#1558 - Fiji - Cyclone Winston - Rakiraki Area - Buildings/Roads (Archived)

Please donate!
russdeffner validated #98

Nedran validated #97

Nedran validated #114

russdeffner validated #79

rivw validated #222

rivw validated #219

rivw validated #218

rivw validated #244

rivw validated #269

rivw validated #285

rivw validated #267

hebolz validated #59

hebolz validated #58

hebolz validated #192

hebolz validated #189

hebolz validated #216

hebolz validated #243

hebolz validated #268

hebolz validated #266

rivw validated #240

This project was archived.

Cyclone Winston hit Fiji on 20 February, leaving at least 21 dead and destroying hundreds of homes.

The Category Five storm (the most powerful on the Saffir-Simpson Scale) is the worst to affect Fiji in decades. It was also one of the most powerful storms the southern hemisphere in recorded history.

Winston made landfall at 06:30 UTC (18:30 local time) on the north of Viti Levu, the main island in the archipelago. From there it moved west, and the large storm narrowly avoided the capital city of Suva.

The government declared a state of emergency and imposed a 36-hour curfew for residents' safety.

The storm has reportedly destroyed entire villages and many homes were damaged or destroyed. Thousands of people are currently in hundreds of evacuation centres until they can return to their homes, if possible.

Winston also damaged power lines, leaving 80% of the population without power and impacting communications across the more than 300 islands in the archipelago. This limited or lack of communication is making it difficult to assess the full scale of the damage in the more remote islands. It is believed that some of the islands may have suffered far more damage than the main island. There are reports that Koro Island, one of the larger islands, may have suffered "catastrophic" damage.

Repair and recovery efforts began on 22 February, and rescue workers priorities are restoring communications and provided clean food and water to the many affected by the cyclone.

Through the MapGive project, the Humanitarian Information Unit (HIU) of the U.S. Department of State is providing the OpenStreetMap community access to updated satellite imagery services to help assist with humanitarian mapping.

Created by bgirardot - Updated - Priority: low


Entities to Map
buildings mainly, roads as needed
Changeset Comment
#hotosm-project-1558 #MapGive Source=WorldView-2, DigitalGlobe, NextView, 19 Sep 2014 #cyclone_winston building and road mapping.
When saving your work, please leave the default comment but add what you actually mapped, for example "added buildings and a residential road".

[imagery url available only after accepting the license]

Access to this imagery is limited by the NextView license agreement.

You need to review and acknowledge the agreement.

Project Specific Mapping Notes

  • UPDATED UPDATED: Some imagery loading problems, switch to Bing if the provided imagery seems too low resolution (iD) or shows errors (JOSM). (iD -> Background Settings (on right, icon looks like 3 stacked sheets) -> Select Bing Aerial)
  • If needed align provided imagery to objects already mapped with Bing imagery (iD -> Background Settings (on right, icon looks like 3 stacked sheets) -> Fix alignment -> use arrows)
  • There are sections of forest with no buildings, just look carefully and then mark task square "Done" and leave a comment "No buildings"
  • Some of the roads are mapped, you will see unmapped roads, please map them. No need to change any road classifications that already exist. See road mapping notes below.
  • JOSM users should put the correct source from the comment field into their changeset upload in the 'source' field and remove it from the comments field.


  • Please accurately outline all the buildings you can find. The outline should be for the full size of the building even if it is partly covered by trees in the imagery.
  • After drawing the outline, use the 's' key in the web editor to "square" the corners. In iD you have to select the building again after tagging it a building.
  • Many buildings are very close, but do not actually touch each other, try to map them close to each other without letting them connect or share nodes with each other, roads or residential area outlines. Zooming in very close to map helps with this!
  • In the iD web editor, the first time you map a building you will use the "Building" category and then at the bottom of the list select "Building" again, this is the most generic building tag we can use as we almost never can tell the more specific use of any building from imagery alone.

Roads and Paths

  • Please map roads and major paths as completely as possible by connecting them to the main road networks where ever possible.
  • Very short segments that do not connect to anything should not be mapped.
  • Short segments of a road or path that you cannot see through the trees but seem very likely to exist should be mapped to keep the road complete
  • Try and map a little beyond your task square so the person who maps the task square next to yours can easily connect them.
  • Always connect roads to other roads where they meet and never end roads on Residential area polygon edges.

You will most often use the following road classifications:

  • highway=residential - Only in the city area for roads that just access buildings, most other roads are one of the classifications below.
  • highway=unclassified - This is for roads that a vehicle with 4 wheels could travel on that connect small villages and settlements or groups of buildings in rural areas.
  • highway=track - This is for roads that a 4 wheel vehicle or farm tractor could travel on and only lead to farm fields or forests. They do not connect settlements.
  • highway=path - This is for small paths that a 4 wheeled vehicle could not travel on and are usually used by people on foot or 2 wheeled vehicles. They may or may not connect settlements or lead to farm fields. Only map the major paths that connect to settlements or other roads.

Imagery Notes

This imagery is from Digital Globe’s WorldView-2 satellite, and is made up of an image taken on September 19th 2014. The image is orthorectified for terrain corrected geographic precision. Additionally, the image has been contrast stretched using a custom stretch and processed into a Tiled Map Service (TMS) for performance.

  Start contributing

Contributors Done Assigned