[R-sig-Geo] spatstat: test plant growth patterns against environmental covariate? - kstest

[Erika Mudrak]

I have a point pattern data set (SPECIES.ppp) of stem locations for a species, and I want to see if they depend on several Environmental variables such as light and soil properties (working one at a time, for now, generally named Envir, of class im, which was obtained using IDW of sample locations evenly distributed throughout the study region).  I realize that since I'm working on the assumption that my variables are continuously varying, I should use KS-tests rather than X-squared quadrat-based tests. 

I would like to use the KS test as outlined in the spatstat notes, but I am confused about the difference in null models for the below tests:

1)
KS=kstest(X=SPECIES.ppp,covariate=Envir)	#Using ppp and covariate
2)
fit=ppm(SPECIES.ppp,~E  , covariates=list(E=Envir))
KSfit=kstest(model=fit,covariate=Envir)         # Using ppm and covariate

I get slightly different results (though both p-values are very low).

Thank you!
Erika

-- 
-------------------------------------------
Erika Mudrak
Graduate Student
Department of Botany
University of Wisconsin-Madison
430 Lincoln Dr
Madison WI, 53706
608-265-2191
mudrak at wisc.edu
http://botany.wisc.edu/waller/lab/Mudrak/