Greetings Valeriia Bondarenko

We will need some more details:

1) I assume that you want a value for each province, and this value must be something like the mean value of all 248 timestamps?

2) I dont really understand what do you mean with "need to find the four closest points to the center of the grid".. closest from what? The centrer of the provinces?

I think you can load the data in R and get their mean temperature value, if this is what you want.

path= "your path to the nc_data"

temperature = brick(path)

temp_mean = mean(temperature)

This would give you the mean temperatures

How to get province centers

https://gis.stackexchange.com/questions/43543/how-to-calculate-polygon-centroids-in-r-for-non-contiguous-shapes/43558

This is a guide that could help you. You load your shapefile with countries and provinces, then you can calculate their centoids, and then get their temperature from your raster

Then, to calculate the temperature value for the center of each province:

prov_center # the center of each province

prov_temp = raster::extract(temp_mean, prov_center)

this can be used for each province center

I am assuming you're interested in the interpolation of your NC file. I would suggest using CDO (https://code.mpimet.mpg.de/projects/cdo/). For example; bilinear interpolation can be achieved by "cdo remapbil, n32, input.nc output.nc".

Hello,Panos

Thanks a lot for your reply.

I need to find 4 points(latitude longitude),which are closest to the center of province and calculate the distance between each point to the centre of province.

For example,for province Bangkok:lat:5.63 to 20,46; lon:97.3 to 105.4.

Calculate the center point of multiple latitude/longitude and calculate the distance between center of province and latitude/longtitude (to find the corresponding grid centroids)

Valeriia Bondarenko

From my understanding, the four points you want to find, correspond to a raster cell right?

And the province is presented by a single point (lon, lat), right?

In essence, you want to find the raster cell where the province point lies in, right?

The raster::exctract does just that!

You input your province point, then it searches for the grid cell where the points lies in and gives you the temperature value from that cell!

If you only care about the value in the cell, you don't need to find the exact lon/lat points!

Is this what you need?

If you want to find the exact four points as well:

Step1: load the lon and lat variables from your nc

nc_name = your path to your nc data

lon

Panos,yes,exactly,this cell in nc file..

Maybe,it's possible to do in R

Valeriia Bondarenko

Yes! It is absolutely possible to do this in R!

You can use the code from my previous replies and adjust it to your needs!

I attached a code sample with all the code I gave you!

This is the general procedure to extract a value (i.e., temperature) from a raster based on specific points (i.e., your province centers).

Hope that helps!

Panos,thank you very much for your help

May I just ask you the last question please,if I just need to get centroids and save the file (if i've a shape file of country),is it correct the code?

library(raster)

library(rgdal)

library(rgeos)

library(sf)

library(GISTools)

library(maptools)

sids

Valeriia Bondarenko

My pleasure to help!

The code seems fine. I cannot tell 100% because I can't run it, but if you calculate the centroids, your code seems to save them in a shp object!