Select to view content in your preferred language

Analysing the surface area coverage of a multi layer polygon map?

675
7
04-08-2024 05:27 AM
clynden
Emerging Contributor

Hello all,

I have an interesting project to generate insights about land use, and I haven't seen any similar examples. I started this project in Tableau but it doesn't seem to have the features required for this analysis, so I am moving to ArcGIS. I'm hoping this group can help.

 Example map.jpeg

I need to do some spatial analysis of the land/surface area coverage of Australia, my home country. Australia has a land area of 7.7 million square km, and I've created a map with two layers (see above). Each layer shows a different set of restrictions on land use. But the two layers (red, blue) do have overlaps - i.e. some land has both sets of restrictions. I need to analyse the area (square km) of the map to understand:

  • how much area is covered only by restriction A - depicted in red
  • how much land area is covered only by restriction B - depicted in blue
  • how much land area is covered by both restriction A and B
  • how much land area is covered by neither A or B

That is a minimum, ideally I also need to be able to also generate those numbers for each of Australia's six states and two mainland territories. Ideally, I'd also like to add in subsequent layers as well and repeat the analysis.

The source data is two separate shape files, and each polygon within the files has area data associated with it.

Thanks for your help in getting started with ArcGIS!

  

AJ

0 Kudos
7 Replies
BobBooth1
Esri Contributor

Hi AJ,

Try the Union tool:

https://pro.arcgis.com/en/pro-app/3.1/tool-reference/analysis/union.htm

This will combine the A and B layers. You can then Select by attributes to get areas that are A, or B, or A and B, or neither.

https://pro.arcgis.com/en/pro-app/3.1/help/mapping/navigation/select-features-using-attributes.htm

To summarize by states and territories, obtain a feature class or shapefile of those and union them to the A+B result. Then select by Attribute by state name, and when you do the other selections use the Selection type to choose to select from within the currently selected features.

You might also look at using the Summary Statistics tool to summarize the data, grouping by state/territory.

https://pro.arcgis.com/en/pro-app/3.1/tool-reference/analysis/summary-statistics.htm 

clynden
Emerging Contributor

Hi Bob,

Thanks for your advice, this is really helpful. I think ArcGIS is the right tool, but I'm enountering some obstacles along the way.

So far I've imported the two layers A and B, and clipped A to exclude polygons outside Australia's land boundaries.

clynden_1-1713484164279.png

When I perform the Union however, the Union can't complete as I get an Invalid Tolpology error.

clynden_0-1713484126527.png

I looked into the error code, but the XY resolution and tolerance values are the same for both layers A and B, although different to the base map.

Do you have any advice on getting the union to work?

Thanks again

 

Alwyn

 

0 Kudos
BobBooth1
Esri Contributor

I'd suggest running Repair Geometry on each layer first.

https://pro.arcgis.com/en/pro-app/latest/tool-reference/data-management/repair-geometry.htm

You might also create a new file geodatabase and import each of the feature layers into it before running the Union tool.

clynden
Emerging Contributor

Thanks Bob,

Your advice has been really helpful, although I'm now stuck again. The steps I've taken so far are:

- repaired geometries

- clipped layer A to remove coastal areas

- union of layer A and B

- union with layer C (states and territories)

By this point, I now have one combined shapefile that I want to analyse to understand the area (as per OP). However

- if I use summary statistics, or if I export the Attribute Table to excel, the area seems to be in degrees or something strange that I can't easily convery into square kilometers

- I looked at using the Calculate Geometry Attributes tool to add an Area column listing square kilometers for each polygon, but I get a "too many values to unpack" error.

- As an alternate approach, I then used the "Project" tool to change the combined shapefile to WGS 1984 UTM Zone 55S. This gave me an export with area in square metres which I converted to km^2 in excel - but the total combined area is ~8.6m km^2, which is greater than the land area of Australia at 7.886m km^2

Very frustrating given I thought this might be fairly straight forward....

Does anyone have some advice on what step I've missed, or how I can easily set up ArcGIS or the analysis so I can view things in square kilometers, and ensure the total land area is correct?

Thanks!

0 Kudos
BobBooth1
Esri Contributor

Your original data layers were probably in geographic coordinates, rather than projected.

You might try projecting each of the original layers to get data with planar coordinates first, then doing the overlay, selection, summary statistics.

You chose a projected coordinate system, UTM Zone 55s, but that is probably not the best one for working at the scale of Australia. UTM zones are optimized for the area within them, where distortion is minimal. However, Australia spans multiple (8, I think) UTM zones, so much of your data is being distorted.

This might be a useful resource:

https://learn.arcgis.com/en/projects/choose-the-right-projection/

I'd suggest an area-preserving projection (like Albers Equal Area) designed for the scale and location of your analysis (Australia, continental scale). 

EPSG 3577 might work. https://gis.stackexchange.com/questions/2909/equal-area-srid-for-australia

Australia_Albers.png

 

clynden
Emerging Contributor

Thanks Bob,

I started again from the original raw data and followed your advice, and I think I'm getting a lot closer. However, I still see a larger total land area than Australia has, but the error is coming entirely from Western Australia and Queensland.

This leads me to think that area is somehow being duplicated. Most of the overlap I expect to see between Layers A and B are in those states. Although I thought the union tool would not produce such an outcome? Or is Union the wrong tool for the job?

Interestingly - Pairwise Clip no longer works (to remove the maritime areas from Layer B), but Clip does.

For completeness, the data sets I am using are:

- Layer "A"

  - Collaborative Australian Protected Areas Database (CAPAD) 2022 - Terrestrial - shapefile

  - https://fed.dcceew.gov.au/datasets/ec356a872d8048459fe78fc80213dc70_0/explore?location=-30.215639%2C...

- Layer "B"

  - Native Title Determinations - shapefile ("NTD_Register_Nat_shp.zip ")

  - http://www.nntt.gov.au/assistance/Geospatial/Pages/DataDownload.aspx

- Outline of States and Territories

  - States and Territories - 2021 - Shapefile

  - https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3...

- Outline of Australia (used to clip and remove the maritime areas from Layer B

  - Australia - 2021 - Shapefile

  - https://www.abs.gov.au/statistics/standards/australian-statistical-geography-standard-asgs-edition-3...

Thanks for your help!

0 Kudos
BobBooth1
Esri Contributor

Union should not be producing overlapping areas. You mentioned you "see a larger total land area".  Where? In the Union result attribute table? In your summary statistics? Maybe you can show screenshots of what you're seeing.

 

 

0 Kudos