Extract monthly temperature data using coordinates an NC file

Question

I am trying to read temperature data in R using a NOAA OI SST .nc file. I have temperature data per month, but, I am having trouble extracting the monthly average temperature data from the coordinates I want and putting it into a dataframe.

I'm new at this, any help or pointers are most appreciated.

setwd("~/temperatura/noaa"
prueba<-nc_open("sst.mnmean.nc")

#EXTRAER DATOS
    lon<-ncvar_get(prueba,"lon")
    lat<-ncvar_get(prueba,"lat")
    time<-ncvar_get(prueba,"time")
    time=as.Date(time, origin="1800-1-1",tz="UTC")
    sst=ncvar_get(prueba,"sst")

unit<-ncatt_get(prueba,"sst","units")$value

I tried to make a matrix but in time I only have numbers and not the months

matriz <- data.frame(cbind(time,lon,lat,sst))
names(matriz) <- c("time","lon","lat","temperature")

 time   lon   lat temperature
1    4352   0.5  89.5       -1.79
2    4383   1.5  88.5       -1.79
3    4414   2.5  87.5       -1.79
4    4442   3.5  86.5       -1.79
5    4473   4.5  85.5       -1.79
6    4503   5.5  84.5       -1.79
7    4534   6.5  83.5       -1.79
8    4564   7.5  82.5       -1.79
9    4595   8.5  81.5       -1.79

like this

Answer 1

The simplest way to read a .nc file to a dataframe in R is tidync. This is done easily. You will probably have to handle times manually, based on your calendar. I don't think tidync currently has the ability to decode them. If I am correct, the files you are using have a calendar such that the dates are saved as the number of days since 1978-01-01. So you will need to calculate the dates in each file based on this. The following should work:

tidync::tidync("sst.mnmean.nc") %>% 
  tidync::hyper_tibble() %>%
  mutate(date = lubridate::ymd("1978/01/01") + lubridate::days(time))

Answer 2

You can use raster package for that like

library(raster)
r <- brick("sst.mnmean.nc")
r
plot(r, 1)

#Provide longitude & latitude
x1 <- c(0.5, 1.5, 2.5, 3.5, 4.5, 5.5, 6.5, 7.5, 8.5)
y1 <- c(89.5, 88.5, 87.5, 86.5, 85.5, 84.5, 83.5, 82.5, 81.5)
points <- cbind(y1,x1)

#Make points as saptial points
spts <- SpatialPoints(points, proj4string=CRS("+proj=longlat +datum=WGS84"))

#Plot those points
plot(spts, add = TRUE)

#Extract raster values
ex1 <- extract(r, spts, fun='mean', na.rm=TRUE, df=TRUE, weights = TRUE)

#To get the date of the raster you can use
idx <- getZ(r)

#Write the extracted vaues in .csv file for further processing
write.csv(t(ex1[-1]), "File_name.csv", row.names = idx)