Some Question the Data in the 2006 Livingston/Penn Paper

Tomas David Hood (NW7US), the director of HFRadio.org and producer/co-host of the NW7US Space Weather and Radio Propagation Podcast, has received correspondence from a number of individuals that calls into question the research discussed in Episode 4 of the Podcast.  For example:

"In the 2006 journal article, I found the last sentence of the first paragraph of Observations troubling:  1.  If no effort is made to prevent counting the same sunspot twice, one could, even subconsciously, favor sunspots that might favor one's hypothesis and 2.  If multiple counting is allowed and long-lived sunspots have different characteristics than shorter-lived sunspots, the data is skewed toward sunspots with longer lives because there are more opportunities to count these sunspots, more times.  If I were a reviewer of this paper, I would be very uncomfortable with this sentence."

Tomas asked Dr. M. Penn, one of the authors of the 2005 paper, about this issue.  Penn responds:

"As you know the Sun is highly dynamic!  The following movie from MDI on SoHO of the 2001 sunspots dramatically shows this:

http://soi.stanford.edu/press/ssu11-01/MPG/Ic.2001.mpg

Bill's observations occur only 60 days per year, so with clouds, he might have 50 snapshots of the sunspot activity during this movie.  Small pores evolve from hour to hour, and even large umbrae change from day-to-day and week-to-week.  His observation plan minimized bias by measuring every sunspot on the visible disk each day he observed.  Without continuous observations, it's impossible to tell if a particular pore has been observed already; and worse, without seeing the far-hemisphere of the Sun, it's impossible to tell even if large umbrae appearing on the eastern limb have been observed on their previous rotation.  Equally important is the fact that not all sunspots could be observed.

Selection bias was extensively tested in this and following work.  Larger sunspots tend to live longer and thus have a higher probability of being observed multiple times.  The data sets were examined to measure the behavior of different-sized spots, and each size bin showed the same time variation; large spots by themselves showed the same trends that small spots by themselves showed.

We know the sample is incomplete, but with the testing we've done we think that multiple measurements do not introduce a bias which would cause the time variation.

As far as an observer-introduced bias... Bill is internationally known as a very patient and excellent observer.  It's hard to imagine that during the course of 13 years he (consciously or not) subtly changed his sunspot selection to introduce a linear trend, and further that his personal bias would eventually agree with the fact that the next solar minimum would be longer than usual!  Furthermore, current work on automatically selected sunspots from archival data bases supports these IR observations."

Reference:

Episode 4: Disappearing Sunspots by 2015?

Comments

Add New Search RSS

Write comment
Name:
Email:	do not notifynotify
Title:
	Please input the anti-spam code that you can read in the image.

Powered by !JoomlaComment 4.0alpha3

!joomlacomment 4.0 Copyright (C) 2009 Compojoom.com . All rights reserved."