I’m TAing a class right now. While most grad students loathe that they have to do this, I love it. I don’t mind the grading, the office hours, the e-mails, (the re-learning of things that I’ve long since forgotten so that I can adequately help undergrads who are learning it for the first time), and I love it when students ask me questions about my research.
Last week someone asked me what prepping for a field season actually means. It took me a while to answer. Ultimately, I told her it meant learning three new computer programs simultaneously, dreaming about statistics, and exercising every marketing guru trick that I know to get someone to listen to me talk about how important this whole project is…. and not just to me. That’s why I love my research; because somewhere at its core it is larger than myself (and not just because my study species is an 80,000 pound leviathan).
In retrospect, while I think that quip suffices to answer the question, I don’t think that it is tangibly helpful for aspiring field biologists (or aspiring grad students). So I thought, since prepping for our field season is in fact exactly what I’m doing these days I’d unpack the details of what that entails. For me it boils down to the following five things
- Software Mastery
- Logistics
- Theory and Questions
- Analysis
- Sampling Protocol
Software Mastery
If I could offer a piece of advice to undergrad, post-bacc students, or whomever may be choosing classes for the future I’d say this: take classes that teach you to do something. While it is obviously valuable to take classes that give you information, information is much easier to acquire than skill. I loved my behavioral ecology classes, but I NEED statistics.
Currently I am learning three computer programs. ArcGIS, Program R, and Ishmael. ArcGIS is a geographic information science software package that I am using to map humpback whale distribution around the lighthouse. It allows me to turn the numbers we so diligently recorded from the lighthouse tower into symbols on a map useful for analysis. While I am typically reluctant to celebrate digitizing nature I must say, seeing those little blue dots on a map of Alaska for the first time, and knowing those were real whales seen from a real lighthouse was so satisfying I danced a little. While a powerful program, ArcGIS is not too complicated for any computer savvy person to learn- particularly in a classroom setting. I’m learning it on my own, and having just mastered the basics, I’m impressed and excited about just what happens next. Now? I need to take that enormous cluster of points (nearly 2,000 in all) and reduce them down to something meaningful. I spend a lot of time working on this.
Why? By creating a visual representation of what we recorded at the lighthouse I can 1) begin to see spatial patterns in distribution based on several variables; 2)I can attempt to gauge the effectiveness of our sampling protocol (ie. are hour long intervals too long to capture behavioral shifts? Do patters of dispersion vary between survey sectors?); 3) I can run preliminary analysis on things like nearest neighbor distances (cluster analyses) to see if, as I hypothesize, something is actually happening when boats come through. My advice to anyone hoping to move forward in marine mammals- take a GIS course.
I addition to ArcGIS I rely heavily on program R for statistical analysis. R, in my opinion, is difficult to learn and difficult to use. However, its free, open to the public, and once you learn how to write code for it (that’s correct, it’s one of those code writing deals) it’s extremely flexible. I learned the basics of R in my statistics courses. This quarter I’m putting that information to work in an effort to learn a few things about my data.
While the purpose of the study is to determine what impact, if any, vessel noise has on humpback whale communication and social behavior, it becomes important to be able to tease out whether or not humpbacks are reacting to boats, or whether they are reacting to their environment. Are group sizes smaller midday because vessel traffic is heaviest? Or is that a function of diel variability in humpback behavior? To determine this a few things have to be done. First our sampling protocol needs to account for environmental variables. For example, I created a sampling protocol that attempted to control for tides, time of day, and vessel traffic. This means that the lighthouse team is often up at 4 o’clock in the morning doing those dawn surveys, and often hauling kayaks and skiffs up rocky intertidal zones at the lowest, and seemingly most inconvenient, of tides.
Now, 9 months later, I look at those variable statistically to see whether or not they had an impact on thing like nearest neighbor distance, group size, frequency of surface behavior, and group composition. If I find now that tides and time of day have no impact on these things: great. I can stop accounting for them in my sampling protocol and just focus on the meat of the matter: boats. If I find they do have an impact (which it’s likely they do), no problem. It just means that our analysis is richer and more complicated, like the whales themselves. Ecology is not a neat science. Nature is more complicated that a laboratory, thankfully.
Lastly, I’m learning Ishmael. Ishmael is a bioacoustics program developed by Dr. Dave Mellinger and the bioacoustics lab at the Center for Integrative Marine Resource Studies (CIMRS) at the Hatfield Marine Science Center in Newport, OR. While Ishmael is also fairly user friendly (and much friendlier when you have experienced mentors around to guide you), the physics behind marine bioacoustics are daunting and complex. So far teasing out sounds with Ishmael is going well. Placing them in the context North Pacific bathymetery…? I’ll get back to you once I’ve finished reading the stack of books I have on my bedside table about marine bioacoustics and communication. I feel confident, so long as my fingers stay crossed.
More to come….
