|
Phenology
|
| Phenology is the study of
the timing of natural events such as migrations, mating, and ice melt.
Recently, long-term phenological records collected by amateur
naturalists, scientists, and other non-traditional sources have been an
important component of our understanding of how the natural world
responds to climate change. Many of these records show long-term trends
toward earlier signs of spring. However, many of the published records
are slightly biased by failing to account for long-term trends in the
date of spring equinox. |
 |
|
The photo
above (by Julie Coghill) shows the people of the town of Nenana, Alaska
raising a wooden tripod on the frozen Tanana river. Since 1917, there
has been a contest to guess the exact minute in spring that the tripod
will fall through the melting ice. The contest, now called the Nenana
Ice Classic, draws hundreds of thousands of entrants with prize money
of over $300,000. The record of the contest is also a great
phenological record, in that it is a very precisely recorded and
consistent record of the time of ice melt each year. Analysis by
Fiorenza Micheli and I shows that on average, ice melt has advanced by
5.5 days relative to the time of spring equinox since 1917, and that
the time of ice melt has closely mirrored climatic trends in the
region.
Read more
about it in our Science article:
Sagarin, R., and F. Micheli. 2001. Climate
change in nontraditional data sets. Science 294:811.
Or check
out the some of the press coverage:
Or check
out the Nenana Ice
Classic web pages.
|
| But all is not exactly
clear in the world of phenology. Specifically, I was troubled
throughout the writing of my paper on the Nenana Ice Classic that
phenological events are typically reported as "number of days since
December 31" (or often simply, and incorrectly, as "Julian Day"). This
reporting leads to strange effects on the data, due to leap years. For
example, the Nenana Ice Classic office reported the earliest year of
breakup as 1940, when the ice broke up on April 20 at 3:27 pm, and the
second earliest year as 1998, when the ice broke up on April 20 at 4:54
pm. When you report these dates and times as elapsed days since Dec.
31, however, 1940 is reported as 110.64 (the 110 days to April 20, plus
the 0.64 of the day needed to get to 3:27 pm) while 1998 is reported as
109.74, an earlier date of break-up. I scratched my head over this for
some time, then I figured out that this is because 1940 is a leap year,
so there is an extra day before April 20 in 1940, but not in 1998,
which is not a leap year. I then thought, "why should leap year, which
is an artificial way of making up for the fact that our calendar
doesn't precisely match the length of the true year, make such a
difference in reporting phenological data?" I finally decided that
phenological events should be reported relative to the timing of vernal
equinox, which is the true start of spring. Here the first problem
appears. It turns out that there are long term trends in the date and
time of vernal equinox - the equinox comes earlier toward the end of
any century (see Figure, below). Because true spring is coming earlier,
reported trends toward earlier spring that don't take this into account
will tend to overestimate the advance of spring in their phenological
data. |
 |
This trend
is driven by leap years (open symbols in figure), which over-correct
for the fact that our 365 day calendar is shorter than the
approximately 365.2424 day "tropical year" (the average time from one
point in the earth's orbit to the same point a year later).
Normally,
this trend "resets" and equinox becomes later at the beginning of any
century because leap years are skipped on most years divisible by 100.
However, years also divisible by 400 are still leap years, so in the
case of our current time, the year 2000 was a leap year, so the equinox
will come earlier until 2096. This will further bias studies of
phenology.
|
|
The figure
at right shows equinoxes (only every 4th year is plotted) from 1000 to
2100. You can see that in centuries that begin with leap years (e.g.,
year 2000), equinoxes come earlier for nearly 200 years.
For more
reading see Duncan Steel's excellent book Marking Time, or read
my short paper in Nature:
Sagarin, R. 2001. False estimates of the
advance of spring. Nature 414:600.
|
 |
|
|
|
|
|