Hi all,

For those of you which had a previous experience with Idridsi TerrSet (from ClarkLabs), I am considering working with LCM (Land Change Modeler) to model land use change over a rural watershed of 40 sq km in the sahelian climate (west africa).

I am considering three different types of land uses. I have the strong belief that all the land uses are being changed under the influence of the same set of driving variables. Amongst them, I have cultivated areas, which are continiously increasing, based on land use demand, which is closely (I believe) related to climate and population size.

Say I have annual population census (or density) values over the whole simulation period. However, these values does not change spatially over the whole area,. They are rather constant in space, but time varying.

How can I input them in the modelling process, using a Multilayer Perceptron (MLP) ?

Roland.

• Risper Nyairo

Check the tutorial. There is a clear explanation on how to add dynamic variables and how to transform categorical variables into continuous ones

This is an interesting question. I've also been trying to use the population variable as a dynamic input with the intention to replace the original population driver layer with new future ones when their time is reached. Referring to the manual and IF DOABLE, this should be performed by writing a macro script for the population driver. However, I tried writing a Macro script but could not find a function that performs a logical test prior to implementing the new population layer. Instead, from the first stage, the prediction model uses the new population layer (although the time of that layer is still in the future). To transfer my text with numerical values:

- Assume my current LULC map is for 2020

- Assume my population density layer is for 2020

- Assume that I have population predictions for 2030 and 2040.

- Assuming that I want to predict LULC for 2050 while using population as a dynamic variable.

-> this would imply that the written Macro should replace the population density layer 2 times, one when the prediction stage is after year 2030, a second when the prediction stage is after 2040. However, I have not found a function in the Macro that checks the time condition prior to recalculating the population density driver.

Have you found a solution to this issue? How did you eventually include the population density as a dynamic variable within LCM interface?