Nigeria Floods Mokwa, 2025

Hi all, the Open Mapping Hub West and Northern Africa (part of Humanitarian OpenStreetMap Team (HOT) is responding to the flooding that occurred on 28 May in Mokwa, Nigeria. Officials have said that the floods killed more than 200 people and confirmed that 1,000 were missing. At least 121 others were injured, and the flooding destroyed 2,000 homes.

On June 3, 2025 the Western and Northern Africa Hub conducted a size-up (OSM needs assessment) of the disaster focusing in on a GDACS polygon. This found that Mokwa town was already very well mapped on OSM due to a successful Tasking Manager project 9417 from Médecins Sans Frontières (MSF) that completely intersects Mokwa Town. The project had used Maxar imagery captured in November 2020 - which was the latest available open imagery for improving OSM in that region at the time (via Bing/Esri). Thus, no immediate needs to improve OSM base map data in response to the disaster were identified.

On June 4, HOT created a custom Humanitarian Data Exchange (HDX) export to allow data users to easily download OSM data clipped to Mokwa Town. This will be available from HDX for at least six months.

On June 5, Maxar released a post-flood scene captured on 2 June under a Creative Commons Attribution Non Commercial 4.0 International following a request from HOT the day before. This scene was then added to OpenAerialMap.

On June 9, the Western and Northern Africa Hub spoke to a representative of Geohazards Risk Mapping Initiative who shared that for their longer term flood recovery efforts it would be valuable for damage tags to be added to building footprints using the post-flood Maxar imagery.

After consulting local organizations and locally based individuals a Tasking Manager project restricted to manually selected very experienced mappers has now been launched to coordinate this damage mapping. Please find the dedicated Wiki page for this response for more info and additional links: https://wiki.openstreetmap.org/wiki/Organised_Editing/Activities/Nigeria_Mokwa_Flood_Recovery_2025

Of key importance, the damage mapping approach will be as follows:

The boundary of this project is focused on UNOSATs flood extent data , derived from the post-flood Maxar scene, ensuring that mapping efforts are targeted to areas affected by the disaster.

The area has been well mapped already on OSM, so the goal here is to add relevant tags to existing OSM building footprints based on the latest damage.

On comparing the custom (post flood) imagery with mapped footprints and Esri World Imagery, if you spot damage or destruction please tag the relevant building footprint with:

  • *building=yes*
  • *damage=minimal*OR*damage=significant*OR*damage=complete*(based on * the * *BAR methodology
  • *source:damage:date=2025-05*
  • *source:damage=Maxar Imagery*

Please pay special attention to any buildings that existed in OpenStreetMap (OSM) prior to the flood but that are no longer visible in the Maxar Imagery provided, they have a high likelihood of being destroyed or heavily damaged (i.e. damage=complete).

Do not delete any building footprints, even if they seem completely destroyed , the resulting data will be used for recovery and it is important to know what existed before the flood event (baseline data).

Please let us know if you have comments or questions about this mapping activity.

1 Like

what with completely destroyed buildings? Marking them as building=yes seems wrong

2 Likes

Good question - I’d refer to this discussion and particularly this line:

When a building is than removed fully - when there is even no rubble anymore to see on the ground - I would propose for these damage tags to be removed and the building can than be retagged with the destroyed:building=* or just deleted from OSM.

So while there is still rubble and potentially walls or structures that cannot be seen overhead then it remains appropriate to leave building=yes

1 Like

Idk what mental gymnastics you are doing, but a pile of rubble isn’t a building.

It seems to me that the instructions on the tasking manager could be improved.

On comparing the custom (post flood) imagery with mapped footprints and Esri World Imagery, if you spot damage or destruction please tag the relevant building footprint with:

  • building=yes
  • damage=minimal OR damage=significant OR damage=complete (based on this interpretation of the BAR methodology)
  • source:damage:date=2025-06
  • source:damage=Maxar

retagging as building=yes seems wrong. damage=complete should also result in the building=* key being changed to destroyed:building=*.

2 Likes

To be clear, no new building=yes tags are being added. This area was well mapped before the event and the pre-event building=yes footprints are already there. Damage attributes are being added to those footprints (or not if there is no visible damage).

In the BAR methodology that is being followed, the prior category to damage=complete states:

This category … signifies that partial but extensive visible damage has been sustained. In these structures, the roof is entirely damaged or missing. The walls of the structure remain upright. However, the interior wall partitions can be partially damaged. Debris inside the structure can also potentially be visible. The structure does not appear to have complete structural integrity and is in need of significant repair.

Any more damage than that will be tagged damage=complete. So damage=complete will include buildings with partially upright exterior walls (i.e. not leveled). See examples of this from (page 3 of this doc)

It is not easy to distinguish buildings into damage categories using 50cm satellite imagery, and if we were to split the final damage category into building=yes damage=complete for less severe damage and destroyed:building=yes damage=complete for even more severe damage I worry that we might push the limits of what can be distinguished from a top-down 50cm view.

Hi @SColchester

I copy a quote from that other discussion:

So maybe damage=complete should be rather mapped as building=ruins than?

Than you would have for

  • The roof and walls remain largely intact: building=yes + damage=minimal
  • The roof is entirely damaged or missing. The walls of the structure remain upright: building=yes + damage=significant
  • The roof is completely destroyed or missing, and the walls have been destroyed or collapsed. The support structures are completely levelled. building=ruins
  • When there is even no rubble anymore to see on the ground: destroyed:building=* or just deleted from OSM

And in addition of course the source tagging for damage

  • source:damage:date=2025-06
  • source:damage=Maxar

What do you think?

The wiki says that “in theory it should only be used for buildings constructed to look like ruins.”.

3 Likes

Is damage=complete for pile of rubble? Then any building= value will be wrong. At least ruined:building or similar

2 Likes

The instructions say explicitly to tag damaged buildings with building=yes , which should be changed.

According to the wiki page for destroyed:*, it’s fine if remains are present. Keeping building=* for effectively destroyed structures which require “comprehensive reconstruction or demolition of the entire structure” seems strange.

@citrula agree that the instructions could be clearer there, I’ve updated by adding this new line (where my cursor is):

3 Likes

Ah yes indeed, I should have looked that up again, apologies. I like however that this would be changed to @Mateusz_Konieczny his suggestion of building=sham_ruins but maybe let’s leave that for another discussion :slight_smile:

@SColchester I went checking the BAR methodology again, which for and damage=complete it would be:

Critical Visible Damage: This category, classified by the color red, signifies severe visible damage has been sustained. In these structures, the roof is completely destroyed or missing, and the walls have been destroyed or collapsed. The support structures are completely leveled, and interior objects have also suffered visibly heavy damage or destruction. The structure does not appear to have any structural integrity and requires comprehensive reconstruction or demolition of the entire structure.

So no standing walls and thus basically a pile of rubble in my opinion. When there are still walls standing it would be rather damage=significant.

This saying, I however do think there is value in making a difference between a pile of rubble and a building that is completely gone.

I do another suggestion - because I understand that mapping building=* + damage=complete sounds like nonsense to people :grin:

What do you think of this?

  • The roof and walls remain largely intact: building=* + damage=minimal
    → not changed from my initial suggestion

  • The roof is entirely damaged or missing. The walls of the structure remain upright: building=* + damage=significant
    → not changed from my initial suggestion

  • The roof is completely destroyed or missing, and the walls have been destroyed or collapsed. The support structures are completely levelled: destroyed:building=*
    → The wiki page describes that remains might still be present so that can work I think. And I decided not to suggest here ruined:building=* because of the potential confusion with buildings constructed to look like ruins.

  • When there is even no rubble anymore to see on the ground: was:building=* or just delete from OSM
    → it seems better to use this according to this wiki page: Key:was:* - OpenStreetMap Wiki

1 Like

Ah yes indeed, I should have looked that up again, apologies. I like however that this would be changed to @Mateusz_Konieczny his suggestion of building=sham_ruins

absolutely, there are so many building=ruins that are just buildings in ruinous state, and very few that are actually built as ruins. There’s a quarter of a million of these. https://taginfo.openstreetmap.org/tags/building=ruins

2 Likes

I suppose a lot of these building=ruins are also those with historic or touristic value. Which is of course as well different than we are speaking of here.

I suppose a lot of these building=ruins are also those with historic or touristic value.

maybe, some probably are, 250000 are a lot though, so I expect most are not. In my part of the world, significant ruins are often tagged as archaeological sites (or with both tags), but building=ruins is also used on fragments (which never were a building by themselves), e.g. https://www.openstreetmap.org/way/720588127

1 Like