maps with different projection ?

Smallpox vs. largepox, gen2 vs gen5, early war vs. peaceful alliances. Which is your favourite gaming style?
Post Reply
Molo_Parko
Hardened
Posts: 158
Joined: Fri Jul 02, 2021 4:00 pm

Re: maps with different projection ?

Post by Molo_Parko »

^ I don't know. I haven't found anything, but that doesn't mean that it doesn't exist. I did find documentation from a 2003 GEBCO data set release, but no explanatory details about data values. In the current data set from 2021, a separate file contains "TID" data which lists the source of the data used for each pixel/cell/area so perhaps looking into the source(s) for the anomalous pixel data might be helpful, but there is a category of "Source unknown" which is only described as "Depth data from unknown sources" which is mind-boggling odd, in my opinion. How do they know that it is valid depth data if they don't know where the values originated???

Maybe it would be better to use the current 2021 data set and see if the problem still exists in that newer data?

https://www.gebco.net/data_and_products ... etry_data/

There is some mention here: https://www.gebco.net/data_and_products ... historical about errors in previous versions of data where land was misclassified as water and vice-versa, but I don't know which areas are wrong in the data you have so not sure if the list explains what you are seeing.
User avatar
meynaf
Hardened
Posts: 155
Joined: Sun Jan 21, 2018 10:27 am
Location: Lyon / France
Contact:

Re: maps with different projection ?

Post by meynaf »

Molo_Parko wrote: How do they know that it is valid depth data if they don't know where the values originated???
Perhaps they know something was invalid when receiving error reports ?

Molo_Parko wrote: Maybe it would be better to use the current 2021 data set and see if the problem still exists in that newer data?

https://www.gebco.net/data_and_products ... etry_data/
Possibly. But this data does not look like what i used. Ice surface elevation, sub-ice topo, tid grid, are nothing looking like ordinary oceanic depth.

Molo_Parko wrote: There is some mention here: https://www.gebco.net/data_and_products ... historical about errors in previous versions of data where land was misclassified as water and vice-versa, but I don't know which areas are wrong in the data you have so not sure if the list explains what you are seeing.
It apparently doesn't explain my conflicting areas. There is one located west to the Nile delta, for example. On the 4320x2160 image it can be found at coords 2490,720.
What could this be ? Swamp ? Lake ? Area with altitude lower than sea level ?
Molo_Parko
Hardened
Posts: 158
Joined: Fri Jul 02, 2021 4:00 pm

Re: maps with different projection ?

Post by Molo_Parko »

Image
^ That area is "Flooded Vegetation" such as a wetland or swamp.

Image
^ Current GEBCO map from http://download.gebco.net shows the area in a way that makes it appear to be water (but is actually only indicating depth.)

That area is called the Qattara Depression.
The Qattara Depression lies below sea level, and its bottom is covered with salt pans, sand dunes, and salt marshes.
https://en.wikipedia.org/wiki/Qattara_Depression

^ The pictures of the area at the Wikipedia page definitely don't look like water though, so the GEBCO map must be indicating it below sea level and using a color that suggests water despite that it doesn't appear to be a notable feature of the area. It doesn't look at all like "flooded vegetation" either. It looks like a desert. :roll:
User avatar
meynaf
Hardened
Posts: 155
Joined: Sun Jan 21, 2018 10:27 am
Location: Lyon / France
Contact:

Re: maps with different projection ?

Post by meynaf »

That appears to settle it. It is land below sea level and should be considered as ordinary land (for our gaming purposes).
If depth is ignored (land always dominates) then this area indeed becomes desert. As altitude isn't handled in civ (at least not for desert areas), this land feature will disappear from the map.

Now that land cover and bathymetry have been handled, i need to mix land cover and land elevation, to make mountains and hills.
Molo_Parko
Hardened
Posts: 158
Joined: Fri Jul 02, 2021 4:00 pm

Re: maps with different projection ?

Post by Molo_Parko »

^ Yes, it certainly does appear to be desert, but somehow it appears to have been misclassified in the "Land Cover of Africa" image as "flooded vegetation". My confidence in the accuracy of land cover data just dropped significantly. How many other areas are grossly misclassified? Is it solely a problem of the source data (ESA CCI 2015) for the "Land Cover of Africa" image, or is it the classification methodology which may affect any or all data sets??
User avatar
meynaf
Hardened
Posts: 155
Joined: Sun Jan 21, 2018 10:27 am
Location: Lyon / France
Contact:

Re: maps with different projection ?

Post by meynaf »

Molo_Parko wrote:^ Yes, it certainly does appear to be desert, but somehow it appears to have been misclassified in the "Land Cover of Africa" image as "flooded vegetation". My confidence in the accuracy of land cover data just dropped significantly. How many other areas are grossly misclassified? Is it solely a problem of the source data (ESA CCI 2015) for the "Land Cover of Africa" image, or is it the classification methodology which may affect any or all data sets??
It seems that these maps are made by something too indirect to be 100% trustable.
I can't tell how accurate it might be. Besides, the data is also subject to interpretation...
Perhaps land cover is just too coarse for the use we need it for.

Currently i have :
- land cover
- ocean depth (aka bathymetry)
- land height
- temperatures
The first has limited value set, the three others have more or less linear value for single data.
I also have a "vegetation" data set. Not converted, not used, but i could if i knew what to do with it.
Molo_Parko
Hardened
Posts: 158
Joined: Fri Jul 02, 2021 4:00 pm

Re: maps with different projection ?

Post by Molo_Parko »

It seems that these maps are made by something too indirect to be 100% trustable.
It varies from one data set / map to the next. Perhaps the misclassification for that location occurred due to the depth of the depression causing shadow, with high levels of mineral salts causing reflection similar to water, resulting in light-spectrum assessments of the land cover being "tricked".

The map/data from year 2000, at this link, shows the area as "sandy desert" with "salt hardpans" at the northwest of the area, which seems accurate.

This one also gets the area right -- desert, but does not specify what source data was used.

At least it isn't thoroughly ludicrous like describing what is obviously desert as "flooded vegetation."

Image
Or as "lake" as it is described by Google's map. Search "Qattara Depression", satellite view, zoom in as much as you like and there is nothing there but desert.
Perhaps land cover is just too coarse for the use we need it for.
What's the alternative??? Some of the newer data sets have 9 meter resolution per pixel which -ought- to improve accuracy, but the data sets are also HUGE because of the level of detail.
User avatar
meynaf
Hardened
Posts: 155
Joined: Sun Jan 21, 2018 10:27 am
Location: Lyon / France
Contact:

Re: maps with different projection ?

Post by meynaf »

Molo_Parko wrote: What's the alternative??? Some of the newer data sets have 9 meter resolution per pixel which -ought- to improve accuracy, but the data sets are also HUGE because of the level of detail.
If i had valid alternatives, i would be already using them ;)
I don't think accuracy is gonna be better with increased resolution. If a data set can't differentiate a forest from a grassland, even one meter per point won't help...

All this makes me wonder how existing maps of Earth for fciv (and other games) have been made...
Molo_Parko
Hardened
Posts: 158
Joined: Fri Jul 02, 2021 4:00 pm

Re: maps with different projection ?

Post by Molo_Parko »

^ The existing Earth maps available for Freeciv are small enough (in tile dimensions) that "small" errors in land cover such as the Qattara Depression would disappear anyway.
User avatar
meynaf
Hardened
Posts: 155
Joined: Sun Jan 21, 2018 10:27 am
Location: Lyon / France
Contact:

Re: maps with different projection ?

Post by meynaf »

Sure, but there are potentially much larger errors than this one.
Post Reply