I am fine with it. I was asked to be a validator but like you found it was easier to map from scratch and do the occasional clean up afterwards.

A note on rural buildings in Zambia, there is usually a main house surrounded by a number of other structures and the use of these will depend on the area and the tribe but I would say, from my experience, generally there will be an outside cooking structure that will be circular and have a thatched roof, a grain store which will appear similar and may be temporary. While the children are younger they will live in the main house but teenage boys will then move into another separate structure to the other 2 mentioned, possibly bigger. There may be another much smaller bathing structure usually square and not roofed.

In parts of Zambia a family may have a walled compound, reeds or other natural vegetation surrounding all of this, particularly in the west. Most isolated settlements will be a cluster of family dwellings with fields surrounding them, a classic African village, these may have a larger thatched communal structure or, more likely, just a large tree in the centre of the village. Closer to the main line of rail it is small family farms with fields nearby or surrounding the buildings.

Due to the construction methods the buildings may not last and are rebuilt and not necessarily on the same footprint and in cases where the land is totally exhausted or the water is a problem the whole village may move away, this is also the case where it is feared there is witchcraft or other nastiness affecting the area. Even in official descriptions of the boundaries of parks and districts landmarks will be given in the form of Manza Village (1971) ie where Mwanza Village was in 1971. Just some notes.

Happy for you to clear out the HOT detritus

Another thought, I know last year teams went around some areas and those along what we call the 'line of rail, and verified that buildings were dwelling houses, I assume this was in preparation for the census which usually takes place in October. Obviously places like the far west of Zambia, west of the Zambezi, are pretty well inaccessible by road due to the sandy nature and poor maintenance of the roads as well as the sparseness of the population so perhaps that is the idea behind mapping the buildings there

No problems there. The HOT mappers are a bit of a headache but also those that use AI generated building footprints as they will not verify it is a building not a billboard, there are many large ones in Lusaka, also bare concrete rectangles that are parking areas. Unfortunately may of the HOT mappers doing this recently are actually Zambian and I suspect University students UNZA is University of Zambia they tend not to respond to messages and suggestions.

I did look at one or 2 of your deletions in Kabwe and you actually deleted the more accurate outline but I know it is a problem as you tend to assume the bad mapper is the one making all the mistakes.

The census mapping l have looked at in the far west is possibly a futile exercise as unless the imagery is up to date the buildings mapped are a snapshot of the past and villages there may move because of lack of water, it is extremely sandy in that part of the world, unfortunately it has been a region underdeveloped since Independence. Generally the imagery in the urban areas is pretty up to date, within a few months in places, not so in the far west.

Hi Johnwhelan,

Just a note, Maxar imagery tends to be more up to date over Zambia at present. But that is not guaranteed, but you may end up deleting buildings that appear in Maxar and not in Bing, there is also a big difference in offset between the 2 in some places. Imagery is often many years out of date in places.

Hi Chabota Munachoonga,

Before adding 'localities' please check whether there is a town or village already on the OSM map with the same name and if so don't add the locality. As per the Wiki (place=locality?uselang=en-GB), a locality is an unpopulated place so a general area not somewhere like Chisamba or Liteta which are towns. I suspect you are mapping constituencies and your source is the Electoral Commission of Zambia (ECZ). I do not think it is a good thing to map constituencies as they will change and are particularly hard to map the proper boundaries, in any case the tag locality doe not accurately represent the data you are mapping.

From my knowledge Chisamba and Liteta are towns and are on the map and both Miswa and Chikonkomene are railway sidings with residential areas with those names already mapped.

Hi mapwitch,

I know this was mapped many years ago but thought I would point out that the 'weirs', such as node/4445249241, are actually fish traps built with reeds, you will notice there are breaks in nearly all of them, in the break a cone of reeds will be placed to trap fish as they swim down stream. Not sure of the correct way to map them

Hi Mar Mar,

This is an example where what appears as scrub is not actually scrub. The boundary between the scrub and woodland is actually the overlap of 2 different sets of imagery taken at different times of the year and a different clarity. The imagery to the right is taken later in the the dry season, hence black burnt dambos and to the left is soon after the end of the rains so trees are still green and dambos look wetter and so are not burnt

I have mapped others as wetland seasonal=yes but not tagged them as wet_meadow as I don't think that is entirely accurate. They should possibly have their own type tag and perhaps be rendered more as a combination of grassland and wetland. The larger wetlands in Zambia and Central Africa are actually classified as Zambezian Flooded Grasslands, this includes areas such as the Okavango Swamps, these areas tend to be flooded all year round just the extent varying through the season. During the dry season the dambos are more like a dry grassland and in dry years may not even become very wet at all. There are many types of dambo, the ones along rivers being the most common but some are more isolated and may have a temporary pool of water at the low point known as a pan.

Don't be fooled as to how dry it looks in the imagery as most imagery is taken during the dry season April to November, even what appears as scrub may appear as heavily wooded during the wet season. What you have tagged as scrub is not technically scrub, it is more wooded than a typical savanna grassland but not as wooded as a true woodland

Another feature to be aware of is the areas surrounding the streams, these are called 'dambos' locally and are a form of seasonal wetland that are dry grasslands in the dry season that are highly susceptible to wild fires and will have a black appearance, have look tot he east of where you are mapping

Hi Mar Mar,

I keep an eye on edits in Zambia as there are many poor edits and it is easier to pick them up straight away before there are too many. I just happened to check and saw your edit. I have created a temp changeset (Changeset: 123036964) that identifies termite mounds. Large ones will have a clump of bushes on them as per the one in your edit area, smaller ones may have been cleared by the small scale farmers. The collections of buildings are small family farms. You may even come back to map again in a few years time to find the buildings have all moved, particularly the round mud brick huts as they are temporary by nature and will only last a few years. Rectangular structures will be more permanent.

Hi Mar Mar,

way/1074716174 does not enclose grassland but an agricultural field that will be planted when the rains come in November/December. Where you have marked 'TEMP REF' is a large termite mound. Traditional or 'small scale' farmers practice a form of slash and burn agriculture which results in large areas of deforestation

Be careful when using mapwithai as the AI algorithms can not distinguish between large billboards and buildings, way/1061958737 is a billboard. Also Some of the imagery used to analyse for buildings is old so there are errors with a rapidly developing city like Lusaka

Be careful using mapwithai as sometimes it shows open concrete areas as buildings as in this case

Hi there is no bridge here just a crossing over a seasonal stream that will be dry for around 6 or 7 months of the year. There will not even be any structural ford

Please do not use roads as part of multipolygons, while the road may be used as a boundary for protected areas in Zambia the actual boundary is usually gazetted as the road reserve to one side of the road not the actual road itself. If the road is re routed the boundary may not follow the new route

While GMAs are protected areas in Zambia they are not of the same category as a National Park, they are a buffer area around the park and controlled hunting is allowed, professional hunting in particular. GMAs are IUCN Management Category VI i.e protect-class 6 not 2

Hi Max Berckmueller,
While Lusaka is the provincial admin centre of the Lusaka Province and the admin centre of Lusaka District it is also the capital of Zambia and so its admin level should be 2. In Zambia admin level 3 is not used, 4 is provincial admin level and 5 is district admin level, while there are further subdivisions in the form of wards these are not mapped as they are more related to political constituencies therefore Zambia has no admin levels above 5

Hi mapwitch,

I see one problem with a 'containing multipoygon' with the outer ways of all multipolygons within, If it is tagged with a feature, in this case wetland this feature now overlays all other features when the map is rendered, including water bodies and islands. The solution to this is to not tag it with any feature but I suspect this may show as an error when validation tools are used but I am only guessing at this, JOSM validation does not complain, I have tried. It did complain where somebody has named one of the elements as 'Lake Shore', I have removed this. I have also repaired some of the role errors shown by OSM Inspector in some of the 'contained' multipolygons. I have not removed the wetland tag from the containing multipolygon just the name on the one element. I hope you do not mind.

Hi SPECTRA995,

Please don't add culverts these are just tracks and the streams will only run when it rains. I can assure you there are NO culverts

I had also considered the idea of a relation made up of only outer most ways and think this is worth a try.