Upside Down Trees: My Personal History with Computer Models
The incessant News-talk these days about the COVID simulation models has gotten me to think about the lessons I learned from writing computer programs and how they are helping me to understand and personally cope with the numbers Drs. Fauci and Birx report on a daily basis. During the 1970’s I learned to write simple computer models in FORTRAN, punched on cards and printed on green-bar machine-punched-perforated computer paper. It was during my initial set of exercises that I learned my first lesson about computer models: programs are largely If/then statements. These statements are what scientists call hypotheses. If ‘this’ happens then ´that’ is the outcome. So lesson #1: We all want absolute dates, numbers, predictions but scientists can only answer our questions with hypotheses.
My second lesson was that for a program to be worth its weight in punch cards, it needs to acknowledge a particular event almost never leads to only one outcome. Outcomes can be more or less likely. Computer simulations help solve this problem by incorporating probability into their program. If x is less than or equal to 5, then y occurs. If x is greater than 5 and less than 10, then z is the likely outcome. We have to tell the computer the likelihood that x is less than or greater than 5. We can do this simply by putting numbers in a hat. If all numbers are equally likely, then each gets its own slip of paper. That way all of the numbers have the same chance of being chosen. If one number is more likely (probable) than another, you simply put more slips of paper with that number in the hat. For example, x is twice as likely to be greater than 5, we simply put 2 slips of paper in for the numbers 6,7,8,9 and 10 into the hat.
The third lesson I learned is the one that most profoundly connects with our current individual, societal and global need to project the future of the viral pandemic. In the 1970’s I used a computer language called DYNAMO to simulate the ecological impact of the first farmers of some 6000 to 7000 years ago had on the temperate forests of Denmark. And here is the connection with COVID: I was studying a colonization process. In my archaeological case, people were colonizing forests. In the COVID case, viruses are colonizing human communities.
The results of my forest-farmer simulations baffled me and my fellow students. Under certain conditions, my model forest ended up with negative numbers of certain species of trees. Sometimes it was elm, sometimes linden, and sometimes as I recall, it was birch.
What on earth (pun intended) was a negative tree? Well, it was suggested to me as a playful critique that perhaps the trees were growing upside down? (I consequently became known as the upside-down-tree-guy). I solved this problem pretty easily by not answering the question: what is a negative tree? I simply inserted a new assumption into the model that the number of trees could not go below zero. I thought I was done, and luckily so did my advisory committee. But was I finished?
Forty years later, I realized that I was not. And I confronted this in a very unexpected place: a music festival sponsored by the band Wilco. The band-members of this group are known as being especially socially and ecologically conscious. So they hold their "Solid Sound" festival on the grounds and in the galleries of the Massachusetts Museum of Contemporary Art (MASS MOCA) in western Massachusetts to contribute to the economy of the region through visual art and music.
As my wife and I entered the Festival a few years ago, we were taken by a large environmental art installation that consisted of a line of live trees hanging from large pots on a line held between two telephone poles. We looked at each other and said: “Upside Down Trees!” (3). It was not until then, 40 years after I completed my studies that I understood that my solution to the problem of negative trees was incomplete.
My simple resolution for dealing with ‘negative trees’ was to put in an assumption that clearance and planting crops could not produce a negative number of trees. But the MASS MOCA experience actually challenged this resolution. No, it did not change the fact that a tree species population cannot go below zero. But recall, the first reaction to this mathematical impossibility, was to characterize the negative numbers of trees as “upside down” trees.
This was probably not of course a serious consideration for characterizing a danish forest, but it was not an impossibility. What I had thought was a full resolution of negative trees turns out to be partial. And here is why.
What I did not fully comprehend until I saw the art exhibit was that I was making the assumption that trees cannot grow upside down. Now is it possible for trees to grow upside down under certain conditions? Off of a cliff face perhaps? Or perhaps even underground in response to being colonized by prehistoric farmers? Not likely. But consider the Quaking Aspen (2). Known as the planet's largest organism, this Aspen adapts to its environment by growing underground to connect colonies of trees. Part of this huge tree is in fact growing down. The point here is not that any of this was likely for prehistoric Denmark. But perhaps it is pertinent to our understanding viruses multiply and colonize the planet. In point of fact, as the number of asymptomatic cases increases, the number of people with COVID immunity increases. So increased cases can lead to a diminishment of disease over the course of the viral spread. Are these immune people, "negative trees"? How do we write them into the computer model?
We have to continuously remind ourselves that a model is an imperfect representation of a real world system. In point of fact the Mass Moca installation empirically demonstrated that trees can grow upside down. For the purposes of my computer model I had to assume away a real possibility, something that could happen. I obviously had good reason to do so for my project. But it was not a good assumption for the artist. For her, upside down trees was both an empirical demonstration (because they were alive and growing) and a metaphor for the complex relationship between the environment and people. Again, are immune people conceptually upside down? For complex processes, like forests and viruses, we can never forget that we are always assuming something.
Nearly every day scientists do their best to give their best model projections of COVID cases and deaths. And every day journalists do their jobs to best report to the public the latest projections of the virus. They ask a lot of “how many” and “when” questions. Reporters want ‘facts’ about the future. But facts do not exist in the future. So scientists can only answer our questions with hypotheses. With If/then statements. This often leads to scientists and reporters becoming frustrated and a citizenry that just wants to how to conduct its everyday lives. But we all have to remember:
Upside-Down trees can mean different things to medical professionals, artists
and anthropologists.
Footnotes:
1. For the purpose of this essay, I am not bringing in the use of the model by politicians because of the ideological and partisan bias many of them bring in ways that offend scientific thinking.
2. Quaking aspens, weighing 13 million pounds, Pando is the world's largest organism by mass (Oregon's “humungous fungus” spans a greater distance). Quaking aspens can reproduce by disseminating seeds, but more frequently, they send up sprouts from their roots and form a mass of trees aptly known as a “clone.”
3. Tree Logic (1999) at MASS MoCA (in which six live trees are inverted and suspended from a truss, displaying the contrived growth responses of the trees over time).