从光栅文件中提取特定坐标的数据并在R中合并

Question

I have gridded data files ( each file consist of monthly rainfall and is label pYYYYMM) https://www.dropbox.com/sh/63z166tjxyu12s5/AAAs3Ccn1zdVoBYMj8Y1o303a?dl=0

library(raster)
files= list.files( ,pattern='*.grd',full.names=TRUE)
df = NULL
for(i in seq_along(files)) {
r2 <- raster(files[i])
# setting the missing values to NA
r2[r2 >= 170141000918782798866653488190622531584.00] <- NA_real_
# Setting the projection
crs(r2) <- "+proj=stere +lat_0=90 +lat_ts=70 +lon_0=-45 +datum=WGS84"
plot(r2)
#Doing this, got me the value for a given point but you will have to manual edit:
d1<- getValuesBlock(r2, row=42, nrows=1, col=26, ncols=1)
# Combine data to get one dataframe 
rbind(df,d1)->df
# then write to a csv file
}

The loop above works for me to an extent but it means that I have to manually change each point and then I will also have manually input the date.

I am kindly asking for some guidance on how I can modify my code to the data for the latitude and and longitude of my interest which is in this file https://www.dropbox.com/s/3p4u4pyxkyo2q15/latandlong.csv?dl=0

Answer 1

This is how you would normally do this, assuming that all your grid files line up

library(raster)
library(rgdal)
files <- list.files(pattern='.grd$', full.names=TRUE)
s <- stack(files)
# points <- ...

# if your points are lonlat
points <- SpatialPoints(points[, c('longitude', 'latitude')], 
          proj4string=CRS('+proj=longlat +datum=WGS84'))
# and your raster data are not they need to be matched
pts <- spTransform(points, CRS("+proj=stere +lat_0=90 +lat_ts=70 +lon_0=-45 +datum=WGS84")

df <- extract(s, points)
write.csv(df, 'file.csv')

Answer 2

Your data is in two different coordinate systems or the .grd files are bad. I had a look at your actual data and my guess is that your data is from Canada, the lat and longs are in wgs-1984 and I'm unsure what is going on with your .grd files. There lower left hand corner starts at 0,0. If you extract correctly you will have all NA values because the data do not overlap. This is a data issue and not an R question.

library(raster)
library(rgdal)
#get extent of grid file
dat.dir<-'filelocations'
a<-raster(paste0(dat.dir,'/p201307.grd')
extent(a)
#class       : Extent 
#xmin        : -0.5 
#xmax        : 124.5 
#ymin        : -0.5 
#ymax        : 94.5 
#get locs
locs<-read.csv(paste0(dat.dir,'/latandlong.csv'))
head(locs)
#longitude latitude
# 1 -116.8736  60.0334