[Vegan-commits] r1348 - in pkg/vegan: inst man

[noreply at r-forge.r-project.org]

Author: jarioksa
Date: 2010-11-04 10:35:16 +0100 (Thu, 04 Nov 2010)
New Revision: 1348

Modified:
   pkg/vegan/inst/ChangeLog
   pkg/vegan/man/SSarrhenius.Rd
Log:
finish commit of SS-species-area models

Modified: pkg/vegan/inst/ChangeLog
===================================================================
--- pkg/vegan/inst/ChangeLog	2010-11-03 16:37:03 UTC (rev 1347)
+++ pkg/vegan/inst/ChangeLog	2010-11-04 09:35:16 UTC (rev 1348)
@@ -4,14 +4,21 @@
 
 Version 1.18-15 (opened November 1, 2010)
 
+	* Self-starting nls() models for species-area relationship: added
+	Gitay and Lomolino models from Dengler 2009 (J Biogeog 36,
+	728-744). Some more should be added, but these were among the best
+	in Dengler's simulation. The starting values of the Lomolino model
+	(SSlomolino) need work, as do several other candidate models. The
+	purpose is to use these also in fitspecaccum(). 
+
 	* rda: the scaling 'const' can now be a vector of two items: the
 	first is used for species and the second for sites. This allows
 	compatibility with Canoco 3.x (but Canoco 4 changed scaling): if
 	'mod' is an rda() result, and 'nr' is the number of rows in data,
 	then "scaling = -2, const = c(sqrt(nr-1), sqrt(nr))" in scores(),
 	summary() or plot() will use default scaling of Canoco 3.x. Help
-	page now points to decision decision vignette explaining scaling
-	and 'const'.
+	page now points to decision vignette explaining scaling and
+	'const' (and the vignette needs updating).
 
 	* prc: coefficients folded in a wrong way. Diagnosed, reported and
 	fixed by Cajo ter Braak. CtB also fixed the scaling of species

Modified: pkg/vegan/man/SSarrhenius.Rd
===================================================================
--- pkg/vegan/man/SSarrhenius.Rd	2010-11-03 16:37:03 UTC (rev 1347)
+++ pkg/vegan/man/SSarrhenius.Rd	2010-11-04 09:35:16 UTC (rev 1348)
@@ -32,7 +32,9 @@
 \details{
   All these functions are assumed to be used for species richness
   (number of species) as the independent variable, and area or sample
-  size as the independent variable.
+  size as the independent variable. Basically, these define least
+  squares models of untransformed data, and will differ from models
+  for transformed species richness or models with non-Gaussian error.
 
   The Arrhenius model (\code{SSarrhenius}) is the expression
   \code{k*area^z}. This is the most classical model that can be found in
@@ -100,7 +102,32 @@
 ## draw a line
 xtmp <- seq(min(sipoo.area), max(sipoo.area), len=51)
 lines(xtmp, predict(marr, newdata=data.frame(sipoo.area = xtmp)), lwd=2)
-## The normal way is to use linear regression on double log data.
+## The normal way is to use linear regression on log-log data,
+## but this will be different from the previous:
+mloglog <- lm(log(S) ~ log(sipoo.area))
+mloglog
+lines(xtmp, exp(predict(mloglog, newdata=data.frame(sipoo.area=xtmp))),
+   lty=2)
+## Gitay: quadratic of log-linear
+mgit <- nls(S ~ SSgitay(sipoo.area, k, slope))
+lines(xtmp, predict(mgit, newdata=data.frame(sipoo.area=xtmp)), 
+  lwd=2, col = 2)
+## Lomolino: using original names of the parameters (Lomolino 2000):
+mlom <- nls(S ~ SSlomolino(sipoo.area, Smax, A50, Hill))
+mlom
+lines(xtmp, predict(mlom, newdata=data.frame(sipoo.area=xtmp)), 
+  lwd=2, col = 3)
+## One canned model of standard R:
+mmic <- nls(S ~ SSmicmen(sipoo.area, slope, Asym))
+lines(xtmp, predict(mmic, newdata = data.frame(sipoo.area=xtmp)),
+  lwd =2, col = 4)
+legend("bottomright", c("Arrhenius", "log-log linear", "Gitay", 
+  "Lomolino", "Michaelis-Menten"), col=c(1,1,2,3,4), lwd=c(2,1,2,2,2), 
+   lty=c(1,2,1,1,1))
+## compare models (AIC)
+allmods <- list(Arrhenius = marr, Gitay = mgit, Lomolino = mlom, 
+   MicMen= mmic)
+sapply(allmods, AIC)
 }
 \keyword{ models }