Counting telemetry intersections with a polygon in R

Question

I am currently tracking animals in national parks and have lat/long and date/time data for individuals and a polygon shapefile for the park boundary. I want to know how many times individuals leave the park boundary each month and at what time of day. In nearly every instance, the individuals return to the park, so I need to account for directionality or count every other intersection.

At the moment, I am converting the raw data into telemetry paths as an ltraj object with the 'adehabitatlt' package. However, I cannot seem to trim the ltraj objects to the polygon, nor does it retain the associated time data for individual segments.

In the below example, I would want to know how many times and when the ibex leave the sps polygon:

library(adehabitatLT)
library(sp)

data(ibex)

##Sample Polygon
x_coord <- c(895000,  898000,  898000, 895000, 895000)
y_coord <- c(2031000, 2031000, 2033000, 2033000, 2031000)
xy <- cbind(x_coord, y_coord)
p = Polygon(xy)
ps = Polygons(list(p),1)
sps = SpatialPolygons(list(ps))

Answer 1

You should not use sp because most of the packages that support it are being retired from the end of 2023, and everyone is being advised to move to sf.

Here's a code to detect when ibex is leaving a polygon :

First, I convert ibex dataset to spatial point. I assume that projection is NTF (Paris) / Lambert zone II (EPSG:27572) because it work with this one but you should provide your projection if you want better answering.

library(adehabitatLT) # get ibex data
library(sf) # spatial vector
library(dplyr) # data manipulation
library(tmap) # plotting result

data(ibex)

# Convert data to spatial point
ibex1 <- ibex[[1]] |> 
   filter(!is.na(x) | !is.na(y)) |> 
   st_as_sf(coords = c("x", "y"), crs = 27572) |> 
   arrange(date)

# Create polygon
xy_coords <- data.frame(x = c(895000,  898000,  898000, 895000, 895000),
                        y = c(2031000, 2031000, 2033000, 2033000, 2031000))

poly <- st_as_sf(xy_coords, coords = c("x", "y"), crs = 27572) |> 
   summarise() |> 
   st_cast("POLYGON") |> 
   st_convex_hull()

# find points inside poly
ibex1 <- ibex1 |> 
   mutate(is_inside = st_intersects(geometry, poly, sparse = FALSE)[,1])

When an ibex is leaving, the next point is not inside poly. I use lead function to shift up column is_inside.

is_inside	lead(is_inside)	is_leaving
FALSE	TRUE	FALSE
TRUE	TRUE	FALSE
TRUE	FALSE	TRUE
FALSE	NA	FALSE

# detect leaving
ibex1 <- ibex1 |> 
   mutate(leave_poly = ifelse(is_inside & !lead(is_inside), "leave", NA))

You now have a column that detects the last points before leaving the polygon, in this case the sixth point on 2003-06-02 at 04:00:00.

You can check the result below by converting point to path.

# check result
path <- ibex1 |> 
   arrange(date) |>
   summarise(do_union = FALSE) |>
   st_cast("LINESTRING")

tm_shape(path)+
   tm_lines()+
tm_shape(poly)+
   tm_borders(col = "red")