I am trying to model the abundance and spatial distribution of a tree species using remote sensing but I don't know if this is feasible and what the limitations of this approach are.
Is there anyone who can suggest some ideas for me to think about?
Yes it possible. ..One can analyse the image of satellite using the ground control point with a suitable suitable software and classification techniques avaliable. ....
Yes you can! From remote sensing images first develop predictor variables based on ecology of your target species. Find relation between abundance and your predictable variables and make predictions!
Are looking for abundances of a certain species (=ratio of certain species in a pixel) or a count of plants?
In the case of counting, you will have to ensure that your satellite resolution is enough to get a certain number of pixels per plants. If the resolution is to low you can then estimate the presence (and percentage) of a plant species by classification/unmixing approaches. Finally, a last possibility is to estimate the density of plants in the low resolution imagery which would provide you with an estimate of count over a given area.