search

PCMDI / pcmdi_metrics

Open-source Python package for Systematic Evaluation of Climate and Earth System Models
http://pcmdi.github.io/pcmdi_metrics/
BSD 3-Clause "New" or "Revised" License
98 stars 37 forks source link

[Feature]: Reduce number of plots from modes of variability #1120

Open lee1043 opened 3 weeks ago

lee1043 commented 3 weeks ago

Is your feature request related to a problem?

Suggested by @zhangshixuan1987 at https://github.com/PCMDI/pcmdi_metrics/pull/1112#issuecomment-2328128362 and discussed in https://github.com/PCMDI/pcmdi_metrics/pull/1112#issuecomment-2328424624.

Describe the solution you'd like

Maybe plot option can be elaborated to plot_eof, plot_additional_eofs, and plot_cbf?

zhangshixuan1987 commented 3 weeks ago

I think a clearer group of the output figures can be enough to address the confusions. Here is my idea:

  1. Take SAM, PSA1 and PSA2 as examples, they were first three modes of EOF analysis from either sea level pressure or 500hPa geopotential height.
  2. In theory, we expect that the first, second and third EOFs are corresponding to SAM, PSA1 and PSA2, respectively. Therefore, to my understanding, we can group the output figures into three groups:
    • “Mode pattern”: save the expected EOF pattern (theoretical), i.e. EOF1 from reference and model as they are supposed to be the SAM.
    • “Best Matching EOF pattern”: i.e. the EOF patterns that can best match expected EOF pattern (theoretical) for SAM, PSA1 or PSA2.
    • “EOF Pool”: the first three leadings EOFs in reference and model as additional information

I am just sharing the thoughts from me as a references for you, and my group maybe not optimal.

However, split plot option to plot_eof, plot_additional_eofs, and plot_cbf also sounds good to me as it ineed allows users to decide if they want the addental figures. Thank you!