Comments on: Combining process-based and spatially explicit statistical modeling

By: Anthony

Anthony — Mon, 20 Oct 2008 16:47:25 +0000

Yeah, that’s correct–well put in fact. I think a big challenge will be to run these models in a spatially explicit way, but if that can be accomplished, then the outputs would indeed be “tied together” with the statistical models. The statistical techniques are designed to tie together predictors and responses using rigorous validation procedures to prevent overfitting.

By: Brian

Brian — Tue, 14 Oct 2008 01:03:55 +0000

So run multiple process-based models for different “areas” and tie all of them together with a statical model?

By: Anthony

Anthony — Thu, 09 Oct 2008 22:12:08 +0000

Brian–you may misunderstand. The process-based models are being incorporated into other modeling efforts, not ignored or destroyed in any way. If this works, it would be a step forward I think. It was tough to explain, so maybe I didn’t make it as clear as possible. thanks for your comments as always! -a

By: Brian

Brian — Thu, 09 Oct 2008 08:14:10 +0000

“Hopefully my nitrogen obsession didn’t bore the participants too horribly.” *yawn* j/k

“…we can use process-based models (like CENTURY) to create spatially explicit layers that can be used in statistical modeling.” Chris is brilliant, but I love process-based models :p

“…the danger that too many cooks are spoiling the broth!” Yup, I’ve been there a few times! Not fun…