Geoffrey Legault
@legault.bsky.social
Data Scientist at Getty Images. Ph.D. in Ecology from CU Boulder.
Google Scholar Profile: https://scholar.google.com/citations?user=1Nh8dD0AAAAJ&hl=en
Google Scholar Profile: https://scholar.google.com/citations?user=1Nh8dD0AAAAJ&hl=en
> CEGS fit to the Barro Colorado Island tree inventory is near-perfect, so nothing more needs to be assumed
...for Barro Colorado Island trees, maybe. How does it perform for other communities?
Presumably you are not arguing that we only need 2 parameters to usefully describe any kind of assembly.
...for Barro Colorado Island trees, maybe. How does it perform for other communities?
Presumably you are not arguing that we only need 2 parameters to usefully describe any kind of assembly.
July 26, 2025 at 2:13 AM
> CEGS fit to the Barro Colorado Island tree inventory is near-perfect, so nothing more needs to be assumed
...for Barro Colorado Island trees, maybe. How does it perform for other communities?
Presumably you are not arguing that we only need 2 parameters to usefully describe any kind of assembly.
...for Barro Colorado Island trees, maybe. How does it perform for other communities?
Presumably you are not arguing that we only need 2 parameters to usefully describe any kind of assembly.
Thanks for the interesting paper!
It seems like one of your arguments is that the parameters of abundance distributions are more useful descriptions of community assembly than H/D/J.
OK. ✅
But are ~2 parameters enough to capture complex assembly processes? Why not fit population models instead?
It seems like one of your arguments is that the parameters of abundance distributions are more useful descriptions of community assembly than H/D/J.
OK. ✅
But are ~2 parameters enough to capture complex assembly processes? Why not fit population models instead?
July 26, 2025 at 1:48 AM
Thanks for the interesting paper!
It seems like one of your arguments is that the parameters of abundance distributions are more useful descriptions of community assembly than H/D/J.
OK. ✅
But are ~2 parameters enough to capture complex assembly processes? Why not fit population models instead?
It seems like one of your arguments is that the parameters of abundance distributions are more useful descriptions of community assembly than H/D/J.
OK. ✅
But are ~2 parameters enough to capture complex assembly processes? Why not fit population models instead?
Would vine copulas be relevant here? I might be misunderstanding the question
December 12, 2024 at 3:22 AM
Would vine copulas be relevant here? I might be misunderstanding the question
What about running some simulations? Start with known h^2, generate a range of sample sizes, then see how well your estimation approach returns the known h^2 for different sample sizes.
November 27, 2024 at 3:29 PM
What about running some simulations? Start with known h^2, generate a range of sample sizes, then see how well your estimation approach returns the known h^2 for different sample sizes.
Like the authors, I think these tools are becoming increasingly important to the field as models become more realistic, and therefore more complex. With recent advances in simulation-based inference, we might even be able to use them to parameterize previously intractable models.
November 22, 2024 at 3:20 PM
Like the authors, I think these tools are becoming increasingly important to the field as models become more realistic, and therefore more complex. With recent advances in simulation-based inference, we might even be able to use them to parameterize previously intractable models.