Changing threat weights, impact on results

Software Support
1

Dear Dave,

I hope this message finds you well. I modelled some ecosystem services for my university research through the InVEST model and I am currently performing sensitivity analysis for the Habitat Quality model. As part of this process, I adjusted the threat weight by an increase of 10% and 30% to observe the effects on the model outputs. However, I noticed that the resulting maps were quite similar to the base map.

Could you please advise on potential reasons for this outcome? I would appreciate your insights and suggestions regarding this matter. Thank you for your time and attention.

Best Regards,

Mitra

2

Hello @Mitra.Gh -

Are you adjusting all of the threat weights by 10% at the same time? Or adjusting the different threat weights individually (for example, adjusting just one threat at a time by 10%)? If you’re adjusting them all at the same time by the same amount, I’m not surprised that the maps look similar. Since the model is all relative, if all weights change by the same amount, the relative rankings will remain similar, and likely even be the same, as the base map.

~ Stacie

3

Dear Natural Capital Project Team,

Hope this note finds you well. To validate the Habitat Quality model, I performed a sensitivity analysis. This is the result:

“To validate the habitat quality model, we performed a sensitivity analysis to identify how the model responded to changes in input data. We increased the weight assigned to highways, a primary threat to habitat quality, by 10 and 30 percent to evaluate the model’s response. The resulting changes in habitat quality were proportional to the weight increase, with a 10 percent weight increase leading to a 0.1 percent decrease in habitat quality, while a 30 percent weight increase resulted in a 3.3 percent decrease within the study area. These relatively small fluctuations in model output indicate that the model is robust and stable, as it is not overly sensitive to variations in input data. The model output followed a similar trend with minimal changes when we altered the weights of other threats, further indicating its robustness.”

Is this validation justifiable? Can I use this in my research? As I am not confident with these findings, I would appreciate your guidance on it which would be really helpful.
Best Regards,
Mitra