The following post is excerpted from an email conversation with my friend Linda Hasenmyer.

Hi Linda,

So, for the last year or so I’ve been on this Quantified Self kick. One of the things I do is wear a calorie tracker gadget which measures, among other things, the calories I burn on a minute-by-minute basis. Every few days or weeks I check my data and try to see what I can learn, and try to adjust my behavior accordingly.

Today I was reviewing my data and remembered the conversation we had last week about calorie burn with regard to rowing and snow shoveling, so I thought I’d share with you what I found.

On the day we spoke (March 5), I went out and shoveled snow for one hour, from 4 to 5pm. Here is what my calorie burn looked like for the full range of that afternoon from noon to midnight (from 12:10 to 4pm I wasn’t wearing the device, which is why it looks like a flatline for that timeframe):

The primary metric being tracked in this chart is calories per minute. The armband also tracks another metric called a MET (or “Metabolic Equivalent of Task”), which measures your rate of calorie burn. According to Wikipedia, one MET is roughly equivalent to the rate of calorie burn of a healthy, 154-pound, 40 year old male sitting on a couch staring off into space (so basically 1 MET is the rate at which someone burns calories by simply having a pulse).

Moderate activity (e.g. walking around) usually burns roughly 3 METs, and vigorous physical activity (e.g. heavy exercise) is typically considered anything over 6 METs.

From previous tests I know that I personally burn somewhere around 1.2-1.3 METs by existing, which makes sense because I’m younger than the control group that originally established the metric (i.e. I probably have a slightly faster metabolism).

On an unrelated note, another thing I find interesting about the chart above is that you can see where I got on the computer a little before 8pm and then the spikes where got up to brew tea at 8:40 and refill my cup at 9:30, 10:10, and 10:40pm, before taking off the device at around 11. It’s amazing how much of a spike can happen just by getting up for a minute or two! It really makes me think about the negative impact that lots of sitting can have.

Okay, back to snow shoveling. Here is a closeup of the snow shoveling time range:

On the chart above, you can see the effect of shoveling (4-5pm) on my rate of calorie burn. Over that period:

• Average METs: 3.7
• Calories burned: 242

So what does this tell me?

• As expected, while shoveling I burn more calories than when I am just walking around, but fewer than when I am exercising in earnest
• Shoveling is more of an endurance task than a high-intensity task (I maxed out at 5 METs at my most intense, which does not break the 6 MET threshold into high-intensity exercise)
• I took mini breaks at the 10-minute mark (*embarrassment*) and 37-minute mark
• I definitely tapered off in the last 20 minutes or so, with a last push in the last few minutes before 5pm

This strikes me as accurate, as I don’t recall being out of breath much, or having too much of an elevated heart rate. I think I went easy on myself (probably too easy).

Effects of Cold

One of the things we also talked about the other day was the question of whether the cold temperature would help burn additional calories (interesting article on that theory). The temperature that hour was right around 31 F and wind speed was negligible (~12mph). So it was moderately cold.

Sadly I don’t have a proper control to compare this against. Also, I may have partially or fully negated the effects of the cold by wearing three layers of clothing, a hat and a hood.

How could I better test this? Purely speculating here: potentially I could try shoveling sand for an hour in the spring when it’s warmer. The remaining problem then is how do I know that I’m putting in the same amount of effort for each trial? One way to mitigate that is if I could do several trials of snow shoveling and several trials of sand shoveling so I can compare the averages of the two groups. I’ll have to hope for another snowfall (the rain last night has done away with much of the snow from last week).

But just for fun, let’s compare snow shoveling to an actual workout. On a different day I did a moderately high-intensity cardio workout for a little over an hour. Here were my results:

The cardio lasted from 5:15 (began with stretching for about 15 minutes), until 6:25 (the last 5 minutes comprised of more stretching). If we take just the time range from 5:15 to 6:15 (so that the time range is 1 hour to match the snow shoveling data sample), we get:

• Average METs: 5.7
• Calories burned: 382

I suppose it makes sense that a targeted workout like this would burn about 58% more calories than something less targeted like snow shoveling. I was hoping the cold would have a clear effect for shoveling, but we can’t know what impact it had because of the factors previously mentioned.

Nick

Hi Nick,

Very intriguing!

Would one shovel of sand be heavier than most snows? One would have to weigh the amount in the shovel. Did you pick up the shovel or just slide the shovel full of snow to the side? Had you looked at the data on snow shoveling prior to exercising or were they independent of each other? Which makes me wonder what muscle groups are not used during shoveling that you used during exercise. And what would make the danger of snow shovel over regular workout since you burned more calories then can we assume you worked hard creating more stress? Inquiring minds want to know. How does the MET change as you grow older or have a larger body mass?

Linda

Hi Linda,

Those are really good questions. I did a little digging (pardon the pun) and combined it with some wild speculation (until I can actually test it):

Sand vs. Snow

I wonder if I can compensate for the relative heaviness of the two media (snow or sand) with a differently sized shovel to ensure better consistency.

Here’s some dubious research on the weight of 1 cubic foot of each material:

• Dry Sand: 100 lbs
• Snow:
  • Light & Fluffy: 7lbs
  • Average: 15 lbs
  • Compacted/wet: 20+ lbs

In an actual experiment I would definitely want to weigh each material.

I think the snow I was shoveling the other day was fairly average (although it was very fresh). Let’s assume 15 lbs per cubic foot. That means that sand is about 6.67 times heavier than snow. Good call.

Expectations

I didn’t necessarily have expectations going into the snow shoveling. The cardio exercise I was referencing happened on January 9, so my results were not fresh in my memory as I was shoveling on March 5. The trials were pretty independent.

Method

I probably did not have a terribly consistent method for picking up the snow. I would generally bend at the knees, push the shovel forward to get the snow on the shovel, and then straighten my legs to fully stand up while lifting it. I would then throw the snow to the side, which used a slight turn at the waist, but mostly from arm movement (I was being careful about my back).

As I got further along, I modified the last bit of my method. Instead of flinging the snow, I began to just lift it and then walk over to the side of the driveway and turn the shovel over to dump the snow straight down onto the banks.

Here is some speculation about the muscle groups used in snow shoveling:

• Snow shoveling:
  • Squat: Quads, Adductors, Hamstrings, Spinal Erectors, Glutes, Abdominals, Obliques
  • Holding the shovel and throwing the snow: Biceps, Triceps

The cardio exercise was a video workout called Kenpo, which is a faux-martial arts routine. So lots of kicking, punching, blocking, and some balancing:

• Kenpo:
  • Quads, Adductors, Hamstrings, Glutes, Abdominals, Obliques, Biceps, Tricepts, Lats, Deltoids

I tend to feel most sore in my Lats and Deltoids the next day after Kenpo. That’s probably because there are a lot of different punches right at the beginning of the video (after stretching), so I put the very most energy into those before I start to fade.

As far as safety, I imagine that Kenpo is safer because it is more controlled and involves no weights. Shoveling can make it easy to hurt your back. I think METs are a pretty good proxy for intensity, but controlled intensity is probably safer than something less intense but less predictable.

METs and Age

As someone ages their Resting Metabolic Rate (i.e. 1 MET) goes down, meaning that their typical MET value would be less than 1 when they’re sitting around. Here is relevant info from the Wikipedia BMR article (BMR approximately = RMR):

BMR generally decreases with age and with the decrease in lean body mass (as may happen with aging). Increasing muscle mass increases BMR, although the effect is not significant enough to act as a weight-loss method.

A person’s metabolism varies with their physical condition and activity. Weight training can have a longer impact on metabolism than aerobic training, but there are no known mathematical formulas that can exactly predict the length and duration of a raised metabolism from trophic changes with anabolic neuromuscular training.

A decrease in food intake can lower the metabolic rate as the body tries to conserve energy. Researcher Gary Foster, Ph.D., estimates that a very low calorie diet of fewer than 800 calories a day would reduce the metabolic rate by more than 10 percent.[25]

The basal metabolic rate varies between individuals. Statistically, the researchers calculated that 62.3% of this variation was explained by differences in fat free mass. Other factors explaining the variation included fat mass (6.7%), age (1.7%), and experimental error including within-subject difference (2%). The rest of the variation (26.7%) was unexplained. This remaining difference was not explained by sex nor by differing tissue sized of highly energetic organs such as the brain.[9]

Thus there are differences in BMR even when comparing two subjects with the same lean body mass. The top 5% of people are metabolizing energy 28-32% faster than individuals with the lowest 5% BMR.[10]

I think ultimately the utility of METs is to compare your own relative energy output for different activities over time. If you were to compare your METs with someone else’s, that’s where issues start to occur.

A few days later:

Comparative Output

So I’ve actually measured the Kenpo workout three times now (most recently today). The duration for each workout varies even though it’s a video because I tend to pause the video when I need a break. But here are the comparisons:

1/9/2013 5:15 – 6:25pm

2/27/2013 6-7:30pm

3/10/2013 4:56-6:15pm (looks like the system is having a problem adjusting to daylight savings today, so everything appears shifted an hour)

Summary Table:

I thought I went fairly all-out today, but I did also feel like I started to lose steam in the middle because I was testing out a new app that measures heart rate, but had to stop moving for 60-90 seconds in order to get the reading, which extended my breaks and undoubtedly had an impact on my calorie burn. It’s hard to draw a lot of conclusions from just three data points, but it looks like shorter workouts, although they burn fewer calories, tend to be more intense in terms of average METs (which is really more what I’m after). I’m disappointed I can’t quite seem to reach 10 Calories/min even during the most intense parts of the workout. I may need to adjust my approach.

Nick

Found the perfect way to shovel the drive, pay the guy with a snow blower! Best $30 I have ever spent!

Linda

Touché!

Nick