seismology.estimate_numax() does not return accurate numax values for low numax red giants.

I am trying to use the lightkurve numax_estimate() to get initial numax values for red giants, but the results for stars with low numax values seem to be wrong. I have included a MWE of a targeted nominal Kepler red giant where this problem occurs below.

from lightkurve import KeplerCBVCorrector
import lightkurve as lk
import numpy as np
import matplotlib.pyplot as plt

lcc = lk.search_lightcurvefile('KIC 2437444').download_all()
lcall = lcc.stitch()
ps = lcall.remove_nans().to_periodogram(freq_unit='microHertz',oversample_factor=5)
sn = ps.flatten()
seis = sn.to_seismology()
seis.estimate_numax()

This should return a numax value of approx. 21 microHertz but the result is instead 13.5 microHertz.

• platform (Linux):
• lightkurve version (1.9.1):

[barentsen]

Thanks for helping out @astrobatty !! 👍

@ojhall94 @danhey Do you think we need to revise the defaults here, or do you think this issue can be closed?

[astrobatty]

You can find the answer in the doc:

A window of fixed width (either given by the user, 25 microhertz for Red Giants or 250 microhertz for Main Sequence stars) is moved along the power spectrum, where the central frequency of the window moves in steps of 1 microhertz (or given by the user as spacing) and evaluates the autocorrelation at each step.

As you do not provide any specific window length and/or step, the default values are used. In your case the autocorrelation function (ACF), from which the numax is estimated, is calculated with a window length of 25 uHz. As your expected numax is less than 25 uHz, you simply oversmooth the PSD.

Solution

Change this line seis.estimate_numax() to this seis.estimate_numax(window_width=5, spacing=0.1).

You can compare the ACFs using seis.diagnose_numax();. Plot the results using the default and new values to see the problem here.

Request

Maybe a warning to change window_width/spacing would be great if estimated numax is low.