# She lives, she breathes, she is beautiful

“What’s new?” people like to ask me about my project. “How’s the research?” they wonder. “Are the birds safe yet?”

When we last spoke, the model was basic. Like, silly basic. It had one isolated group of birds that lived in paradise and could breed all day and night until there were literally so many that my computer refused to count them — something like 10^248 of them after a couple hundred years. There were no other populations for them to interact/mix with, and the havoc that mercury laid on their eggs stayed at a constant level throughout time. I spent a good day looking at the eigenvalues and eigenvectors of the system (rate of growth and population levels), but other than that it wasn’t a very useful model.

A simulation with two environments: top is contaminated, bottom is uncontaminated. Left graph shows population levels, right graph shows the proportion of individuals resistant to mercury. This simulation has 10% of individuals leaving the uncontaminated patch every year to go to the other patch, which is why there are still a lot of non-resistant birds in that patch. I need to fix the way the model makes birds disperse because right now it’s unrealistic.

Now the model actually resembles real life a little bit more. Not a whole lot more, but a little bit more. For one thing, there are now two populations in the simulation; one that lives with mercury that reduces the probability that their eggs will survive to fledge, and another that lives totally free of mercury. This in itself can be fun to play around; when do these populations behave the same way? When do they behave differently? Predictably, the population living with mercury tends to shrink for a while until the population is made up mostly of individuals with genes to tolerate it, then it reaches a tipping point and starts growing again. That’s really cool! It introduces questions, like how can we describe that tipping point? How long and how quickly does the population shrink? But you get more interesting questions when these populations start to mix with each other. Birds sometimes move breeding territories; what happens when a normal bird settles down in the mercury contaminated environment, or if a mercury contaminated bird settles down in the normal environment, and they start breeding and having offspring that are not equipped for living in there? That’s kinda the million dollar question.

The model also features density-dependent growth. That’s a fancy ecology term for “the population stops growing if it gets too big”. The birds reach “carrying capacity” now, meaning that there are only a limited number of spots for birds to breed, so when they are all full the population stops growing. So here’s what probably happened on the South River: populations were at this “carrying capacity” but then mercury was introduced to the environment by a coal plant. So the birds along the river started having fewer kids, and the population may have declined for a bit while natural selection guided the population towards being more tolerant of mercury. Meanwhile, birds from surrounding, clean habitats took advantage of the vacancies along the river and started moving in there and spreading non-resistant genes throughout the population. Will tolerance ever overtake the population? What is the new carrying capacity going to be? What happens when the mercury slowly disappears? This is where I’m going next with these questions. First, I’m working with professors here who do GIS work to get a big map that I can run simulations on, so instead of having this naive two-habitat model I can have many realistically sized and spaced habitats and play around with movement between these patches. I also have to refine the way birds in the model move between patches, because that stuff is complicated and hard to write up on a computer. Then I’m going to play around with varying the effects of mercury over the course of the simulation, to simulate the accumulation of and slow clean up of the mercury in the environment. This will make simulations more realistic and give us more predictive power.