2

I'm trying to extract climate variables in QGIS for sample points with a raster dataset. The raster comes from ClimateNA (https://adaptwest.databasin.org/pages/adaptwest-climatena) and is available as in ASCII (.asc) and NetCDF (.nc) formats.

When I load either file format to QGIS, I get "CRS was undefined: defaulting to CRS EPSG:4326 - WGS 84," which is my project CRS. The raster is visible in my Layers tab but not on the map, and when I use the Point Sampling Tool to extract data for my points, the output CSV lists all my sample points but has an empty column for the variable.

Does anyone know what I need to do to fix this?

1
  • You need to figure out what CRS the raster should be in, and put it in that CRS. That information should be somewhere on the website where you downloaded it. CRS stands for "coordinate reference system", and it's sometimes also called "spatial reference system" (SRS) or it might be broken out into "projection" and "datum". – csk Feb 3 '20 at 21:31
2

For me, the clue/solution was in the metadata linked here:
http://www.cacpd.org.s3-website-us-west-2.amazonaws.com/climate_normals/NA_ReadMe.txt

Coordinate Systems
------------------

X,Y: 
Projection: Lambert Conformal Conic
False Easting: 0.00000000
False Northing: 0.00000000
Central Meridian: -95.00000000
Standard Parallel 1: 49.00000000
Standard Parallel 2: 77.00000000
Latitude Of Origin: 0.00000000
Linear Unit:  Meter
Datum: D WGS 1984

Using the CRS Selector dialog in QGIS, I set my project to the following Lambert Conformal Conic (North America): ESRI:102009 enter image description here

After that, the data was aligned properly with the Stamen WMS layer. The PRJ file included with the ASCII raster does load properly into QGIS, however as a custom projection.

I am using the latest non-LTS build of QGIS 3.10.2-2, and I recommend you do the same. There have been many fixes to GDAL+PROJ which have improved CRS handling over prior releases.

enter image description here

2
  • 1
    Thanks for the reply! This definitely moves me in the right direction. I got QGIS 3.10.2, and was able to reproject the CRS for the ASCII and NetCDF rasters, base map, and sample points to Lambert Conformal Conic (North America): ESRI:102009. As far as I can tell, I've replicated what you did, but the rasters don't line up with the Google base map, and the sample points coordinates are all at "Null Island" (0,0) aka off the west coast of Africa. Any ideas? – Brian Allen Feb 3 '20 at 23:26
  • Don't change the CRS of the layers, just the project. Start clean and try again. It should line up fine. – Saijin_Naib Feb 4 '20 at 0:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.