ben18785
commented
5 years ago Yes, interesting. I agree with your idea for rerunning the inference.
Perhaps the best way to handle this would be by individual based simulation. Where each seed disperses from its parent, then the tree grows (according to its particular characteristics) and competes for local space once it has reached a given size. The idea would then be to compare individual based simulations with the actual data and fit their parameters.
This may well have been done but could be a fun summer project for a student!
On Fri, Mar 29, 2019 at 5:11 PM Armand1 notifications@github.com wrote:
There is an issue that has been in the back of my mind re the BCI data, namely, whether the trees that we are counting can, in fact, contribute to the next generation.
Currently, we are modeling all the trees, regardless of how big they are. But this is not realistic. A seed falls, grows into a shoot, it gets counted; but it cannot contribute progeny to the next census interval since it's still immature.
One way to deal with this is to only count trees of a certain size. In fact, every tree at every census comes with a "bhd" value, breast-height-diameter. I think that one could exclude all trees with bhd<10 cm say, and that would give us reproductives. In effect, a tree would be invisible until it reaches that size and suddenly appear in the population as reproductive individual. I just raise this now, since it may be that we need to re-shape the data and rerun the models with this constraint. I might also affect the frequencies of species that you start with in any simulations that you do. I will ask James and Chisholm what they think.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ben18785/Selection_simulations/issues/8, or mute the thread https://github.com/notifications/unsubscribe-auth/AESFqKB_noALPVLsSRXN5vSgwRfvTPG9ks5vbklDgaJpZM4cS2oa .
Armand1
There is an issue that has been in the back of my mind re the BCI data, namely, whether the trees that we are counting can, in fact, contribute to the next generation.
Currently, we are modeling all the trees, regardless of how big they are. But this is not realistic. A seed falls, grows into a shoot, it gets counted; but it cannot contribute progeny to the next census interval since it's still immature.
One way to deal with this is to only count trees of a certain size. In fact, every tree at every census comes with a "dbh" value, "diameter at breast height". I think that one could exclude all trees with dbh<10 cm say, and that would give us reproductives. In effect, a tree would be invisible until it reaches that size and suddenly appear in the population as reproductive individual. I just raise this now, since it may be that we need to re-shape the data and rerun the models with this constraint. I might also affect the frequencies of species that you start with in any simulations that you do. I will ask James and Chisholm what they think.