April 6, 2009
The unexamined conservation genetics lab is not worth leading. It's time to start thinking about the footprint of science, especially in labs that ostensibly are working toward conservation and an appreciation of the processes that generate and maintain biodiversity. I've brought my Kill-A-Watt into work to start evaluating ways to make my lab more efficient, with a smaller carbon footprint. I pay for carbon offsets, but that is just a financial trick to help me sleep at night (it doesn't work very well; I'm an insomniac!). This week I'm starting with the old chromatography refrigerator I inherited when I took over the lab. These giant glass-doored refrigerators can be useful, but we aren't running allozymes or SSCP gels in there, and really it may be that we don't need the thing at all. I've got it hooked into the Kill-A-Watt and will report on its cost (in dollars and in carbon molecules) soon. My goal is to reduce the kWh of my lab by 10% this month...I'm not sure if that is possible, so let's start crunching numbers. I'm working on the ballpark rate assumption of $0.10 per kWh, which may be a little high for Georgia but keeps the math easy (my friend Jason in Engineering Outreach tells me now that UGA's cost per kWh is $0.065).
Given that, after 24 hours that old chromo fridge used 5.49 kWh - about 1/100 of a typical monthly household energy use. That is actually surprisingly low, I think - that means it would probably cost about $197 a year to just leave it on. It isn't like there is an entire graduate fellowship lurking in that outlet. But, of course, add these things up across campus, or just across a lab...and that is a few pounds of CO2 (about 1.7 lbs CO2-equivalent per kWh). Now I'm going to check out my 'tip incubator', an old drying oven that we use just to make sure our pipet tips are dry - but that sits around heating the air ever so slightly, constantly. More on that tomorrow.
April 7, 2009
Okay, the incubator/oven set at lowest temperature point (holding at 29° in a lab that is probably about 24° ambient) - looks like that costs $28/year, about a sixth of the chromatography fridge (for a much smaller gizmo, that is also not really necessary for my lab). To establish some benchmarks for energy reduction, I guess I need to estimate kWh for a given time period, large enough to tell differences among the gizmos. The fridge used 5.49kWh in just over 24 hours, or about 2000kWh in a year. The oven on the other hand will use about 260kWh in a year (I didn't leave the K-a-W on for very long, so this is obviously ballpark).
A nice reference will be the one thing that I *know* I need for productivity from my lab - my computer. I have a Mac G5 with an LCD monitor, both about 4 years old. They are plugged into an uninterruptable power supply (UPS - fortunately, so I didn't have to shut anything down to plug in the K-a-W). We'll see just how much damage this setup does (to the environment, I mean - my eyes and wrists are certainly getting tired).
April 8, 2009
3.44kWh so far on my computer, in 16 hours. That is dismaying. For all intents and purposes it is almost equivalent to that giant refrigerator, despite having an LCD monitor that shuts off all night. But the computer doesn't - it is often running processes (number-crunching), or I access it from home, and for a few other reasons I leave it on for weeks at a time. There are certainly plenty of scheduling scripts that could be set up to shut it down each night, at least for a few hours, but I'm not sure that wouldn't mess with my (admittedly not high) productivity. I'm going to leave the K-a-W on for another day so it picks up a full day of me working on figures for the Notochthamalus paper, playing Guest Editor for Estuaries & Coasts, and of course checking si.com for Braves stories.
The trade-off may be that my lab currently has 8 computers - more than one per student. A couple are devoted to tasks, like the Gel Doc machine that is a real clunker. So I can't get rid of those, but that one in particular could be put on a shut-down schedule (except it is Windows, so that probably wouldn't work...). And another old machine is only serving information, that function could be transferred to another machine. Lots of options, for now I'll just keep crunching numbers.
Update: I found a python script that was able to rip the MySQL data from the old server G4 into HTML format. This is really dry reading for you, but I know someday at least *I* may find it useful. Anyway, the old G4 and its monitor are now shut down and unplugged from the wall, and I can access all that info (even with pictures) here on my main desktop. If you take nothing else away from reading this wiki-blog, its that learning scripting languages like perl and python is really, really useful in this information age. So, I'm going to assume that whatever kWh load (about 1980 kWh/year) my G5 puts on the world, I just saved by (hopefully permanently) unplugging my 9-year-old G4.
April 13, 2009
The weather outside is gloomy, but through the fog - is that the Flying Spaghetti Monster? No, just a Monday looming. Final acceptance on Clare Scott's honors thesis came today, it is now "in press" at Ecology of Freshwater Fishes, good work Clare! This time of year seems to be manuscript-juggling: two dissertations, our paper at MEPS, a project starting up for J. Biogeog., and two Estuaries & Coasts manuscripts in revision. That means I need to be reminded every five minutes which project I'm working on.
As for the Kill-A-Watt, the most recent info comes from my printer - which is a Lexmark that can print on both sides, so already a green step-up from our older laser printer. After 20 hours it had consumed 0.4 kWh, costing about $20/year to just leave it on. Certainly I could imagine figuring out a way to put it on a kill switch at night though, could cut that price in half. I'll add up those things later on in this project. The other thing that I measured over the weekend was my bank of MJ thermal cyclers. Obviously these are the bread-and-butter of the lab, so their usage is not really that negotiable. But, sadly, their job is to cycle back and forth between heating things and cooling things well past ambient, and they aren't insulated (they need to be able to change temperature rapidly, after all). So, 3 thermal cyclers under more-or-less normal usage for 94 hours used 8.32 kWh...about 775 kWh a year at a cost of about $77 (by my $0.10 guesstimate, which is a bit high - see post from a few days ago - but keeps math easy).
Now we are starting to get into the not-negotiable power consumption. For one, my ultracold freezer. That is necessary for long-term storage of valuable (hard to get again) DNA samples, and some enzymes. It churns pretty much constantly to keep the interior of the freezer at -80°C, and it is big - about 480L interior volume, and that is about 80% full (probably could be cleaned out and be 50% full). It runs 220V so I can't use the K-a-W, but online estimates from manufacturers suggest it is going to use 16.3 kWh per day, or about 6000kWh/year. Obviously these ultracolds are a big part of our footprint. The only way I can think of, off the bat, to deal with this is that there is a "departmental emergency" freezer in the same hall, that is almost empty and also running at -80°C. Maybe we need to negotiate a different arrangement? But, then, we've all had to deal with the crisis of a freezer that needs repair. At that time you need to move your irreplaceable samples immediately so they don't thaw, and so probably cannot wait for a freezer to get to temperature. At least, that is our fear talking...
SO the current tally: chromo fridge (now unplugged), incubator (now unplugged), G5, G4 (now unplugged), printer, ultracold, thermalcyclers used about 13000kWh/year. By unplugging the unnecessary stuff (for now), I'm helping...but there are a lot of gizmos to go. Assume the other 6 computers in the lab are also about 1800kWh/year - that is 10800kWh/year, a pretty massive load. So how do we cut back on computer load, without affecting computational power and analytical needs? That's the next thing to tackle. It will be, uh, a little easier once my two students finish up this summer...another G4 can be retired, for sure. But it is also important to figure out how some of these machines can be put on SHUTDOWN schedules, while still maintaining appropriate access to some data/servers at night.
Also, I sent this email today:
To whom it may concern,
every week I receive an issue of Columns, specifically addressed to me. I do find the information in Columns of value when I have time to read it, but I've noticed that dozens of copies in my building alone are immediately recycled or thrown away. Would it be possible for the University community to receive Columns as a formatted HTML email? Or to unsubscribe from the paper copy if they do not intend to read it? I notice in addition to the copies placed in each of our mailboxes (addressed to us, so our staff takes the time to sort them into mailboxes), there is a rack with extra copies in our mailroom.
At a time when the University is facing enormous financial burdens, changing the delivery default of Columns could be an enormous financial turnaround, not to mention an environmental shift that is in line with President Adams' stated goals. I'll look forward to your thoughts on this matter.
There are lots of stones left to turn over at this University. If you are interested, check out http://gogreen.uga.edu/
April 16, 2009
Monitors are a big part of the power suckage of computers, but I'm surprised it isn't worse. I hooked the K-a-W up to the 24" LCD on our Mac Pro for a couple of days; looks like it is responsible for about 260kWh/year, or about $26 in operating costs (and about 445 lbs CO2e). So that leaves me pondering why the computer itself is so power-hungry at night? Well, it isn't set to go to sleep. So, whether I'm typing on it or not is immaterial. But I hesitate to set it to actually go to sleep, since that can grind analysis to a halt (this week for example I had a long PAML run and a long IMa run going; plus, I contact my G5 from home on occasion). There are ways to script a UNIX-based machine (like Mac OS) to check its processor load before going to sleep, but I'll need to do more research before applying this to my machine (for one, that link is 7 years old - OS X has changed a bit!). There are funny solutions to this problem, including the application Jiggler, which apparently mimics the mouse being jiggled often enough that the system won't go to sleep. This seems to be more for people who can't control the sleep/no sleep function, i.e. they aren't administrators. I had this problem with a Windoze machine while at UNM, I could *not* get it to stop going to sleep during a long analysis - so I put its mouse on a shaker table!
Anyway the trick for some of our machines may be to tell them explicitly to go to sleep - and warn folks not to use those machines for extensive analysis. The trick is, as haphazard as some of the development of my lab has been, it is hard to make major cuts without more control over the environment of the lab. Can't control the heat/air. The fume hood is OFF about 99.99% of the time because we don't use nasty chemicals - we just culture algae under a gro-lite in there. That's good because constantly-running fume hoods have to move so much air, they can be as energetically expensive as three U.S. households!
What I've cut so far - about 4500kWh/year - is about 7500 pounds of CO2 equivalent each year, and about $450. If you don't count heating and cooling but just the stuff we plug in, there is still another 12000 (computers), 6000 (ultracold), 6000 (other freezers and refrigerators), and probably another 6000 overall in other gizmos, or about 30,000kWh/year of energy use. So, for now I've met the 10% goal in a very coarse way. Many of us are able to ride our bikes in most days (I ride in at least 300 days a year)...more later...
April 20, 2009
As Earth Day approaches, continuing the energy-saving theme of this page seems good. At one faculty retreat a couple years ago, we began to discuss ways in which our department or building could reduce our footprint a bit. One thing that came up was a rant against using the elevator - at least, if you are reasonably healthy and not going too far or carrying too much, right? Well, I just came across this post that explains that elevators are one of your smaller concerns in a day's business. On to the next problem.
For now I'm leaving aside the consumption of kilowatt-hours by my lab; most remaining gizmos are somewhat necessary or infrequently used (and off/unplugged). The waste stream is probably the next good place to start. We try to maintain a very low-toxicity lab: no ethidium bromide (we use GelRed for staining gels), no radiation, extremely infrequent use of polyacrylamide or TEMED, and so on. Probably about 90% of our chemical usage is water, common salts, and naturally-derived enzymes in small volume. Our chemical usage could still be tweaked, but by and large mostly benign.
Plastics bugs me. Lots of virgin polypropylene. I estimated about a year ago, based on purchasing for the lab, that the common yellow tips used on 20µl and 200µl pipettors added up to about 4kg a year - the equivalent of 420 20-ounce soda bottles (at 28g each). Given the soda consumption of people in my lab (I'm just as guilty), I suspect we also drink and recycle about that many bottles a year. Which is worse? Washing and reusing the tips would definitely take a cut into our productivity (and would probably cause my students to mutiny), and recycling them is almost impossible (I think polypropylene is #5 plastic, which is not easily recyclable - and of course even if we think that our tips are only touching benign substances, would you buy a teething ring for your kid that advertised it was made of recycled scientific labware?). But, then, cutting out soda might put a notch in our productivity too, given the admitted caffeine addiction we've all got.
Wow, I didn't solve any problems with this information - bummer.
May 15, 2009
Wow, the end of a long semester and the beginning of summer. It is good to have a bit more leisure to consider how my projects are doing; just had my (almost) global phylogeny study on Chthamalus published in PLoS ONE, and today finishing up revisions on a Mark Vellend-inspired paper on genetic diversity in the Georgia Coastal Ecosystems domain.
I'm continuing the assault on electricity usage by my lab; right now our temperature-controlled room is not harboring any experiments or anything living (intentionally). Usually this 200-square-foot space is cooled to 18°C; it usually breaks about once a year because of the load of the first really hot days on the cooling system. If we aren't using the space, why not turn it off? That's gotta be worth something.
The question is, how does this energy usage affect the science? One could argue that my scientific program hasn't expanded to fill the resources I had available, or that I can only cut back in inverse proportion to my productivity. Only time will tell! We may have to devise a metric for productivity per kWh, perhaps the Hirsch index per kWh? Right now I'd be at about 0.00043 - no matter how it compares with others, that is a pretty unimpressive looking stat. I better get back to work.