What I actually do…

by Aaron · Published July 10, 2012 · Updated November 16, 2012

The meme is a little outdated, but oh well. The time has come, this week is my last week in Ithaca, and Friday is my last day at work. So, I figured I would devote a blog post or two to explaining what I’ve been doing for the last year. I’m doing this because when I explain what I do to friends and family I’m pretty vague, usually along the lines of just “climate research” or just “research”. If anyone wanted to know more I would usually make a very bad attempt at elaborating and end with something like “I work with computers a lot” …probably the understatement of the year.

Bluefire Supercomputer! (from ncar.ucar.edu)

I work with computers all day everyday and one really important computer in particular, Bluefire. Bluefire is a supercomputer at the National Center for Atmospheric Research (NCAR) in Boulder, CO. It is powered by 4,096, 4.7 GHz (dual core) processors, about 12 terabytes of memory, and hooked up to a storage system that can hold more than 12 petabytes of data (1 petabyte = 1,048,576 gigabytes). To give you an idea of how powerful that is the highest end desktop computer I could find online was a $10,000 Apple Mac Pro with 2 (6-core) 3.06 GHz processors and 64 GB of RAM, and Bluefire is over four years old! Only this year was it knocked off the list of the top 500 supercomputers, and if you think that’s impressive, Bluefire’s successor, Yellowstone, will be 29 times more powerful than Bluefire (if you want to know more watch this SciShow video)!

So, what did I do with access to this computing behemoth? I ran simulations of the Earth…a lot of simulations. Over the course of the last year I must have run at least 300 years worth of simulations. I did this using the Community Earth System Model (CESM), which is actually several computer models (of the land, atmosphere, ocean, etc) that work in tandem to simulate the Earth. The CESM is an incredible complex model, and I got intimately acquainted with its inner workings as I tried to port (basically install on a new computer) it to a computer system here at Cornell. That venture ended in a stalemate…I got everything working that I could, but the computer system just wasn’t powerful enough to handle all of the features of such a resource intensive program.

Dust off of Western African from a NASA satellite
(By NASA’s Earth Observatory)

Random cool looking graphic!

After the semi-successful (I consider it successful anyway) project of porting the CESM I went back to working on Bluefire. I ran simulations of the present day climate as well as a couple paleoclimate simulations (climates of the past, one of which will be the subject of another blog post). The purpose of these simulations is to improve our understanding of mineral dust in the atmosphere, including its distribution and its effects on the Earth’s energy balance. Dust can affect clouds and other atmospheric processes, as well as ocean and land ecosystems through the movement of nutrients around the globe. Humans can affect dust through land use and the effects of climate change. The work I’ve been doing is part of a larger overarching goal of the research group I’m in, which is to better understand dust’s (and other aerosol’s) variability and their role in the climate system. By increasing our understanding of these physical processes we can apply that knowledge to better understand past, current, and future climates.

Since my work has been wrapping up in the last couple months I’ve been spending more time looking at output data, and making cool looking graphics. After all I have at least 10 TB of data from all the simulations I ran. Before getting to work with data I felt more like a computer science major than an atmospheric scientist. I made changes to the model settings, and then shifted through FORTRAN code to try and figure out what I broke in the model when changed said settings. Not that I’m complaining, gaining more experience with programming and models is one of the reasons I joined Dr. Mahowald’s group in the first place. I didn’t take as many CS courses as an undergrad as a probably should have, and there was a period of time where I even avoided it because I assumed I wouldn’t like it. Not so much the work itself, but the part where I had to sit in front of a computer all day. After my experience in the Student Airborne Research Program I was more interested in field campaigns. It turns out I like modeling a lot more than I thought I would. I found it challenging and fun, even if it was extremely frustrating at times. I can’t thank Dr. Mahowald enough for giving me the opportunity to learn and explore an area of atmospheric science I might not have actively pursued otherwise. After this experience I’m definitely considering a PhD in a similar research area following my MESPOM degree.

My desk in Bradfield Hall. There is an organizational system here…I swear!

If you want to know more about what I do, post questions in the comments below and I’ll do my best to answer them. My next post will be less about my work and more about a paper I read to prepare for some of the paleoclimate modeling I did on a period in time called the African Humid Period.