I would like to know whether starting from a map (shown below; retrieved from Harper et al. 2007) I can use some kind of algorithm (I read something about "midpoint algorithm") that simulates a different landscape, that is, expanding the forest cover (green part) randomly.
In addition, I was wandering whether is also possible to change the rate of expansion of forest cover, based on environmental data (which suggest moments in the past more favourable for forest expansion).
You can run an iterative simulation yourself.
Split the landscape into cells (pixels work, but you can also pick some points of interest, as long as they are more or less equally distributted).
Then you run a simulation in steps. At each step for every cell or point you check the neighbouring cells and apply some logic (check for the conditions for the forest to expand), and modify the cell data accordingly.
Repeat this a number of time to get the simulation you need.
I'm not a forestry expert but one simple algorithm would be to check how many other cells are forest in the next 2-3 cells, and use that as a ratio for a uniform (0-1) random number to say if it will expand in the next cycle there(if random < ratio it expands). The randomness make the result less stable but you can get a feeling of how reliable the forest is to expand or shrink (by checking results of a few runs).
