raadvrt

VERY MUCH WIP
Don’t use the example code here it’s very likely to crash your session.

The goal of raadvrt is to collect VRT versions of raadtools data sources.

WHY

• can GDALwarp a VRT (so we can get any layer we want in any projection/grid on-demand)
• VRT can be used by other systems (QGIS, Manifold, raster, terra, stars)

This gets us heaps of leverage over raadtools itself, because currently raadtools augments data in R code only. It does this by

• reading raw binary and converting to in-memory raster, applying extent and SRS
• using raster to read netcdf (which uses ncdf4)
• applying known geotransforms and/or crs when needed (using raster::crs(), raster::extent())
• using raster::flip() where needed
• examples include polar data stored with longlat arrays, or old Mercator grids stored with latitude vector
• using raster::rotate() where needed, this is much more efficient via VRT depending on the ultimate target
• multiple proper raster files as tiles in one larger one (e.g. GEOID)
• other weird cases …

In theory we could parameterize the VRT creation, but that’s more lofty and no one else cares it seems.

These fixes are R-only, not communicable to users of other systems, can’t be streamed through GDAL itself, or through other R GDAL wrappers. We won’t give up on our raadtools ways (they are still better some of the time), but we get to choose when streaming through GDALwarp is the better approach.

GDALwarp

I mean with the GDAL API, via the vapour package, not bound by any other GDAL wrapper or in-memory format. But, these will work through other GDAL wrappers as well.

Installation

You can install the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("AustralianAntarcticDivision/raadvrt")

Working cases

NSIDC

These are binary sources with the byte offset and orientation specified, along with the output dimensions, geotransform and SRS for both north and south hemispheres.

CERSAT

Use of a warped source to flip the geotransform from positive y read from the bottom. The files need the projected transform applied, and the SRS because it’s not recorded (only there are longlat arrays), but also to warp with the y-negation to flip.

AMSRE

This is very similar to the CERSAT case, but the source has no longlat arrays to fool tools with, it’s purely in index space (so needs flipping, applying extent and SRS the same).

OISST

There’s oisst_pacific.vrt and oisst_atlantic.vrt, the former merely has the SRS added and the latter has two window sources, to flip from Pacific view (0,360 longitude) to Atlantic view (-180, 180). (There’s no flip needed so there’s no warp with transform negation. )

GEOID

We use VRT currently, a funny case as it’s composed of multiple source files, so to fit the sprintf templating here we need that parameterized out.

SMITH and SANDWELL

Another example where we use VRT currently, because Mercator grid in raw binary.

Example

This is a basic example which shows you how to solve a common problem:

library(raadvrt)
## basic example code

Code of Conduct

Please note that the raadvrt project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.