mf_import_data: Import datasets downloaded using 'modisfast' as a...

In modisfast: Fast and Efficient Access to MODIS Earth Observation Data

Import datasets downloaded using modisfast as a terra::SpatRaster object

Description

Import datasets downloaded using modisfast as a terra::SpatRaster object

Usage

mf_import_data(
  path,
  collection,
  output_class = "SpatRaster",
  proj_epsg = NULL,
  roi_mask = NULL,
  vrt = FALSE,
  verbose = "inform",
  ...
)

Arguments

path	Download root folder supplied to [mf_download_data()], or a folder containing the NetCDF files for one ROI and collection.
collection	Collection identifier, e.g. '"MOD11A1.061"', '"VNP43MA4.002"' or '"GPM_3IMERGDF.07"'.
output_class	character string. Output object class. Currently only "SpatRaster" implemented.
proj_epsg	numeric. EPSG of the desired projection for the output raster (default : source projection of the data).
roi_mask	SpatRaster or SpatVector or sf. Area beyond which data will be masked. Typically, the input ROI of mf_get_url (default : NULL (no mask))
vrt	boolean. Import virtual raster instead of SpatRaster. Useful for very large files. (default : FALSE)
verbose	Character string: '"quiet"', '"inform"' (default), or '"debug"'. Controls progress messages.
...	not used

Value

a terra::SpatRast object

Note

Although the data downloaded through modisfast could be imported with any netcdf-compliant R package (terra, stars, ncdf4, etc.), care must be taken. In fact, depending on the collection, some “issues” were raised. These issues are independent from modisfast : they result most of time of a lack of full implementation of the OPeNDAP framework by the data providers. Namely, these issues are :

• for MODIS and VIIRS collections : CRS has to be provided

• for GPM collections : EPSG:4326 is assigned to the grid

The function mf_import_data includes the processing that needs to be done at the data import phase in order to safely use the data as terra objects.

Also note that reprojecting over large ROIs using the argument proj_epsg might take long. In this case, setting the argument vrt to TRUE might be a solution.

Examples

## Not run: 

### Configure an Earthdata bearer token for LP DAAC Cloud
Sys.setenv(EARTHDATA_TOKEN = "your Earthdata bearer token")

### Set-up parameters of interest
coll <- "VJ121A2.002"

bands <- c("LST_Day_1KM", "LST_Night_1KM")

time_range <- as.Date(c("2026-01-01", "2026-01-30"))

roi <- sf::st_as_sf(
  data.frame(
    id = "roi_test",
    geom = "POLYGON ((-5.82 9.54, -5.42 9.55, -5.41 8.84, -5.81 8.84, -5.82 9.54))"
  ),
  wkt = "geom", crs = 4326
)

### Get the URLs of the data
(urls_vj121a2 <- mf_get_url(
  collection = coll,
  variables = bands,
  roi = roi,
  time_range = time_range
))

### Download the data
res_dl <- mf_download_data(urls_vj121a2)

### Import the data as terra::SpatRast
modis_ts <- mf_import_data(dirname(res_dl$destfile[1]), collection = coll)

### Plot the data
terra::plot(modis_ts)

## End(Not run)

modisfast documentation built on Oct. 5, 2026, 1:06 a.m.