lizzieinvancouver
commented
2 years ago @DeirdreLoughnan That's awesome! I would actually do couple simpler things before adding on ....
- [ ] Run the model on east and west separately and make sure the answers make sense
- [ ] Decompose the models in plot (e.g., in OSPREE force_bdecompSeedMass_log10_stanfit_37spp_wp.RDS.pdf) ... this is a good model check (and hopefully has also been done for the updated OSPREE traitors models?)
- [ ] Other model checks you can think of?
@lizzieinvancouver I have been working one day a week on my traits chapter and running our trait data using the joint traitors model. I did include site instead of study in the trait model, but otherwise did not change the model.
Most models run well (and quickly!), but the CN and SSD model have low ESS that I can try to fix.
My initial findings are...mixed. I made a plot similar to the cue slope by trait plot we made for traitors and it can be found here. I was disappointed to see there were no relationships with LMA, but not surprised SSD was not interesting. The maroon crosses are estimates of the eastern species and teal crosses the western species.
In terms of next steps. I am curious how hard it would be to add interactions to the phenology model and try adding a second trait. Happy to discuss this if you have time!
Thanks!