[R-sig-Geo] Predict gam in a loop on multiple raster stacks and keeping identical layer names

[Robert J. Hijmans]

Kristin, I think you can do something like this:

lst <- list()

for(i in 1:96) {
  Predictors <- stack(stackchl[[i]], stacksst[[i]], stackpar[[i]], lat, lon)
  names(Predictos) <- c('chl', 'sst', 'par', 'lat', 'lon')
  lst[[i]] <- predict(Predictors, gammodel, na.rm=TRUE, type="response")
}

s <- stack(lst)

On Tue, May 5, 2015 at 8:54 AM, Kristin <nessnjor at gmail.com> wrote:
>
> I have a gam model with predictor names (variables) par, chl, sst, lat and
> lon. The model I called "gammodel".
> Then I have predictors for 96 day period in raster stacks: "stackpar",
> "stackchl", "stacksst"
> -that is 3 different raster stacks  - each one with 96 layers.
> lat and lon are fixed in time and are each a separated raster with the
> central value of each cell as a value in the raster.
> All stacks and rasters have same extent, dimention, projection etc.
> dim (84 840  96)
>
> I want predict for each day for each pixel in a loop and create an output as
> a raster stack of predictions for each of the 96 days
> Something like this:
>
> Stack <- stack()
> lenght(Stack) <-96
>
> for(i in 1:96) {
>   Predictors <- stack(stackchl[[i]],stacksst[[i]],stackpar[[i]],lat, lon)
>   Pred <- predict(Predictors,gammodel, na.rm=TRUE, type="response")
>   Stack[i] <- Pred
> }
>
> I´m not sure I´m doing the loop correct? Any suggestions of improvement?
>
> BUT the main problem is that the layers for each predictive varaible have
> names which
> do not fit with the model. When I run it outside the loop for a single
> layer:
>
> Error in eval(expr, envir, enclos) : object 'par' not found
> In addition: Warning message:
> In predict.gam(model, blockvals, ...) :
>   not all required variables have been supplied in  newdata!
>
> I have tried to rename the layers in each raster stack like this:
> names (par)=rep("par",96)
> But the names are changed to: par.1, par.2, par.3, etc..
> So they do not fit the variable name in the gam model itself.
>
> all the best and thanks in advance
>
>
>
>
>
>
>
>
> --
> View this message in context: http://r-sig-geo.2731867.n2.nabble.com/Predict-gam-in-a-loop-on-multiple-raster-stacks-and-keeping-identical-layer-names-tp7588176.html
> Sent from the R-sig-geo mailing list archive at Nabble.com.
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo