Save a plot for each observation then make a video out of it! Full article at: https://firsttimeprogrammer.blogspot.com/2017/03/resizing-spatial-data-in-r.html

qnet_videos_pictures.R

# Load required libraries
require(dplyr)
require(raster)
require(rasterVis)
require(ncdf4)

# Clean workspace
rm(list = ls())

# Open nc file and get the data
a <- nc_open("qnet_2008.nc")
qnet <- ncvar_get(a, "qnet")
lon <- ncvar_get(a, "lon")
lat <- ncvar_get(a, "lat")
time <- ncvar_get(a, "time")
nc_close(a)

# Generate the expanded grid
grd <- expand.grid(lon, lat, time)
# Set names
names(grd) <- c("lon", "lat", "time")
# Put the data for qnet in a dataframe
vdata <- data.frame(qnet = as.numeric(qnet)) %>% tbl_df()
# Bind with the grid
data_ <- grd %>% tbl_df() %>% bind_cols(vdata)
# Set levels for levelplot
lvls <- pretty(range(na.omit(data_$qnet)), 50)
# Remove variables not needed anymore
rm(a, qnet, lon, lat, grd, vdata)

# Loop over the number of days and save a picture for each day
for(i in 1:length(time))
{
    # Get the data for the current day
    current_data <- data_ %>% filter(time == i)
    # Generate the raster
    x <- raster(nrows = 180, ncol = 360, 
                ymn = min(current_data["lat"]),
                ymx = max(current_data["lat"]),
                xmn = min(current_data["lon"]),
                xmx = max(current_data["lon"]))
    # Assign values to the raster
    values(x) <- as.vector(unlist(current_data["qnet"]))
    # Flip the raster to get the correct plot
    x <-  flip(x, direction = "y")
    # Level plot
    plt <- levelplot(x, contour = FALSE, par.settings = RdBuTheme, at = lvls)
    print(plt)
    # Save the plot
    dev.copy(png, paste("plot_2008_", i, ".png", sep = ""))
    dev.off()
    # Print day
    print(i)
}