Problems calculating home ranges

Question

I am using GPS data and the ctmm package to calculate home ranges for multiple individuals. The code runs without errors, but does contain warnings and produces incorrect results. A colleague used the same code and the same csv file on another computer without issue. I have gone as far as to copy and paste their code from email to ensure it was correct and tried multiple versions of R studio and have yet to find the issue. I have included a small sample of the data. I have also included the code and warnings after the line of code where they occur.

Id              DateTime           Latitude      Longitude   YearMonth
151_259_45_1    11/14/2007 0:00   27.4922397    -97.6223344   2007-11    
151_259_45_1    11/14/2007 1:30   27.4922361    -97.6224614   2007-11    
151_259_45_1    11/14/2007 2:00   27.4922217    -97.6225903   2007-11   
151_259_45_1    12/4/2007 0:30    27.4851964    -97.6170192   2007-12    
151_259_45_1    12/4/2007 3:00    27.4846739    -97.6277725   2007-12   
151_259_45_1    12/4/2007 3:30    27.4854247    -97.6346442   2007-12    
151_259_45_1    1/3/2008 14:30    27.4840406    -97.6196067   2008-01   
151_259_45_1    1/3/2008 15:00    27.4840003    -97.6195778   2008-01   
151_259_45_1    1/3/2008 15:30    27.4840381    -97.6196153   2008-01

#####Packages
library(sp)
library(rgdal)
library(move)
library(amt)
library(ggmap)
library(dplyr)
library(purrr)
library(tidyr)
library(lubridate)
library(adehabitatLT)
library(adehabitatMA)
library(raster)
library(mapview)
library(sf)
library(FNN)
library(ctmm)

#########Import Data
df<-read.csv("C:collars_sub.csv")

names(df)
unique(df$Id)
unique(df$YearMonth)

df$id2<-paste0(df$Id, df$YearMonth)
unique(df$id2)

df$DateTime2<-as.POSIXct(df$DateTime, format="%m/%d/%Y %H:%M")
df$DateTime2

############AKDE  
deers<-df%>%
  make_track(Longitude,Latitude,DateTime2,id=id2, crs = CRS("+proj=longlat"))

deers<-transform_coords(deers, sp::CRS("+proj=aea +lat_1=29.5 +lat_2=45.5 +lat_0=23 +lon_0=-96 +x_0=0 +y_0=0 +datum=NAD83 +units=m +no_defs +ellps=GRS80 +towgs84=0,0,0"))

Warning messages:
In showSRID(uprojargs, format = "PROJ", multiline = "NO") :
  Discarded datum North_American_Datum_1983 in CRS definition,
 but +towgs84= values preserved

deers<-deers%>%
  nest(data=c(x_, y_, t_))

#########calculate akde
akde<-deers%>%
  mutate(hr_akde = map(data, ~ hr_akde(., fit_ctmm(., "ouf"))))

Warning message:
1: Problem with `mutate()` column `hr_akde`.
i `hr_akde = map(data, ~hr_akde(., fit_ctmm(., "ouf")))`.
i Discarded datum Unknown based on GRS80 ellipsoid in CRS definition,
 but +towgs84= values preserved

########reformat to calculate hr sizes
hr_sizes_2 <- akde %>% 
  pivot_longer(hr_akde, names_to = "estimator", values_to = "hr")

#########calculate hr area
hr_sizes_2 <- hr_sizes_2 %>% 
  mutate(hr_area = map(hr, hr_area)) %>% 
  unnest(cols = hr_area)

###########We're not interested in LCI and UCI right now... get rid of

hr_sizes_3<-hr_sizes_2[hr_sizes_2$what=="estimate",]

#########look at each HR
for (i in 1:8) {
  telem<-as_telemetry(hr_sizes_3$data[[i]])
  plot(telem, UD=hr_sizes_3$hr[[i]]$akde, 
       main=paste(hr_sizes_3$id[[i]],hr_sizes_3$area[[i]]*3.861e-7),  cex=.5)
  compass()
}

user_5329073 · Answer

This expands on my comment.

When I asked about package versions, I said you should see if you have the same versions as the other R-installed computer. The computer where you said this code works as expected. If their versions aren't the most recent that might be the difference.

When I look at the functions you used and the libraries you called, there are a lot of conflicts. However, when I looked closer, I noticed that you called libraries that are not needed.

The libraries needed for this code to work are here, along with those that are not needed commented out:

##### Packages
# library(sp)
# library(rgdal)
# library(move)
library(amt)
# library(ggmap)
library(dplyr)
library(purrr)
library(tidyr)
# library(lubridate)
# library(adehabitatLT)
# library(adehabitatMA)
# library(raster)
# library(mapview)
# library(sf)
# library(FNN)
library(ctmm)

If there is more code, make sure you only add the libraries you actually need. Sometimes this gets a bit convoluted (as you try different packages, for example).

I get a lot of deprecated errors when I run this code. I receive other warnings as well. If I run the code too quickly I get errors that the code didn't work. However, I run the line again and it works.

Since you know how this should look when it's finished, I do wonder if this is the output you expect:

I changed the limits so you could see the plot a bit better

If this is what you expected, then only use the libraries that are listed here, and wait for each function to complete before running the next one. I am guessing that the package amt isn't really stable and is doing a ton of work in the background.

Problems calculating home ranges

1 Answers