Friday, July 14, 2017

Dietze Lab Internship

I’m Katie Ragosta, and this summer I’m interning with Professor Dietze and working on PEcAn, the Predictive Ecosystem Analyzer. While ecologists have collected a large amount of data, and are capable of collecting more with relative ease, that data is only useful for the kinds of large scale predictions that affect policy if we have a way to synthesize multiple data sets effectively, since no one data set will give anything close to a complete picture. Essentially, the Dietze lab aims to develop a model capable of integrating multiple data sources for more accurate predictions of the carbon cycle and biodiversity.
Screen Shot 2017-06-29 at 11.42.50 AM.png
I’m majoring in Physics and Mathematics here at BU. While it sounds like those majors might not fit perfectly with an environmentally focused project, they provide a lot of experience with systems modeling and programming, both of which are important in the Dietze lab. I took Computational Physics last semester, which gave me experience with Fortran and C. I also have a job in the physics department during the school year which uses a lot of C++, and I learned some Python in high school. The experience has definitely helped a lot with my work here, but if you don’t have much formal training and still want to work in the Dietze lab, there are lots of free resources online. The Dietze lab actually mostly uses R, which I learned some of in the first few days of the job, but what I do specifically doesn’t require much R.
The goal of PEcAn is to synthesize multiple data sources, so one of the ongoing projects in the Dietze lab is incorporating new models into ours. There’s a to-do list of pre-existing models that are being gradually added. My job is to add one called CABLE, or the Community Atmosphere Biosphere Exchange Model.
Screen Shot 2017-06-29 at 11.58.39 AM.png
The files for CABLE are written mostly in Fortran, which is why I was assigned this particular job. The CABLE website in that screenshot has some documentation on how you should go about building and running the model, so my first step was to read through that. Then, I got to work on trying to build it. There are two versions of CABLE, an offline version that uses its own input data and an online version that can use any input data. I’ve been focusing on the online so far. Getting the model to build involved a fairly large amount of debugging, and I actually just finished building the online version today.

Now I need to run the model as a test to make sure I built it correctly and see how it’s supposed to work. This is important because my next step will be converting CABLE to the format that PEcAn’s models are supposed to use, and we don’t want anything to get lost or changed in translation. I’ve been told that transitioning into PEcAn generally goes more smoothly than the initial stages of building the model, so I’m looking forward to that. When (or if) I get CABLE fully integrated into PEcAn, I’ll take another model from the model to-do list and repeat. Working in the Dietze lab has been a great experience, and I’m really glad to have this opportunity through BURECS.

No comments:

Post a Comment