Extract time series of a point ( lon, lat) from netCDF in R

Question

I am relatively new on R. I am trying to get time series of different points ( lat, lon) of temperature data from a netCDF file. My sample data file is here and here is the small file . I have tried netCDF package and the code i have used so far

library(ncdf)
obsdata = open.ncdf("obs.nc")

print.ncdf(obsdata) 

obsdatadates = obsdata$dim$time$vals
obsdatadates = as.Date(obsdatadates,origin = '1950-01-01') 
obsdatadates
obsoutput = get.var.ncdf(obsdata, varid = 'tasmin', start = c(1,1,1),
                         count = c(1,1,22280))
dim(obsoutput)
datafinal=merge(obsdatadates,obsoutput)

Can anyone help me to get a dataframe of timeseries ( first column) and value of data in another for a particular points( lat, lon) of that data. In this case I am looking for time series ( 1950-01-01 to 2010-12-31 for which the data is ) for a particular lat lon point ( and repeat for many points of interests) and for given variable(in this case tasmin). Your help would be appreciated. Thank you, aseem

Answer 1

Perhaps use the raster package, this won't work for all NetCDF files but it does for yours:

library(raster)
## brick reads all 22280 layers
r <- brick("obs.nc", varname = "tasmin")
## extract works for all time steps
vals <- extract(r, matrix(c(-120, 52.5), ncol = 2))

dim(vals)
## [1]     1 22280

Note that gives a 1-row, many column matrix because I only gave a single point to extract().

(The extraction is simple with direct copy from the nearest cell, use method = "bilinear" to do interpolation). See ?extract for other options.

Answer 2

Here's how I would proceed to do this with ncdf:

library(ncdf)
obsdata = open.ncdf("obs1.nc")
obsdatadates = as.Date(obsdata$dim$time$vals,origin = '1950-01-01')
#Get the whole data first
obsoutput = get.var.ncdf(obsdata, varid = 'tasmin')
#Prepare your points of interest
points_of_interest = data.frame(lat=seq(1,8,by=2),lon=c(1,5,3,6))
#Subset your data accordingly
data_at_point = apply(points_of_interest,1,function(x)obsoutput[x[1],x[2],])
#Turn it into a dataframe
data_at_point = as.data.frame(data_at_point)
#Add the dates to the dataframe
data_at_point$Date = obsdatadates

Answer 3

the 'ncdf'-package is deprecated: http://cirrus.ucsd.edu/~pierce/ncdf/

update with the ncdf4 package:

library(ncdf4)
obsdata <- nc_open("obs1.nc")
print(obsdata) # check that dims are lon-lat-time

# location of interest
lon <- 6  # longitude of location
lat <- 51 # latitude  of location

# get dates
obsdatadates <- as.Date(obsdata$dim$time$vals, origin = '1950-01-01')

# get values at location lonlat
obsoutput <- ncvar_get(obsdata, varid = 'tasmin',
                  start= c(which.min(abs(obsdata$dim$longitude$vals - lon)), # look for closest long
                           which.min(abs(obsdata$dim$latitude$vals - lat)),  # look for closest lat
                           1),
                  count = c(1,1,-1)) #count '-1' means 'all values along that dimension'that dimension'
# create dataframe
datafinal <- data.frame(dates= obsdatadates, obs = obsoutput)