transparency are not reliably represented.
Where this combination is missing, the map continues to show terrain shadows from Mapterhorn, with AWS Terrain Tiles as a fallback when needed. A country's presence in the source catalog alone does not confirm usable canopy heights at the selected location.
What the shadow model can actually resolve
The browser checks whether terrain or available tree crowns block the sun. National data is reprojected and resampled more coarsely for the map: near the view, the ground used in Switzerland has roughly 13 m spacing and the finest crown grid roughly 6.5 m. These values vary with latitude and zoom. Even a 0.5 m source is therefore not evaluated here as a 0.5 m shadow model.
A separate building model is currently disabled, including in Switzerland. Prepared Swiss data is retained. The finer national sources in the table therefore do not guarantee building or individual-tree shadows.
Shadow calculation pauses when zoomed far out. It uses available detail nearby and coarser terrain up to 100 km away. Sun times for the map centre use the finest available data regardless of zoom. Weather and clouds are not calculated; times and shadow edges remain planning aids.
Expansion and first visit
Switzerland and eligible areas around existing photospots are prepared. Expansion continues outwards from Switzerland and in parallel in the USA from West Lafayette. Further eligible regions follow. A country source in the table means available elevation data, not a completed crown download.
Missing crown tiles are prepared on the first visit with shadows enabled. While “More detailed data is being downloaded.” and the progress bar appear, the map already shows terrain shadows. Tiles are added automatically when ready and saved for later visits. Without suitable ground or crown data, the terrain model remains.
Planned: finer national models
Direct integration of finer national elevation and surface models is planned. These models are currently inactive as a separate shadow layer; the prepared Swiss model is also switched off. National data already used by Mapterhorn remain part of the current terrain base.
Grid spacing and elevation error are different things. A 1 m grid contains an elevation every metre; it is not automatically accurate to 1 m. Survey method, acquisition year, vegetation and steep slopes affect the error. National models have no single elevation error valid for an entire country. The linked providers describe the quality of their surveys.
The following overview documents national sources from the Mapterhorn catalog checked on 7 September 2026 for the planned expansion. Grid spacing describes the source data, not the current resolution of our shadow model. The finest value often applies only to some areas.
View Mapterhorn coverage · Source catalog
| Country or territory | Source grid spacing | Coverage and sources |
|---|---|---|
| Australia | 2 / 5 m | Selected areasProviders, datasets and licences
|
| Belgium | 0.50 / 1 m | Regional modelsProviders, datasets and licences
|
| Denmark | 0.40 m | National modelProviders, datasets and licences
|
| Germany | 0.25 / 1 m | Regional modelsProviders, datasets and licences
|
| Estonia | 1 m | National modelProviders, datasets and licences
|
| Finland | 2 m | National modelProviders, datasets and licences
|
| France | 0.50 / 1 / 5 m | National model, finer in some areasProviders, datasets and licences
|
| Faroe Islands (Denmark) | 10 m | National modelProviders, datasets and licences
|
| Greenland (Denmark) | 2 m | National modelProviders, datasets and licences
|
| Iceland | 10 m | National modelProviders, datasets and licences
|
| Israel | 10 m | National modelProviders, datasets and licences
|
| Italy | 1 / 2 / 2.50 / 5 / 10 m | National model, finer in some areasProviders, datasets and licences
|
| Japan | 1 / 5 / 10 m | National model, finer in some areasProviders, datasets and licences
|
| Canada | 2 m | Selected areasProviders, datasets and licences
|
| Latvia | 20 m | National modelProviders, datasets and licences
|
| Luxembourg | 0.50 m | National modelProviders, datasets and licences
|
| New Zealand | 1 m | Selected areasProviders, datasets and licences
|
| Netherlands | 5 m | Selected areasProviders, datasets and licences
|
| Norway | 1 m | National modelProviders, datasets and licences
|
| Austria | 1 / 5 / 10 m | National model, finer in some areasProviders, datasets and licences
|
| Poland | 1 / 5 m | National model, finer in some areasProviders, datasets and licences
|
| Portugal | 0.50 / 10 m | Selected areasProviders, datasets and licences
|
| Romania | 0.50 m | Selected areasProviders, datasets and licences
|
| Rwanda | 10 m | National modelProviders, datasets and licences
|
| Sweden | 1 m | National modelProviders, datasets and licences
|
| Switzerland | 0.25 / 0.50 m | National model, finer in some areasProviders, datasets and licences
|
| Slovakia | 1 m | National modelProviders, datasets and licences
|
| Slovenia | 1 m | National modelProviders, datasets and licences
|
| Spain | 0.50 / 2 / 5 m | National model, finer in some areasProviders, datasets and licences
|
| Svalbard (Norway) | 20 m | National modelProviders, datasets and licences
|
| Taiwan | 20 m | National modelProviders, datasets and licences
|
| Czechia | 5 m | National modelProviders, datasets and licences
|
| USA | 1 / 10 m | National model, finer in some areasProviders, datasets and licences
|
| United Kingdom | 1 m | England, Wales and parts of ScotlandProviders, datasets and licences
|
| Cyprus | 1 m | National modelProviders, datasets and licences
|
These values come from the source catalog, not the displayed map pixel size. For example, the catalog lists 1 m for Slovenia while the provider overview already lists 0.5 m. The source actually available at each location remains decisive.
Morocco and other areas using the global base
Morocco remains covered by Mapterhorn. Where no finer national model exists, we use Copernicus GLO-30 with roughly 30 m spacing. It is a surface model: trees and buildings may already be included in its elevations. It represents mountains and valleys, but not reliable individual walls, trees or narrow cuttings. We therefore add no further tree heights there; Morocco is excluded from the planned crown expansion.
Copernicus specifies an absolute vertical error below 4 m at the 90 percent error level. This is a global product specification, not a guarantee for every point in Morocco. Acquisition was mainly in 2011–2015, with some gaps filled by older data. This can improve elevations over the previous AWS/SRTM data; we have not measured a local improvement.
Map sources and licences
The map draws nothing itself, it assembles other people's map services. The edge of the map always credits the layers that are currently visible. Here they are in full, in the wording each provider asks for.
The base maps "Map", "Satellite" and "Terrain" come from Google Maps, which writes its own credit into the map image. Without a Google key, and for search engine crawlers, the same map shows free sources: OpenStreetMap (© OpenStreetMap contributors, ODbL), Esri World Imagery ("Tiles © Esri, Maxar, Earthstar Geographics") and OpenTopoMap ("© OpenStreetMap contributors, SRTM · © OpenTopoMap (CC-BY-SA)"). Roads and place names over views that carry none of their own then come from Esri World Transportation and Esri World Boundaries and Places ("Tiles © Esri").
The extra layers "Swisstopo national map", "Swisstopo aerial image" and "Swisstopo hiking trails" come from the Federal Office of Topography: "© swisstopo". The "track rendering" is OpenRailwayMap: "Data © OpenStreetMap contributors, Style: CC-BY-SA 2.0 OpenRailwayMap". So the tracks are OpenStreetMap data and the rendering is under CC BY-SA 2.0.
Mapterhorn · Copernicus · Meta / WRI
Mapterhorn supplies the preferred elevation data. National providers and their licences are linked in the country table. Elevations have been reprojected, resampled and processed for our shadow calculation. The providers have not reviewed RailPics’ calculation.
Mapterhorn · © swisstopo · Open Government Data
produced using Copernicus WorldDEM-30 © DLR e.V. 2010-2014 and © Airbus Defence and Space GmbH 2014-2018 provided under COPERNICUS by the European Union and ESA; all rights reserved
© Meta / World Resources Institute (WRI), 2026 · CHMv2 · CC BY 4.0
AWS Terrain Tiles · Tilezen
If Mapterhorn does not respond, AWS Terrain Tiles provide a fallback. Depending on location and zoom they include Copernicus EU-DEM, SRTM and GMTED2010, among others. Sources are older and some grids coarser; zooming in does not create new terrain detail.
- Produced using Copernicus data and information funded by the European Union - EU-DEM layers.
- GMTED2010 and SRTM terrain data courtesy of the U.S. Geological Survey.
- © offene Daten Österreichs – Digitales Geländemodell (DGM) Österreich · CC BY 3.0 AT
This elevation data is not used unchanged: for the tiles it was reprojected, resampled and combined from several sources, and the evaluation for light and shadow is RailPics' own. Neither the European Union nor the European Environment Agency nor the U.S. Geological Survey produced or reviewed it. The dataset's full list of sources is kept by the project itself: Tilezen, attribution.md.