Negative values at x-axis response curves

[Iwan van Veen]

Hi maxent users,

I have a question regarding the values on the x-axis of the response
curves. I'm using maxent for rhino habitat modelling and have
difficulties interpreting the response curve results in the way that I
don't understand why the x-axis of the graphs of the response curves
for the separate variables start at a negative value instead of
starting at zero. For example looking at distance to water, which I
use as one of the important variables to explain rhino occurrence, it
seems the x-axis of the response curve starts at -1500 meter. The
graph then show a straight line from this negative value on the x-axis
to the zero point on the x-axis (so the y-axis value stays constant)
and then from the zero point onwards show the expected values.

So my question is: why does the response curves show negative values
on their x-axis while in my original data I only use positive values?

Hopefully anyone is able to help me out, Would be really appreciated.

Regards,
Iwan

[Anna]

Dear Iwan and other group members
Did you ever work this one out?  I too am confused - at the moment the response curves are, just as you say, going beyond 0 on the x-axis.  Some even do this above 0.5 on the y-axis, suggesting, in the case of depth for example, I might find my marine organisms on land!?  Not sure what to make of that. 
Any suggestions much appreciated
Thanks
Anna

[Alex T]

Dear Anna,

It is my turn to ask, did you [or anyone else] ever figure this out? Why is it showing negative values on the response curves when there should be no negative values at all!?! I get the same thing with summer-time temperatures in Haiti, certainly not -150 C.

with regards,

- Alex

[Michelle Moeller]

Hi MaxEnt users,

I also do not get why it starts at negative values. Also which response curves are best to use for comparison of models, the one which shows the response of each factor individually or of all the factors together.

All the best,

Michelle

[John Baumgartner]

The curves are just showing you how the marginal predicted probability of presence varies with different values of predictors. If there was a site with temperature -150 C, and all other predictors were at their sample means, then the curve indicates what the predicted probability of presence would be. This is extrapolation beyond your data, and the default behaviour is for Maxent to "clamp" predictions to the values predicted for the lowest and highest sample values for the predictor - this is why you see "a straight line from this negative value on the x-axis to the zero point on the x-axis (so the y-axis value stays constant)". 

If you want to customise the plots, go to Settings > Advanced and tick "Write plot data". You can then read the data into other software packages for plotting.