scale-dependence of establishment probabilities
Dominic Cyr
Hi everyone,
I've been thinking for a while about the scale-dependence of the species establishment probabilities. The SEP is defined as the probability of a cohort establishing [on a pixel, during one timestep] given a sufficient seed rain and adequate light conditions. Intuitively, that probability should indeed increase with pixel size and timestep length. But are absolute SEP values really that important for simulating stand and landscape dynamics with LANDIS-II? Or is it good enough to make sure that they represent the gradients of relative performances among species on given land types, as well as some kind of fertility gradient among land types.
Often I suspect there are some scale mismatches between the source of information that are used for setting SEP and the spatial scale at which LANDIS-II operate. Can anyone share his-her views about that. I'm pretty sure most of us encountered that problem, but as far as I know that has never been discussed on this forum.
Thanks a lot
Dominic
Caren Dymond
In coniferous forests of British Columbia I've found it is important to have data and expert knowledge to calibrate the natural regeneration rates in LANDIS-II. Having estimates of SEP was not enough because of how the model uses them. We got way too much natural regen under established cohorts. We divided the SEP estimates from TACA by 3. Talking to Craig he said they were probably more like 3yr estimates rather than the annual time step we are using in LANDIS. We also adjusted the light establishment table.
Hope that makes sense!
Caren