SDG 15.3.1 Computation for Large Area (>70,000 km²) at 30m Resolution — Feasibility and Workflow Question

[Sinisa Polovina]

Dear TrendsEarth Development Team,

I am currently working on the computation of the SDG 15.3.1 indicator (Land Degradation Neutrality), which covers an area of approximately 70,000 km². I am using national datasets at 30-meter spatial resolution for all three sub-indicators:

- Land Cover Degradation (Corine Land Cover)
- Soil Organic Carbon (SOC) Degradation — custom national dataset
- Land Productivity Dynamics (LPD) Degradation

I attempted to run the SDG 15.3.1 final indicator computation locally within QGIS using the TrendsEarth plugin. However, after approximately 6 hours of processing, QGIS became unresponsive ("Not Responding" status in Task Manager, 0% CPU usage) and the computation did not complete. My hardware configuration is: Intel Core i9-13900HX, 32 GB RAM, Windows 11 64-bit.

I have the following questions:

1. Is it technically feasible to compute the SDG 15.3.1 final indicator locally for an area exceeding 70,000 km² using 30-meter resolution rasters? Is there a known limitation in TrendsEarth for this scale of computation?

2. If local computation at this scale is not feasible, is it possible to divide the study area into smaller regions (e.g., administrative units or sub-regions) and run the computation separately for each region? If so, what is the recommended approach for splitting the area of interest?

3. After computing the SDG 15.3.1 indicator separately for multiple sub-regions, is there a built-in tool or recommended workflow in TrendsEarth to merge the results and generate a single summary report — including all transition matrices and area statistics — for the entire country?

4. Is there a way to automate the merging of sub-regional results (rasters and summary tables) to produce the final national-level SDG 15.3.1 output with complete transition matrices?

Any guidance or recommendations regarding the optimal workflow for national-scale assessments at high spatial resolution would be greatly appreciated.

Thank you very much for your time and support.

Kind regards,

Sinisa

[Tom Kiptenai]

Hi Sinisa,

Thank you for bringing this matter. The best approach when running analysis for a bigger region will be to use GEE, the team can share a script for this. Fir my case I ran analysis for Kenya starting with the national land use land cover maps 30m resolution. I'll try to complete the analysis sdg 15.3.1 and share my experience. 

Regards 

Kemboi Kiptenai Tom
Tel: (+254)725301045
Geographic Information System (GIS) and Remote Sensing Specialist
P.O Box 100-50302
Kapsokwony,
Kenya.
Skype: tom.kiptenai

--
You received this message because you are subscribed to the Google Groups "Trends.Earth Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to trends_earth_us...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/trends_earth_users/91f94a5f-a595-497d-ab90-5a64306e9273n%40googlegroups.com.

[Tom Kiptenai]

Hi Sinisa,

I was able to run the analysis for Kenya, an area of over 580,000 km², I was able to get the output maps and layers without any problem. I utilized 30m national land use land cover maps to generate land cover degradation layers, then used them to compute sdg 15.3.1. 

Regards,

Tom

[Sinisa Polovina]

Hi Tom,

Thanks for the update, that’s great news about the Kenya analysis.

I’ve also completed the calculations on my end. The process took about 30 minutes to run. Regarding the procedure, I noticed a specific technical requirement: when importing 30m national datasets (Land Cover, Soil Organic Carbon, and Land Productivity) via the 'import custom data' tab in TrendsEarth, the software reclassifies categories and handles bit-depth conversions (e.g., from 8-bit to 16-bit).

I found that for the process to work correctly, I had to close the current QGIS workspace and open a fresh, empty one before importing the reclassified rasters. It wouldn't function otherwise.

Also, just a heads-up: it’s necessary to update TrendsEarth and install the latest QGIS version (OSGeo4W-3.44.9-1) for everything to run smoothly.

Kind regards, 

Siniša

You received this message because you are subscribed to a topic in the Google Groups "Trends.Earth Users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/trends_earth_users/NoNiRMmQzTM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to trends_earth_us...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/trends_earth_users/86e9911b-a111-45b3-8c37-33ffbb352e6fn%40googlegroups.com.

[Shabana Nasreen]

Im also working on LDN of Pakistan. Can you please suggest me is it possible to find maps in GEE through overlay weighted analysis of LULC, soc and NDVI??

To view this discussion visit https://groups.google.com/d/msgid/trends_earth_users/CAM9i1T5H%3DtUbrRgsQiDFd2LPMgO8PPdcNp0kfip4tcKerLdpXQ%40mail.gmail.com.

[Tom Kiptenai]

Super! Thank you, Siniša, for sharing this. It will definitely help other users. @Shabana, you can choose available data from whatever source and prepare them following the video on how to use a custom LULC map and perform the analysis in QGIS. Make sure you follow the hints from  Siniša. 

Regards,

Tom

--