Can't return the standard variable name and reture NAN value When calculate the SPEI index using method ‘MB05’

[CGL5230]

Generic Issue

• xclim version:0.46
• Python version:3.11
• Operating System:Linux

Description

This issue related to this post and this post.

I found I can't get the standard SPEI name,but you can see the variable name is "xarray_dataarray_variable" and the window is 1. The variable name should be 'spei' as document reported and the window should be 6(not 1).

When I try to clip the SPEI dataset, I got the error:"NoDataInBounds: No data found in bounds. Data variable: xarray_dataarray_variable".
It seems the calculation for SPEI index work incorrectly. But I can't find What's wrong.

What I Did

wb_BCC_CSM2MR_ssp585= xclim.indices.water_budget(pr=pr_BCC_CSM2MR_ssp585.pr, tas =tas_BCC_CSM2MR_ssp585.tas, method="MB05")
SPEI_BCC_CSM2MR_ssp585=xclim.indices.standardized_precipitation_evapotranspiration_index(wb=wb_BCC_CSM2MR_ssp585,freq='MS',window=6)
SPEI_BCC_CSM2MR_ssp585.to_netcdf('/path/BCC_CSM2MR2_ssp585_SPEI.nc')
BCC_CSM2MR_ssp585_SPEI = xr.opendataset('/path/BCC_CSM2MR2_ssp585_SPEI.nc')
BCC_CSM2MR_ssp585_SPEI_clip_data = BCC_CSM2MR_ssp585_SPEI .rio.clip(gdf,all_touched=True)

What I Received

"NoDataInBounds: No data found in bounds. Data variable: __xarray_dataarray_variable__".


### Code of Conduct

- [X] I agree to follow this project's Code of Conduct

[aulemahal]

Hi @CGL5230,

I'm not sure this errors comes from xclim. I think the "No data in found in bounds" error relates to your polygon data gdf not overlapping with the grid. If that's the case, that would be an issue for rioxarray. Your grid looks pretty good, so I suspect something's wrong with your gdf ?

The reason for your issue with the variable name and attributes is that you are calling the low-level compute functions¹ instead of the "indicators". You can simply replace .indices. by .atmos. in your case:

wb_BCC_CSM2MR_ssp585= xclim.atmos.water_budget(pr=pr_BCC_CSM2MR_ssp585.pr, tas =tas_BCC_CSM2MR_ssp585.tas, method="MB05")
SPEI_BCC_CSM2MR_ssp585=xclim.atmos.standardized_precipitation_evapotranspiration_index(wb=wb_BCC_CSM2MR_ssp585,freq='MS',window=6)
SPEI_BCC_CSM2MR_ssp585.to_netcdf('/path/BCC_CSM2MR2_ssp585_SPEI.nc')

This way, you'll have better metadata, better missing value checks and a better variable name.

Footnotes

1. We call the low-level compute functions "indices" and the high-level ones "indicators". I agree that might be misleading at first. ↩

[coxipi]

window = 1 instead of window = 6

Thanks for noting this! At least, it's just another error in the attrs, so the computation "does" use window=6. I will fix this ...

[CGL5230]

Thanks for your reply. I will check my gdf again. At least, the result from xclim is good.