Maximize

 "Let me swim against the tide...

Maximize me 

Let's go beyond reality

If you can see what I see

Hypnotize

Maximize me"

- "Maximize" by Amaranthe

Every time I have a few spare minutes, I pull out an old project and try to take it one step forward. It's the only way I ever manage to finish things. The project I'm trying to finish now is my "dock study," which some of you might remember from 2017. I spent the whole summer collecting data on fouling communities with my intern, Nicole, and we found some pretty cool stuff - bryozoan, hydroids, sponges. The project even had a surprise twist at the end when most of our animals died and were replaced by recruiting barnacles. I started the summer with all sorts of fancy hypotheses about how the species were going to interact in intricate, mechanistic ways, but by the end of the summer, I had come to the conclusion that succession in marine fouling communities happens in one of two ways. Sometimes, the first species to arrive wins. Other times, the best competitor wins. 

I was honestly pretty disappointed with my conclusions, because the whole system was a lot simpler than I expected. I had this nagging thought at the back of my mind that I was missing something. My data set was incredible - probably the cleanest, most well-balanced, solidly-grounded data set that I will ever produce in my career. I just couldn't let it go. 

During covid lockdown, I tried making a model to simulate succession in fouling communities. I thought that if I ran the model using different starting conditions - to match the conditions in my study - that I might learn something new about succession. Making the model was an incredibly valuable experience, but after consulting a statistician at WHOI, I realized that it was complete garbage. 

I met with the statistician, Andy, several times over the course of a year, and we came up with a new plan. He would build a mathematical model that included only very simple processes - a null model - and then we could compare it to my data. Honestly, one of the major challenges in this process was that Andy and I do not speak the same language. He speaks Numbers; I speak Species. Our meetings moved along super slowly because we had to constantly stop and explain to each other what we meant. 

I had a bit of a break-through when I went back to the scientific literature and read more papers on succession. There's this amazing paper from 2015 that lays out a key, unresolved question: why does the first species to arrive dominate a community sometimes but not others? As it turns out, the two situations I had observed in my study (the first species wins versus the best competitor wins) occur in a lot of ecological systems. There are theories about environmental and biological conditions that could push a system in one direction or the other, but so far, nobody's really tested those theories in the field. Meanwhile, here I was with the perfect dataset to test those predictions; I just didn't know it yet. 

Progress has been slow, but I am finally making some headway on the study. I have clear hypotheses. I have my data. I have my model - and the model is even giving me realistic, meaningful numbers! My task this week was running the model with 500 different scenarios and picking the ones with the maximum fit to my data. (For you statisticians out there, I was selecting parameters to maximize my log-likelihood values.) It's finally coming together - maximum model fit, maximum value for my study, maximum dominance of the community for some species. 

Maximum science.