Skip to content

Calibration period / prediction period #81

@fmaussion

Description

@fmaussion

Is there an existing issue for this?

  • I have searched the existing issues

Describe the question

I'll start by saying I'm not an expert in applying SPEI, so my question might seem naive, but:

I’m wondering if it would be meaningful to add an option to fit the SPEI distribution once over a fixed calibration period (e.g., 1981–2010), and then apply the fitted parameters to all time steps in the series (or across multiple scenarios).

Unless I'm mistaken this is a use case in climate change studies, where we want to compare future projections (e.g., different SSPs) relative to a consistent historical baseline. It's also an option in climate-indices (calibration_start_year, calibration_end_year).

If I understand this correctly, the fit_window parameter enables a moving calibration window, which is great for non-stationary real-time monitoring — but makes direct comparisons across time or scenarios more difficult? :

Thanks again!

Metadata

Metadata

Assignees

Labels

questionFurther information is requested

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions