The AstroStat Slog

When I was studying astronomy, during when I once became a subject for a social science survey study about life in a department where gender bias is extreme (I was only female), people often asked me how to forecast weather or how to predict future (boys often get questions related to becoming astronauts in addition to weather men and astrologists). Relating astronomy to earth science still happens. Statisticians that I met at conferences, often tried to associate my efforts on astronomical data with those of geologists and meteorologists, who often use stochastic models and spatial temporal models, dimensional extensions of models in time series. Because of this confusion between astronomy and meteorology/geology/oceanology, and the longer history of wide statistical applications found from the latter subjects (a good counter example is the least square method by Gauss but I cannot think more examples to contradict my statement that statistics is used widely among earth scientists with rich history), from time to time my attention has been paid to various applications and models in those subjects so as to find a thread for similar applications for astronomy. Although I don’t like the misconception of astronomy equal to meteorology or geoscience, those scientific fields, what so ever, share at least one commonality that statistical methods are applied to analyzing satellite data.

There is a special issue on Atmospheric Science from the Annals of Applied Statistics, offering me intriguing discussions for finding a common ground between atmospheric science and astronomy. If the general public perception cannot tell the difference between meteorology and astronomy, despite the fact that my affirmative reply to statisticians’ comments on my interests in astronomy always has been “Astronomy and meteorology are very different scientific disciplines,” let’s find out some similarities from how statistics is applied. Astronomers can find more useful applications in the issue from their ends. Here, provided are some interesting ones from my judegment with their [arXiv] links. The whole issue’s table of contents given here: AoAS, vol 2, issue 4 (2008). Most of articles are now to be located at arXiv.

[arxiv:0901.3665] Parameter estimation for computationally intensive nonlinear regression with an application to climate modeling
by D. Drignei, C. E. Forest, and D. Nychka
:I wish for your attention to the sections about constructing a surrogate for the nonlinear complex climate model
[arxiv:0901.3670] Interpolating fields of carbon monoxide data using a hybrid statistical-physical model
by A. Malmberg, A. Arellano, D. P. Edwards, N. Flyer, D. Nychka, C. Wikle
:many astronomers would find more similarities in approaches by reading the abstract than I would. The only difference would be that they are using Carbon Oxide data as a result of the earth green house effect
[arxiv:0901.3494] Interpreting self-organizing maps through space–time data models
by H. Sang, A. E. Gelfand, C. Lennard, G. Hegerl, B. Hewitson
:a good reference for astronomers interested in SOM for high dimensional data and dimension reduction