Variance partitioning

[esgarbanzo]

Bruce et al.,
I'd like to make a summary figure that illustrates the tabular results of my many NPMR models. I was thinking of creating a bar graph with stacked variances illustrating the changing roles of predictors for each response variable and spatial scale. Is there a way to get a variance explained value for individual predictors in 2- or 3-predictor models? I suppose I could use sensitivity values instead, but variance seems more straightforward.

Thanks,
Jeremy