Issue
I have downloaded multiple Lm3 4 km netcdf files (n = 4640) for sea surface temperature from the Ocean Colour Project, NASA from 2012 to 2024.
I have put all my files into a raster stack:
class : SpatRaster
dimensions : 766, 709, 1 (nrow, ncol, nlyr)
resolution : 0.04165021, 0.04165796 (x, y)
extent : 24.61, 54.14, 5.8, 37.71 (xmin, xmax, ymin, ymax)
coord. ref. : lon/lat WGS 84
source : AQUA_MODIS.20120101.L3m.DAY.SST.x_sst.nc:sst
varname : sst (Sea Surface Temperature)
name : sst
unit : degree_C
> sst_stack
class : RasterStack
dimensions : 766, 709, 543094, 4640 (nrow, ncol, ncell, nlayers)
resolution : 0.04165021, 0.04165796 (x, y)
extent : 24.61, 54.14, 5.8, 37.71 (xmin, xmax, ymin, ymax)
crs : +proj=longlat +datum=WGS84 +no_defs
names : Sea.Surface.Temperature.1, Sea.Surface.Temperature.2, Sea.Surface.Temperature.3, Sea.Surface.Temperature.4, Sea.Surface.Temperature.5, Sea.Surface.Temperature.6, Sea.Surface.Temperature.7, Sea.Surface.Temperature.8, Sea.Surface.Temperature.9, Sea.Surface.Temperature.10, Sea.Surface.Temperature.11, Sea.Surface.Temperature.12, Sea.Surface.Temperature.13, Sea.Surface.Temperature.14, Sea.Surface.Temperature.15, ...
During fieldwork, we collected data for counts of dolphins at coral reefs per month per year. My dataframe spans from 2012 to 2024, and winter is 7 months and summer is 5 months.
I want to split the raster stack into two seasons (summer and winter):
I have read many sources to see how to do this and this solution looks like the closest. I have no idea how to perform this action as I'm a novice
I'm sorry, I can't share the data due to ownership issues
Many thanks if anyone can point me in the right direction.
R-Code
#Download Sea Surface Temperature
folder_SST<"~/Documents/GIS_Data/SST."
files_SST <-list.files(folder_SST, pattern='*.nc', full.names ="TRUE")
for (file in files_SST) {
nc_SST=open.nc(files_SST)
print.nc(nc_SST)
}
#Stack the AQUA-MODIS files with raster stack
sst_stack<-raster::stack(files_SST)
This is what my dataframe looks like
Survey_Number Dates Longitude_E_DD Latitude_N_DD SST
1 2012-08-01 33.89083 27.26778 23.635
2 2012-06-02 33.86782 27.40854 23.640
3 2012-02-07 33.86230 27.44623 23.690
4 2012-02-12 33.88653 27.26957 23.635
5 2012-02-13 33.88766 27.26848 23.635
6 2012-02-14 33.85000 27.36111 23.780
7 2012-02-15 33.86177 27.41302 23.640
Assuming your SpatRaster layer names:
Sea.Surface.Temperature.1
Sea.Surface.Temperature.2
...
correspond to:
AQUA_MODIS.20120101.L3m.DAY.SST.x_sst.nc
AQUA_MODIS.20120102.L3m.DAY.SST.x_sst.nc
...
or whatever your actual file dates are, you could first rename your layers so they are 'intuitively' named with date strings corresponding to their .nc source files, then use grep() and regular expressions to split the stack. Ultimately, it would be better to have date strings in the original SpatRaster to remove any uncertainty, but if my assumption is correct, this will work.
First, an example SpatRaster stack with two years of data:
library(terra)
r_list <- paste0("sst.", 1:731)
set.seed(42)
rr <- lapply(r_list, function(x) {
r <- rast(nrows = 100, ncols = 100, ext(-180, 180, -90, 90))
values(r) <- round(runif(10000, 5, 30),2)
names(r) <- as.character(x)
return(r)
})
sst_stack <- rast(rr)
sst_stack
# class : SpatRaster
# dimensions : 100, 100, 731 (nrow, ncol, nlyr)
# resolution : 3.6, 1.8 (x, y)
# extent : -180, 180, -90, 90 (xmin, xmax, ymin, ymax)
# coord. ref. : lon/lat WGS 84 (CRS84) (OGC:CRS84)
# source(s) : memory
# names : sst.1, sst.2, sst.3, sst.4, sst.5, sst.6, ...
# min values : 5.01, 5, 5, 5.01, 5, 5, ...
# max values : 30.00, 30, 30, 29.99, 30, 30, ...
Now create a vector of new layer names with dates the same length as nlyr(sst_stack) and rename your SpatRaster layers. In this example, I have consecutive dates. If your dataframe contains all 4640 dates and these dates correspond, in order, to all of the layers in your SpatRaster stack, for you it will be something like r_list1 <- paste0("sst.", df$Dates) (assuming df is your dataframe).
# Create vector of equal length to nlyr(sst_stack) to rename SpatRaster layers with dates
r_list1 <- paste0("sst.", seq(as.Date("2012-01-01"), as.Date("2013-12-31"), by = "day"))
# Rename SpatRaster layes
names(sst_stack) <- r_list1
sst_stack
# class : SpatRaster
# dimensions : 100, 100, 731 (nrow, ncol, nlyr)
# resolution : 3.6, 1.8 (x, y)
# extent : -180, 180, -90, 90 (xmin, xmax, ymin, ymax)
# coord. ref. : lon/lat WGS 84 (CRS84) (OGC:CRS84)
# source(s) : memory
# names : sst.2~01-01, sst.2~01-02, sst.2~01-03, sst.2~01-04, sst.2~01-05, sst.2~01-06, ...
# min values : 5.01, 5, 5, 5.01, 5, 5, ...
# max values : 30.00, 30, 30, 29.99, 30, 30, ...
# Note the names are truncated in the print(), but they are like:
head(names(sst_stack))
# "sst.2012-01-01" "sst.2012-01-02" "sst.2012-01-03" "sst.2012-01-04" "sst.2012-01-05" "sst.2012-01-06"
Now use two regular expression to create two separate SpatRaster stacks based on month string matches:
# Create two separate SpatRasters using regex to match month values in layers
sst_summer <- sst_stack[[grep("-0[5-9]-", r_list1, value = TRUE)]]
sst_winter <- sst_stack[[grep("-(0[1-4]|1[0-2])-", r_list1, value = TRUE)]]
sst_summer
# class : SpatRaster
# dimensions : 100, 100, 306 (nrow, ncol, nlyr)
# resolution : 3.6, 1.8 (x, y)
# extent : -180, 180, -90, 90 (xmin, xmax, ymin, ymax)
# coord. ref. : lon/lat WGS 84 (CRS84) (OGC:CRS84)
# source(s) : memory
# names : sst.2~05-01, sst.2~05-02, sst.2~05-03, sst.2~05-04, sst.2~05-05, sst.2~05-06, ...
# min values : 5, 5.01, 5, 5, 5.00, 5, ...
# max values : 30, 30.00, 30, 30, 29.99, 30, ...
sst_winter
# class : SpatRaster
# dimensions : 100, 100, 425 (nrow, ncol, nlyr)
# resolution : 3.6, 1.8 (x, y)
# extent : -180, 180, -90, 90 (xmin, xmax, ymin, ymax)
# coord. ref. : lon/lat WGS 84 (CRS84) (OGC:CRS84)
# source(s) : memory
# names : sst.2~01-01, sst.2~01-02, sst.2~01-03, sst.2~01-04, sst.2~01-05, sst.2~01-06, ...
# min values : 5.01, 5, 5, 5.01, 5, 5, ...
# max values : 30.00, 30, 30, 29.99, 30, 30, ...
One advantage of renaming each layer is that you can then check the correct layers have been carried using names(sst_summer) and names(sst_winter).