Trouble using akde as shapefile

[Malena Candino]

Hi Chris,

I'm trying to work with the akde contours to be able to filter locations and extract only those that fall within the 95% home range. I have calculated the akde's, and wrote them as a shapefile using for example: writeShapefile(AKDE$G01_Winter_2019, folder, file=NULL,level.UD=0.95,level= FALSE) I set level to false because I don't want the confidence intervals.

When I preview this shapefile in QGIS, it looks exactly what it should look like, and I don't have the confidence intervals. If I read that shapefile back into r as an 'sf' to mask locations I don't want, it has the confidence intervals again for some reason, and it also appears to be distorted. If I read it as a spatial polygon data frame it also has the confidence intervals back, and I can't seem to be able to plot the locations on top of it.

Any ideas as to how to best do this and why I may be encountering these problems would be really useful!

I tried attaching my code, the shapefile and the locations I'm using- in case they helped- but I wouldn't let me post this (I can share by regular email if it's useful)

Thank you,

Malena.

datos <- read.csv("G01_ctmm.csv",header=T)
head(datos)
str(datos)
#---------------#
# create AKDE---#       
#---------------#

datos$id<-datos$Burst
burst<-as.telemetry(dat, projection = CRS("+proj=utm +zone=19 +south +datum=WGS84"))

#plot all guanacos, and ind guanacos
plot(burst)
COL <- rainbow(length(burst))
plot(burst,col=COL)
# look back at G01
COLG01 <- color(burst$G01_Winter_2019,by="time")
plot(burst$G01_Winter_2019,col=COLG01)

SVF <- list()
for(i in 1:length(burst)){
  print(i)
  SVF[[i]] <- variogram(burst[[i]])
}
names(SVF) <- names(burst)
zoom(SVF$G01_Winter_2019)
## fit models to all animals
FIT <- list()
for(i in 1:length(burst)){
  print(i)
  GUESS <- ctmm.guess(burst[[i]],interactive=FALSE)
  FIT[[i]] <- ctmm.select(burst[[i]],GUESS,verbose=TRUE,trace=2)
}
names(FIT) <- names(burst)

AKDE <- list()
KDE <- list()
for(i in 1:length(burst)){
  print(i)
  AKDE[[i]] <- akde(burst[[i]],FIT[[i]][[1]])
  IID <- ctmm.fit(burst[[i]],ctmm(isotropic=TRUE))
  KDE[[i]] <- akde(burst[[i]],IID)
}
names(AKDE) <- names(burst)
names(KDE)<-names(burst)
# AKDE versus KDE area
summary(KDE$G01_Winter_2019)
summary(AKDE$G01_Winter_2019)

plot(AKDE$G01_Winter_2019)
# comparison plot
EXT <- extent(list(burst,AKDE$G01_Winter_2019))
par(mfrow=1:2)
plot(burst$G01_Winter_2019,AKDE$G01_Winter_2019,ext=EXT)
title("AKDE")
plot(burst$G01_Winter_2019, KDE$G01_Winter_2019,ext=EXT)
title("KDE")

par(mfrow=c(1,1))
folder <- paste(getwd(),"/shp-output2", sep="")
plot(AKDE$G01_Winter_2019)
str(AKDE$G01_Winter_2019)
writeShapefile(AKDE$G01_Winter_2019, folder, file=NULL,level.UD=0.95,level= FALSE)

cont<-st_read("G01_Winter_2019.shp")
cont<-cont[2,]# I don't know how else to loose the CI
plot(cont)
winter<-subset(datos, datos$id== "G01_Winter_2019")
coordinates(winter) <- c("long","lat")
proj4string(winter) <- "+proj=longlat +datum=WGS84 +ellps=WGS84 +towgs84=0,0,0"
plot(winter, add=TRUE) # when adding my locations I see the HR polygon appears to be distorted
pol<-shapefile("G01_Winter_2019.shp")
plot(pol)
plot(winter, add=TRUE) #nothing happens

[Christen Fleming]

Hi Malena,

I don't think level=FALSE would turn off the CIs, I think it would just set them to 0%, which would make them infinitesimally narrow. Also, it sounds like you would rather want to export the AKDE to SpatialPolygonsDataFrame (sp) format and then to sf. I haven't gotten around to making a single function call to export directly to sf, but it's pretty easy. Then, with either an sp or sf format object, it should be pretty easy to extract the point-estimate polygon that you want to work with.

It looks like your plotting at the end is not in ctmm. The distortion could be because you are plotting in long-lat, rather than a locally flat projection? On most of the globe, plotting long-lat is distorted unless you do some calculations to figure out what the correct aspect ratio should be.

Best,

Chris

[Christen Fleming]

I coded up sf exporting with the as.sf() function on GitHub. Feel free to provide feedback on how well that works or what it needs.

Best,

Chris