Hello,
I am hoping to make future projections of LULC for Riau, Sumatra based on three future scenarios (I am currently thinking of doing them for 2050). To calculate changes in carbon stocks and sequestration future LULC raster maps are required. How can I make these in QGIS based on my own set of scenarios?
Any advice or suggestions would be great! Thank you!
Hi @SamStrangward ,
Here’s a post I found by searching the forums:
Hopefully this helps get you started. I’m sure there are other posts on the forum that address this issue as well.
Cheers,
Doug
@SamStrangward along with @dcdenu4’s pointer to the Scenario Generator, I’ll chime in a little more about creating scenarios. How they are created varies widely, depending on what kind of data you’re starting with, what kind of scenario it is, and what kind of information you’re incorporating into the scenario to change it from baseline conditions. So it’s really hard to provide generic guidance.
If you already have spatial data about where changes occur (in a shapefile or raster, for example), you can probably just use GIS techniques to change the appropriate pixels in the baseline landcover map, then update the biophysical table if needed. If your change information is not yet spatial, you’ll probably need to make it spatial by reclassifying certain LULC types based on how they’re likely to change, or whatever is appropriate for visualizing the elements of your scenario. Perhaps only certain areas will change, like a buffer around towns or rivers, in which case you’ll need to make these buffered features, and use them to change the pixels in the baseline landcover map. There are many possibilities for what’s needed to realize your scenario, thus many possible GIS techniques.
It is also possible that your scenarios only change the biophysical table, not the LULC map. For example, if you want to simply age all of your forest by 20 years, you can provide different carbon pool information without needing to alter the LULC map.
~ Stacie
Hello,
Thank you so much for such detailed advice it’s really helpful!
I am hoping to measure how different scenarios of palm oil expansion (I don’t currently have this scenario information but I am hoping to use different projections from literature to create scenarios) will impact peatland carbon stocks and rates of sequestration.
I will explore these different techniques to see what’s best for my idea. Thank you!
Hello @swolny,
I’m also at the stage of trying to create a future LULC map. I have LULC rasters for 2010, 2015 and 2020 that I could use for change detection. I would like to try to find a GIS technique that would allow me to change the pixel values in the class of my baseline land cover map. Since the last post, have you come across any examples of geoprocessing workflows using ArcGIS Pro or QGIS that could allow me to generate changes in pixel values to simulate future land use for 2030?
So far, I have been looking at various raster tools (Reclassify, Expand, Nibble, Raster Calculator) but I can’t seem to find a logical workflow to perform this at the moment. I was also maybe thinking of converting to polygon to do a few vector manipulations and change the land cover class areas…
Any helpful videos or tutorials are welcome.
Thanks
Hi @jonathan.pierre -
I have not done the process of projecting past land cover change maps into the future, so I don’t have a workflow to recommend. It seems complex, if you want to make future changes based on past trajectories, possibly including elevation, slope, social factors, and whatever other factors go into making a pixel change more or less likely. While I haven’t used it, IDRISI has a Land Change Modeler that I think does something similar. And there may be plug-ins for QGIS that could either work directly or provide inspiration for your own workflow.
This is something I’d definitely like to hear recommendations from others, especially if you know of free/open source methods or tools.
~ Stacie
Thanks @swolny, I’ll have a look at IDRISI. I will definitely share any methods if I end up finding anything working. 
Jonathan