What Two Frozen Lakes Taught Me About Perception Versus Data

Ten Mile Lake in winter. Photo: Samantha Oliver

by Samantha Oliver

Today, I woke up at my favorite place on earth – Ten Mile Lake, in Hackensack, Minnesota. On the coldest day of the winter thus far, the frozen lake is so bright, so sparkly, it is almost hard to look at. And when you are from a place where 30 below is often the actual temperature (NOT the wind chill), you rely less on the calendar to define the seasons and more on observations of specific events that are tied to the climate.
Summer, to me, never began on the day with the most sunlight, but rather on the day the dock was installed (and ended, of course, when the dock was removed). Winter and spring were always announced by the day the lake froze and thawed.
Then I moved to the relatively tropical city of Madison, where the winter is shorter, but the lakes still provide a checkpoint in our urban phenology. But now my observations are set up as a study of contrasts: my asparagus came up two weeks sooner than my mom’s or simply subtract 10 degrees from Madison’s temperature to guess Hackensack’s temperature. On December 15th this year, I texted my mom: “Is Ten Mile frozen?” “Almost”, she responded.
We have this conversation every year, but this year it felt different. I asked because Lake Mendota seemed close to freezing. I couldn’t remember a time when the two lakes froze within a week of each other. Mendota always seemed to be frozen a month less than Ten Mile. But lakes are losing ice cover1 and colder places are warming faster. Maybe, I thought, the two lakes’ ice seasons are converging because of climate changes. Was my hometown becoming more like my new town?
Lake Mendota on December 15th, 2016. Photo: Samantha Oliver

I was hesitant to believe my own conclusions because human perception of climate change is complicated. Our beliefs are influenced by our experiences1 – if you can remember ice fishing in November as a kid, but not as an adult, you might feel certain that climate change is affecting ice duration. But, if you already believe climate change is not real, you might instead remember the polar vortex of 2014, and conclude that the climate fluctuates and warming isn’t happening2.
But these perceptions of change cannot protect us from the observable fact that we are living in a warmer and more extreme climate. And that fact has consequences. For example, the data tell us that Lake Mendota has lost a month of ice cover in the last 150 years3.
I wanted to move beyond my experience and perception – what do the data say?
Ten Mile Lake doesn’t have such a long record, but it does have dedicated citizen scientists who have tracked ice on and ice off dates since 1988. I made plots to compare the ice on, ice off, and ice duration of Ten Mile and Mendota to see if my own perception matched reality.
(1) Ice on (top), off (middle) and duration (bottom) records for Ten Mile Lake (blue) and Lake Mendota (red). The difference between the two lakes (in days) is in the right panel, with the 5-year rolling mean shown in the dashed red line. Data came from the North Temperate Lakes Long Term Ecological Research program and the Ten Mile Lake Association. Graphic courtesy: Samantha Oliver

The data both validated and disproved some of the conclusions I made from my personal experiences. Or, to put it in less scientific phrasing, the answer to my question was both “yes and no.”
First, I learned that the two lakes can and have frozen on the same date. But, it’s rare, and this year Mendota did not succumb to the cold snap in mid-December, freezing instead on January 1st, 15 days after Ten Mile and consistent with the record average of 16 days. The data show that the freeze date for both lakes appears to be getting later, but the difference in freeze dates is highly variable and not declining.
Where the lakes might be getting more similar is ice off date. While my memory of the two lakes being “two weeks” apart held up for ice on, it turns out that ice off difference are much larger (29 days apart on average), and I was shocked to learn the difference in total ice duration between the lakes can be as high as two months!
In my childhood, the lakes frequently thawed more than 40 days apart. But that large of a difference hasn’t been documented in the past 15 years and the 5-year average difference in ice off dates was lower in the second half of the record. Given the small differences and relatively short time frame of those observations, I guess I’ll have to patiently wait for the next 30 years of data to conclude if these ice off dates are converging.
Humans can sense change, but complex cause-and-effect processes, combined with change that happens slowly, are difficult to perceive. Even as a limnologist, my perceptions based on observing two lakes I know well did not match the reality of the data. This shouldn’t be too surprising, I guess, as Lake Mendota has already taught us that 30 years of data may not be enough to reveal the slow, variable, but very real changes taking place around us.
Lake Mendtoa in winter. Photo: A. Hinterthuer

1Magnuson, J.J., Robertson, D.M., Benson, B.J., Wynne, R.H., Livingstone, D.M., Arai, T., Assel, R.A., Barry, R.G., Card, V., Kuusisto, E. and Granin, N.G., 2000. Historical trends in lake and river ice cover in the Northern Hemisphere. Science289(5485), pp.1743-1746.
2Myers, T.A., Maibach, E.W., Roser-Renouf, C., Akerlof, K. and Leiserowitz, A.A., 2013. The relationship between personal experience and belief in the reality of global warming. Nature Climate Change3(4), pp.343-347.
3Howe, P.D. and Leiserowitz, A., 2013. Who remembers a hot summer or a cold winter? The asymmetric effect of beliefs about global warming on perceptions of local climate conditions in the US. Global environmental change23(6), pp.1488-1500.