The data and methods behind the Great Vic Gravel Route

Posted byLittleMapsPosted inMapsTags:, , , , , ,

This post provides detailed information on the data and methods that were used to develop the maps and routes in the posts on the Great Vic Gravel Route and the how the gravel route was created. The Q&A page provides answers to a lot of less technical questions.

In all of these posts, I have used the terms gravel, dirt, unpaved and unsealed interchangeably to refer to any and all unsealed roads and tracks. Similarly, the terms paved, sealed, pavement and asphalt are used interchangeably to refer to all sealed roads.

Data sources and manipulation

  • All road and track data were downloaded from OpenStreetMap on 10 December 2021. Overpass Turbo was used to download the data.
  • The download included all ways tagged with highway=motorway, trunk, primary, secondary, tertiary (including link roads), residential, unclassified and track.
  • Walking paths and bike paths were not downloaded or analysed.
  • Roads and tracks were clipped to the Victorian border using a 500 m buffer to ensure that boundary tracks that crossed the Vic-SA border were not excluded.
  • Roads and tracks were grouped into three groups: sealed, unsealed and unknown surfaces, based on OpenStreetMap surface tags. These tags have been shown to be highly accurate in Victoria.
  • Tracks without a surface tag were classed as unsealed. According to the Australian tagging guidelines, tracks in Australia are unpaved unless tagged otherwise.
  • Major roads (highway=motorway, trunk, primary and secondary, plus link roads) that lacked a surface tag were classed as sealed. Most of these were in Melbourne.
  • To avoid private roads and closed tracks, all roads and tracks that were tagged with access = private or access = no in OpenStreetMap were excluded from analysis.
  • All data analyses were conducted in QGIS. The maps of gravel islands and gravel zones were developed using the excellent Disconnected Islands plugin.

Map 1. All unpaved roads and tracks in Victoria

This map shows all unsealed roads and tracks in Victoria, as described in the Data Sources section above. Roads tagged as access = private or access = no are not shown. Not all of these tracks will be open to the public. Some roads and tracks are closed seasonally and many tracks in parks and reserves are open to management vehicles only.

Map 2. All sealed roads in Victoria

This map shows all sealed roads in Victoria, as described in the Data Sources section above. Roads tagged as access = private or access = no are not shown. Most, but not necessarily all, of these roads will be open to the public.

Map 3. Paved and unpaved roads near Korumburra

This map uses a section of the sealed and unsealed road data, as described above. Some residential and service roads in the town of Korumburra did not have a surface tag. I have assumed they are sealed for illustrative purposes.

Map 4. Small gravel islands in Victoria

All unsealed roads in Victoria were analysed using the Disconnected Islands plugin. This plugin groups all interconnected roads into a single group, while disconnected roads are classified into separate groups. The input data included unsealed roads only. Sealed roads were not used. This map shows all groups that had a total road length less than 10 km. Extremely small groups < 100 m long are not shown for clarity.

Maps 5-7. Large gravel zones

Through a series of manipulations, I created a statewide dataset that contained all unsealed roads and all sealed roads ≤ 1 km long that created links between unsealed roads. This dataset was then analysed using the disconnected islands plugin to identify all groups of inter-connected unpaved roads (plus linking paved roads). Thus, all roads and tracks within each network group can be reached without riding for more than 1 km at a time on sealed roads. The total length of roads and tracks in each group was calculated, and the largest groups were identified. Map 5 shows the largest zone, the North West gravel zone. Maps 6 & 7 show all zones that contained more than 100 km of unpaved roads and tracks and linking paved roads up to 1 km long.

Map 8. The Great Vic Gravel Route

The shortest route between Mallacoota was identified using the shortest path (point to point) function in QGIS. The approach used to create the Great Vic Gravel Route is described at greater length in a separate post. The Great Vic Gravel Route was created using the fastest route option in the shortest path (point to point) function in QGIS, using speed settings of 1 km/hr for sealed roads and 50 km/hr for unsealed roads. The same route was obtained for all speed settings from 35-50 km/hr on unsealed roads (see here for more details).

Route quality control

A number of steps were taken to ensure that the roads and tracks in the Great Vic Gravel Route were mapped accurately.

  • To ensure that roads that lacked a surface tag did not influence the route, each analysis was conducted twice. In the first run, roads with no surface data were presumed to be sealed and the fastest route was identified. In the second run they were presumed to be unsealed and the fastest route was created again. The two routes were then compared. If they differed, all roads without a surface tag that were close to the route were inspected on satellite images. The correct surface tag was added on OpenStreetMap and in the QGIS database. The analysis was then re-run using the updated surface tags.
  • To ensure that all road surface tags along the route were accurate, the entire route was inspected in satellite images. Any incorrect surface tags were updated in OpenStreetMap and the QGIS database. The route analysis was then repeated using the updated data.
  • To ensure that the route did not travel along private roads (such as farm tracks and driveways), the entire route was inspected on satellite images. Any questionable sections were tagged as no access in the QGIS database. The route analysis was then repeated using the updated data. I cannot guarantee that this precaution identified all private roads.
  • To ensure that the sections of the route on public land did not travel on permanently closed roads or on roads open to management vehicles only, I compared the route against road information shown in the Victorian government’s topographic maps. All roads identified as permanently closed or open to management vehicles only were tagged as no access in the QGIS database and the analysis was re-run. I cannot guarantee that this precaution identified all roads that are closed to public access.

Each time that access or surface data were updated using the steps described above, a new route was created and the above steps were repeated on the new route. It took many iterations to create the final route. To the best of my knowledge, the final route shown includes public roads and tracks that are open to the public only.

Route access

The following access restrictions should be noted:

  • To the best of my knowledge, all roads and tracks in the Great Vic Gravel Route that are on public land are mapped as being open to the public on the Victorian government’s topographic map. However, many tracks are closed seasonally every year, especially those in Gippsland.
  • Some tracks in NW Victoria are sandy and may only be rideable on a fat bike.
  • Many small dirt roads in north-west Victoria are shown as ‘dry weather only’ on topographic maps and may not be passable after wet weather.
  • The route includes many fords across creeks and rivers which may not be passable when river levels are high. In particular, the Gippsland section includes a ford across the Snowy River at Jacksons Crossing. This ford is on a well-known 4WD Track and is described in the excellent Hema Victoria High Country Atlas and Guide. If this ford is not passable, then the Gippsland section of the route would require a major re-route to the north (across or north of McKillops Bridge) or south through Orbost.
  • Notwithstanding the checks described above, I cannot vouch for the accuracy of the Great Vic Gravel Route or other routes described in these posts. Should anyone be interested in travelling along along them, they do so at their own risk.

Second post: creating a bike route with the shortest possible distance of paved roads

This separate post provided a broader background to the methods used to develop the Great Vic Gravel Route. All of the maps, routes and tables in this post used the same data and methods as described above.

All routes were developed using the fastest route option in the shortest path (point-to-point) network analysis function in QGIS. The same input data was used for each route used except for the speed limit used on unsealed roads, which ranged from 1 to 10,000 km/hr. I ran more analyses with different speed limits than are shown in the post, and presented a small number of routes for clarity.

The final route that was used for the Great Vic Gravel Route was subject to a raft of repeated quality control measures, as described above. The access restrictions that were used to refine the GVGR were applied to all other routes in this post. Thus, any roads that were tagged as no access for the GVGR were not available for other routes. However, I did not check road access or surfaces across the entire length of the alternative routes, as I did for the GVGR. They are shown for illustrative purposes only and may not all be open to the public.

Acknowledgements

All maps and analyses are based on data from OpenStreetMap. It is a pleasure to acknowledge the work done by the many thousands of mappers who have created this invaluable resource.

More info

Leave a comment