The idea makes sense and this is where process models can be especially powerful in helping to identify the potential combinations that will be most effective. Models should be a numerical representation of the same conceptual (or perceptual) model you use to select your combinations of factors. Even if models are not perfect (all are wrong,some are useful), the process of integrating a modeling step into experimental design like this would benefit experiments by exploring a wider factor and parameter space than is physically possible and lead to a better perceptual model that leverages both. Ultimately that feeds back to a better numerical model also and the benefits that would bring.
A globally distributed and coordinated set of experiments to test this approach would be fantastic - would you want to run the same set of interventions globally, maybe the principles are the same but the parameters vary according to species traits, system context etc.
I don't think the same interventions would be applicable globally, but the principle could certainly be applied. I lack the vision how models could be integrated in this, but it might be a great topic on which to collaborate...
I’d be happy to chat. I’m not an expert in ecosystem modeling by any means but have a reasonable sense for how the parts could come together, e.g. using a trait based approach. I found this piece very valuable in the way it expresses how modelers and experimentalists can come together to improve our perceptual models. It’s hydrology focused by totally generalizable. https://wires.onlinelibrary.wiley.com/doi/full/10.1002/wat2.1550
I like the idea. To me, it'd make for a great global coordinated distributed experiment. Everyone runs the same set of interventions on their own set of species locally, replicated 100s of times around the world. Just needs someone with a lot of time to organise it!
The idea makes sense and this is where process models can be especially powerful in helping to identify the potential combinations that will be most effective. Models should be a numerical representation of the same conceptual (or perceptual) model you use to select your combinations of factors. Even if models are not perfect (all are wrong,some are useful), the process of integrating a modeling step into experimental design like this would benefit experiments by exploring a wider factor and parameter space than is physically possible and lead to a better perceptual model that leverages both. Ultimately that feeds back to a better numerical model also and the benefits that would bring.
A globally distributed and coordinated set of experiments to test this approach would be fantastic - would you want to run the same set of interventions globally, maybe the principles are the same but the parameters vary according to species traits, system context etc.
Thanks for writing.
I don't think the same interventions would be applicable globally, but the principle could certainly be applied. I lack the vision how models could be integrated in this, but it might be a great topic on which to collaborate...
I’d be happy to chat. I’m not an expert in ecosystem modeling by any means but have a reasonable sense for how the parts could come together, e.g. using a trait based approach. I found this piece very valuable in the way it expresses how modelers and experimentalists can come together to improve our perceptual models. It’s hydrology focused by totally generalizable. https://wires.onlinelibrary.wiley.com/doi/full/10.1002/wat2.1550
Thanks for sharing this paper. I like the term 'perceptual model' in this paper. That's totally my thing. :D
I like the idea. To me, it'd make for a great global coordinated distributed experiment. Everyone runs the same set of interventions on their own set of species locally, replicated 100s of times around the world. Just needs someone with a lot of time to organise it!
Thanks! Oh yes, this would be great....