CODEWITHSUNDEEP
rgis
Sophie Père
Sophie Père

Reputation: 47

Calculating average latitudinal difference from two data frames and converting to kilometers

I have two data frames, DF1 and DF2, containing spatial data with longitude/latitude at the scale of North America, and their projection is EPSG:4326. I would like to calculate the difference between the mean value (M1) of the y coordinates of DF1 and the mean value (M2) of the y coordinates of DF2, and I want the difference to be expressed in km. I am not sure which projection system to use to convert longitude and latitude to km in my case.

I discovered the function terra::expanse() to compute the area covered by raster cells that are not NA by using the longitude/latitude CRS. It is stated that it is better to use the longitude/latitude CRS to calculate areas.

I could not find the source code for this function to see if there might be a way to convert the values M1 and M2 to km. I am a novice in GIS, so thank you for your help. Here’s a summary of my data frames and the values M1 and M2:

> summary(DF1 [,c(“x”, “y”)])
 decimalLongitude  decimalLatitude
 Min.   :-171.62   Min.   :14.88 
 1st Qu.:-118.62   1st Qu.:41.88 
 Median :-103.38   Median :53.38 
 Mean   :-104.81   Mean   :51.95 
 3rd Qu.: -87.88   3rd Qu.:63.62 
 Max.   : -52.88   Max.   :76.38
 
 
> summary(DF2 [,c(“x”, “y”)])
 decimalLongitude  decimalLatitude
 Min.   :-171.62   Min.   :15.62 
 1st Qu.:-118.62   1st Qu.:42.12 
 Median :-103.12   Median :53.38 
 Mean   :-104.70   Mean   :52.04 
 3rd Qu.: -87.88   3rd Qu.:63.62 
 Max.   : -52.88   Max.   :76.38
 
 
 
> M1 <- mean(DF1[,c(“y”)], na.rm = TRUE)
> M1
[1] 51.95062
 
> M2 <- mean(DF2[,c(“y”)], na.rm = TRUE)
> M2
[1] 52.03755
 
> M2 - M1
[1] 0.08692784 ## how to convert to km ?

Upvotes: 0

Views: 51

Answers (1)

Robert Hijmans
Robert Hijmans

Reputation: 47146

You can compute the distance between longitude/latitude points with terra::distance or geosphere::distm

Example data

d1 <- data.frame(lon=0:1, lat=0:1)
d2 <- data.frame(lon=30:33, lat=20:23)

To get the actual distances between the points, you could do

x1 <- terra::distance(d1, d2, lonlat=TRUE, unit="km")
mean(x1)
#[1] 4088.824

x2 <- geosphere::distm(d1, d2, distGeo)
mean(x2) / 1000
#[1] 4088.824

Since you only care about the mean latitudinal distance, we can set longitude to zero and use the mean latitudes.

md1 <- data.frame(lon=0, lat=mean(d1$lat))
md2 <- data.frame(lon=0, lat=mean(d2$lat))

terra::distance(md1, md2, lonlat=TRUE, unit="km")
#[1] 2323.15
geosphere::distGeo(md1, md2) / 1000
#[1] 2323.15

You can also compute that like this

d1$lon <- 0
d2$lon <- 0
dm <- terra::distance(d1, d2, lonlat=TRUE, unit="km")
mean(dm)
#[1] 2323.158

Your data are longitude/latitude coordinates. You should not project them to compute distance, as that could lead to inaccurate results due to distortion.

Upvotes: 3

Related Questions