Harvard-Smithsonian Center for Astrophysics Eric's Cool Astronomy Plots

I love to make plots. How often they get published is another matter. Here are a few of my favorites, some of which might even be useful.

• Distance (parsecs) vs. age (in billions of years; Gyr) for the nearest 100 solar-type dwarf stars. Plot made from data in Table 13 of Mamajek & Hillenbrand (2008). The ages were inferred from chromospheric activity levels from the F7-K2 main sequence stars, using the revised rotation vs. age and rotation vs. activity calibrations from this paper. You can think of this as the distribution of ages of the nearest (potential) planetary systems to the Sun, for the nearest Sun-like stars in our Galactic neighborhood.
  
  
• Color-magnitude diagram (B-V vs. Mv) for stars within 80 pc, with color coding by spectral type
  
  
• Color-magnitude diagram (B-V vs. Mv) for stars within 30 pc, with solar metallicity evolutionary tracks
  
  
• Color-magnitude diagram (B-V vs. Mv) for the young (~5 million-year-old), nearby (145 parsecs) OB association Upper Scorpius.
  
  
• Bluest Main Sequence B-V color for a given age/isochrone
  
  
• Pre-MS contraction time versus stellar mass: How long does it take a pre-main sequence star to contract and reach the zero-age main sequence? It takes a 1 solar mass star roughly 44 million years to contract to the point at which hydrogen fusion stabilizes (reaches the main sequence). Plot was contructed using the D'Antona & Mazzitelli evolutionary tracks.
  
  
• Watch Proxima Centauri run!
  
  
• Cumulative number of exoplanet discoveries versus time. It appears that the number of known extrasolar planets is doubling every 27 months -- displaying a behaviour similar to Moore's law, but with a slightly longer time constant. For illustation's sake, if you assume that the Milky Way has 400 billion stars, and each star averages one planet per star, then this relation(?) would predict that we will complete our census of these worlds in late 2073 (hmm, I wouldn't bet on it). Data taken from the Extrasolar Encyclopedia.
  
  
• The distribution of known O-type stars, viewed from above the Galactic plane, with spiral arms (from Vallee 2002). O-stars are from the Maiz-Apellaniz et al. catalog, where I calculated distances using the Mv and (B-V)o values from Martins et al. 2005. Here I assume the Sun is 8 kpc from the Galactic center. The anticorrelation of the O-stars with the arms appears to be due to the magnitude-limited nature of the O-star catalogs. There tend to be more dark molecular clouds in the "gaps" where there are no O-stars.
  
  
• B-V vs U-B color-color plot of OB and A0V stars. The plot gives an improved fit for deriving intrinsic (B-V) colors for OB stars using Johnson's Q-method (I had noticed that some of the formulae for deriving intrinsic B-V from the Q-method for high-mass members of the Sco-Cen OB association were giving more unphysically negative reddening values (E(B-V)) than one might suppose just from photometric errors. This plot shows why -- the previous calibrations do a somewhat poor job of fitting the "blue envelope" of colors for unreddened nearby B-type stars by attempting to force their fit through (B-V, U-B = 0, 0) for A0V stars.
  
  
• "The Lithium Plot": A crude age indicator for cool stars. This is a plot of stellar effective temperature (Teff) versus the equivalent width of the Li I 6707A line for stars in clusters of "known" age. Stars appear to be born with a more-or-less "cosmic abundance" of Li (roughly 1 Li atom for every 500 billion hydrogen atoms!). Li is burned in stellar interiors at relatively low temperatures (~1-2 megakelvin), but it is burned relatively slowly in stars like the Sun since they have thin convective shells that do not allow the Li to reach great depths and high temperatures.
  
  I'm also a weather and tropical cyclone nut... so here are some plots.
  
  
• Atmospheric CO2 concentration vs. annual global mean temperature. The statistical correlation is very strong, however correlation does not necessarily imply causation (but in this particular case, there is a well-studied physical mechanism for understanding why more CO2 might lead to higher temperatures!). There is, however, not a well-studied physical mechanism for understanding why the number of pirates and global mean temperatures should be correlated (save perhaps that pirates are "cool"). Arrr!
  
  
• Atmospheric CO2 concentration vs. yearly accumulated cyclone energy (ACE) for the Atlantic Hurricane Basin (1851-2007) and satellite era (1959-2007). X-axis is the atmospheric concentation of carbon dioxide, and Y-axis is "ACE" - accumulated cyclone energy -- a convenient way of measuring the strength of a given hurricane season. Essentially ACE scales as the product of the kinetic energy produced by tropical cyclones times their longevity (storms that are strong or long-lived contribute most to ACE). The pre-~1960 data is less reliable as some at-sea storms may have been missed.
  
  Given the number of times I've heard non-experts say on TV that higher atmospheric CO2 "will" translate into worse hurricane seasons to come, I'm amazed that I've never seen this plot anywhere. Arguably, ACE is the best metric of hurricane season strength -- as the number of types of storms of different strengths may be biased due to what technology is available at the time to assess their status. The plots above are very interesting. As the previous plot showed a strong correlation (r ~ 0.9) between CO2 and global temperatures, both the 1851-2007 and 1959-2007 CO2 vs. ACE plots show substantially weaker correlations (r ~ 0.22) [perfect anticorrelation r = -1, no correlation r = 0, perfect correlation r = +1]. But how significant? Given the large sample size, r appears to correspond to a significant correlation. For the satellite era sample (1959-2007) r = 0.25, with the probability of zero correlation being ~8%. For the whole sample (1851-2007) r = 0.23, with the probability of zero correlation being only ~0.4%. So while not overwhelming, the current data are suggestive of a positive correlation. Taken at face value, the measured slopes are delta(ACE [kt^2])/delta(CO2 [ppm]) = 0.46+-0.19 (1851-2007) and 0.73+-0.46 (1959-2007). Given a mean ACE value of ~86 kt^2, the 1851-2007 trend suggests that each additional ppm of CO2 added to the atmosphere enhances the potential accumulated cyclone energy for an Atlantic Hurricane season by ~0.5 +- 0.2%. Over the past 10 years, the CO2 in the atmosphere has increased on average by 2.0 ppm/yr, i.e. ~1% increase in potential ACE per year. Here is a dangerous extrapolation of these trends, which assumes that CO2 increases at the pace seen between 1959 and 2006, and the Atlantic Hurricane ACE increases gently as measured in this figure. Extrapolating the quadratic trend in increasing CO2 between 1959-2006 would suggest that by year 2100, the atmosphere will have ~673 ppm of CO2 (assuming status quo increase) and a normal Atlantic season would have ACE of ~250 -- equivalent to that for the hyperactive 2005 hurricane season. Is this possible?
  
  The crazy thing is that the Atlantic Hurricane seasons have been unusually active since 1995 (ostensibly the Atlantic Multidecadal Oscillation has returned us to a period of enhanced cyclone activity). If you measure the same trends during the 1851-1994 period -- the probability of no correlation between CO2 and ACE for Atlantic Hurricane seasons was 94% (r = 0.006)! During the 1851-1994 period, the trend of CO2 vs. ACE was essentially flat: delta(ACE [kt^2])/delta(CO2[ppm]) = 0.02 +- 0.17. Yet atmospheric CO2 increased by 26% from 285 ppm (1851) to 360 ppm (1994) during that period! This suggests that the majority of the positive CO2 vs. ACE trend has come just from adding the hurricane seasons over the past decade or so. It will be interesting to see in the coming decades whether the mean ACE decreases again to levels seen during the quieter 1970-80's, or whether it continues to march upward...
  
  
• Mean annual temperature for Pittsburgh from 1871-2006. I was curious to see if there was a demonstrable long term trend in temperature for my hometown over the past century (ostensibly due to global warming). The answer was surprising. If you fit a slope to all of the yearly data back to 1871, one finds a negative slope (i.e. cooling), however the 90's and 00's clearly buck this trend in the warming direction. Another interesting find: the 90's were slightly warmer than the first 7 years of the 2000's (by ~0.3 deg F), however this is at the level of the uncertainties in the decadal mean temperatures (~0.3-0.5 deg F). Most interestingly, according to the NOAA data, the warmest decades were (in order of highest temperature): 1890s, 1880s, 1930s, 1900s, 1920s, 1910s, 1990s, 1940s, 2000s (2000-2006). So while the mean decadal temperatures in Pittsburgh during the 1990s and 2000s are higher than during the 1960s, 1970s, and 1980s, they are clearly not unprecendentedly high when compared to the late 19th and early 20th century.
  
  
• Annual precipitation for Pittsburgh from 1837-2006. The slope of annual rainfall vs. time is slightly positive (wetter vs. time), but statistically consistent with zero. So, the NOAA data suggests that there has been no significant long-term trend in annual precipitation in Pittsburgh over the past century and a half. Also, the variance (i.e. year-to-year precipitation totals; a possible bell-weather of "extreme" changes) does not appear to be significantly different in the last three decades compared to before.
  
  
  
• Cumulative distribution of monsoon starting dates for Tucson, AZ (1947-2005) and the relevant data and statistics
  
  And more random stuff...
  
  
• Gold Prices 1970-2006.
  
  
• The Fed lowered interest rates to 2.0%. Since 1954, the rate has only been at or below this level 10% of the time. Fed Interest Rate 1954-2008 (monthly average)
  
  
• Age distributions for 3 very different cities, all with populations around 30,000: Cambridge MA (college town), Bethel Park (suburb), and Sun City AZ (retirement city). Peak ages are 20, 42, and 77, respectively.
  
  
  
• Basic Astronomical Data for the Sun (BADS?). If you work on stars other than the Sun (like me), and occasionally need to know the solar value, this may or may not be useful.
  
  

Back to main page Comments or Questions? Contact www-admin@cfa.harvard.edu