Calculating polygon area with R geosphere package

Question

I'm having trouble with the areaPolygon function in R. Sometimes it appears to produce the correct result, and sometimes the results seem orders of magnitude off.

E.g. I have a polygon with these points:

lng <- c(-51.74768, -51.74768, -51.74735, -51.74735)
lat <- c(-0.1838690, -0.1840993, -0.1840984, -0.1838682)

and combine them into a dataframe

data <- data.frame(lng, lat)

Then I try to get the area:

area <- geosphere::areaPolygon(x = data)

which comes out to 326928.8 m^2.

I was expecting something more in the 1000 m^2 range.

What's weird is that this seems to work fairly well for about half the cells I'm calculating, but not the other half.

A similar StackOverflow question noted strangely small values, due to the points being in an incorrect order. However, when I plot, e.g.

plot(data, type="l")

The polygon seems to be drawing correctly.

Anyone know what might be wrong here? Thanks!

Answer 1

With your example data, result is as expected, as noted in the comments.

library(geosphere)
lng <- c(-51.74768, -51.74768, -51.74735, -51.74735)
lat <- c(-0.1838690, -0.1840993, -0.1840984, -0.1838682)
d <- cbind(lng, lat)
geosphere::areaPolygon(d)
#[1] 935.2693

The idea that rounding plays a role seems far-fetched.