Thursday, April 3, 2014

Nenana Ice Classic - Early Breakup?

An entertaining ritual of spring in Alaska is the famous Nenana Ice Classic guessing game, wherein hopeful participants submit estimates of the precise time that the ice will go out on the Tanana River at Nenana.  Last year's breakup was the latest on record, although the breakup in 1964 was technically a day later if you account for the fact that 1964 was a leap year.

The ice at Nenana this year is a little thinner than usual, which has led some to speculate that the breakup will be early, although the historical correlation between early April ice thickness and the breakup date is very small.  A much more important factor is the temperature in April - and in early May if the breakup is late.  Of course, for those playing the Classic, the submission deadline of April 5 makes it impossible to know with confidence how temperatures will evolve in the last 2 or 3 weeks prior to breakup.  However, I find it interesting to note that the long-range temperature forecasts from the National Weather Service's CFSv2 model are modestly skillful and therefore allow for a slight edge over random chance in the guessing game.

The chart below shows the breakup date (day of the year, vertical axis) against the forecast temperature for April, according to the CFSv2 forecasts as of April 1 of each year since 1982.  The forecasts for prior years are obtained from the historical "reforecasts" provided by NOAA - these are the retrospective forecasts used to calibrate the operational forecasts of today.  I extracted the forecast average temperature over the Tanana River drainage basin upstream of Nenana, with the idea that upstream temperatures are probably important as well as temperatures at Nenana itself.  There is a modest negative correlation between the forecast temperature and breakup date (R=-0.65), which is not particularly impressive but is better than nothing.  (Note that the actual April temperature in Fairbanks predicts the breakup date with R=-0.85, and the forecast predicts the actual with R=+0.77.)

Simple linear regression allows for a crude estimate of the potential effect of the expected temperature on the breakup date.  This year's forecast temperature for April is 23.0 °F, or 1.9 °F above normal, which translates into a breakup date of April 30, or 2 days earlier than the 1982-2013 mean.  Of course, this is only a slightly better guess than simply taking the mean; the error distribution on the regression indicates that the probability is about 66% that the breakup will be earlier than May 2.

From a subjective standpoint, I would guess that the breakup will be earlier than the forecast indicates, because there is very little snow on the ice (see below).  This is unusual for the time of year, and with no reflective snow pack, the sun will be unusually effective in melting the ice.  My best guess for breakup is between April 23 and April 28.


  1. The pic above is telling. Note the absence of snow on the south facing hillside, and the visible lack of snow cover on the river. All contribute to warming of the ice, especially for the adjacent Nenana River that joins the Tanana just below the tripod's location. Its inflow action can compromise the Tanana ice downstream of the marker and create instability of the main ice pack.

    Part of the lack of snow cover is due to the persistent easterly winds and occasional rain Nenana experienced this winter. Same for the barren terrain. Snow was meager as well.

    I saw about 32-36" of ice at 60 26 N, 150 50 W, with a similar lack of snow cover over the lake ice (6-12" max in places with some bare spots developing). The ice crust from winter rain is about 1/4", and when present, is slowing the sublimation and melting due to increasing insolation.

    I was able to maintain a water hole for 18 days just by covering it with a pile of snow and chipping away any new ice encroachment. Very atypical with temps that rarely rose above freezing, and approached 0 deg F nightly. The daily Sun is working on the surface ice temp.

    So, if it remains cold at night, breakup may be delayed. If it warms with wind and clear skies, breakup may be hastened, especially if the Tanana River upstream of Nenana starts to melt prematurely.

    Gary (not an Ice Pool gambler)

  2. Replies
    1. I don't have a ticket, but my parents were in Fairbanks in February and picked up a few tickets... so they and I will be watching carefully!

  3. I am wondering if the thin snow pack will assist in delaying the ice rotting out. When warm weather causes rapid snow melt and runoff into the rivers, it quickens breakup. Here in South Central the cold mornings have more than offset the (relatively) warm afternoons. Also, the abundant sunshine has enabled more sublimation of the snow – as opposed to melting.

    1. Yes good points Brian. The situation as of 4/2 as I flew over the Nenana area was as seen in the pic...locally wind blown ice and terrain devoid of snow.

      However that's a local Nenana condition due presumably to previous focused wind events and solar exposure on the adjacent hillside. Both upstream and downstream the river and terrain are snow covered. And now some light snow followed by cold next week are forecast.

      If the upper Tanana/Delta river melts, if the Wood River upstream and the Nenana just downstream flow, and if surface runoff from adjacent terrain starts running, it may enhance the ice breakup in Nenana.

      There are really two melt water sources in under and one (usually lesser in effect) potentially on top of the exposed ice. Local surface melt will stay on top until drain holes in the Tanana River upstream form, and sub-surface flow will be driven by increasing flow from tributaries and the main Tanana's sources. Downstream damming can create localized surface flow as well.

      An increase in either or both water sources can melt, structurally weaken, and lift the ice from the shore. If the downstream channel gets compromised from melt and flow, the whole pan lets go and someone(s) get $.


    2. Brian, good point, I agree a thin snow pack would favor a later breakup - the volume of melt water flow is really a critical variable. The question is what is the snow pack like over the drainage basin upstream. I wonder if the computer models have any grasp on this at all, or if it might be better estimated from total precipitation and solar radiation over the past several months. Earlier in the winter it was wet over the area, but of course some of that was rain.

    3. Might be some info here eventually:


      PS: The geese arrived yesterday in Fairbanks:

  4. Some easy reading references about ice on rivers. There's more but this will get the job done:


  5. And that's that: