calculating SAC from model residuals

[Areej]

Dear Maxenters

I have been reading all the previous conversations regarding this topic but I'm still in doubt of my understanding. So if someone can revise my understanding and the method I followed below.. that would be of great help, not only for me but for the future users too. 

 The first thing is that the model prediction values comes in ascii format. So I was thinking on how I'm going to substract that from the observed values (i.e cells with value of 1)? 

Since I'm using ArcGIS, my initial trials were to create a layer with only presence cells that are equal to 1. All other cells are left empty not zero (as I read the residuals are only calculated for presence cells only). Then I tried subtracting maxent ascii file (which contains the predicted values) from this layer. The obtained difference layer is then converted to a point/polygon feature format so I can conduct Moran's test (Moran's test accepts features only). 

Are these steps correct?  and if you are not familiar with ArcGIS can you explain your alternative method even if it was manual?

Best

[JEON]

Hi, It's a bit late but hope this mail helps you out.
There was same question on Facebook group "Ecology in R" so I commented there.
Here is link:https://www.facebook.com/groups/ecologyinr/permalink/928961311300098/

My answer was 

"To measure SAC in residuals, first need to measure residuals. <residuals = observation value - model prediction> you may have model prediction already, but I believe we should think about their definition first.

For some papers, they measured residuals as "1(for presence) or 0(for absence) - model prediction(0~1 value)". But I strongly doubt about this way. Model prediction could be interpreted as "habitat suitability" or "relative occurrence probability".

Does presence n absence data could be interpreted as same way? 1(for presence) means their relative occurrence rate or probability? if there is presence point, That xy coordinate have highest habitat suitability? if there is absence point, that coordinate indicating 0% of occurrence probability?

I believe they just simply meaning "presence or absence" not "probability" or "habitat suitability". To solve this issue, I believe we need very clear information about species prevalence first. 

I am also very curious about how to measure "probability" or "habitat suitability" in specific coordinates or research area too. 

I think except for the case of remote sensible species such as tree, it must include very intensive field work to measure "residuals"."

Simply, I wonder if you have prevalence information(Tau) and absence data. 

Without that, it would be hard to measure residual from model prediction whatever in their form(raw, logistic, etc)

You can check Merow, 2013, Ecography. This will be more helpful.

2022년 2월 10일 목요일 오전 6시 56분 54초 UTC+9에 gentl...@hotmail.com님이 작성:

[Areej]

 Thank you JEON for the thoughtful answer. Considering your explanation, this method won't be an option as I don't have true absence data and its difficult to obtain since I'm modelling a seabird. 

Thank you JEON

Areej