Closed cecileherr closed 2 months ago
FYI: for most of the input layers, we should be able to work with Int format. Seepage and nitrogen_atmospheric are layers that as Float could come
I can not reproduce this locally. Can you show which versions of rasterio and gdal you use. If you just type simple
after running the code above, this should be reported. Eg:
# Niche Vlaanderen version: 1.2b1
# Newer niche_vlaanderen 1.2b2 available
# Run at: 2023-03-08 12:40:22.688925
package_versions:
pandas: 1.4.0
numpy: 1.22.1
rasterio: 1.3.4
gdal: 3.5.3
python: '3.9.2 (default, Feb 28 2021, 17:03:44)
[GCC 10.2.1 20210110]'
model_options:
deviation: false
full_model: false
name: ''
strict_checks: false
model_properties:
model_extent: ((216590.0, 198380.0), (216905.0, 198170.0))
input_layers:
mhw: ../testcase/zwarte_beek/input/mhw.asc
mlw: ../testcase/zwarte_beek/input/mlw.asc
soil_code: ../testcase/zwarte_beek/input/soil_smallerextent.tif
# Model run completed
Running conda list in my niche environment gives me this: gdal 3.5.2 py39h34c8707_1 conda-forge rasterio 1.3.2 py39h223f425_0 conda-forge
Some additional testing on windows/linux show the error. No data is correctly parsed on the input layer, but not after calculations.
Expected behaviour: if there is a NoData in any input raster for a given pixel, the output should be NoData
This works (nodata in any input raster gives nodata in output)
This doesn't work (NoData in soil map becomes no potential)
soil_smallerextent.zip
Is there something we can do about it? Or should we add a warning (avoid Float), at least in the documentation?