My watch said I burned 612 calories on my Tuesday morning run. The temperature was 102°F. The humidity was 78%. I was soaked through my shirt by mile two, and my heart rate was sitting at 178 bpm — a number I normally only see during sprint intervals. I looked at the 612-calorie reading and laughed. Not the good kind of laugh. The kind of laugh you make when your $400 fitness tracker treats a heat-stress response like a personal best.
I ran the numbers. That is what I do. And the numbers were a mess.
Here is the thing about exercise physiology in extreme heat: your body is not working harder. It is working differently. When the ambient temperature climbs above 95°F, your cardiovascular system shifts resources away from oxygen delivery to the muscles and toward thermoregulation. Blood flow redirects to the skin. Sweat production ramps up. Your heart rate elevates to maintain cardiac output despite reduced venous return. Your watch sees a higher heart rate and assumes you are crushing a tempo run. You are not. You are just hot.
I spent three weeks in July — the worst month in Austin, which is saying something — collecting data. Same route. Same pace. Same time of day. The only variable was the temperature. I ran the Butler Trail loop at 6:30 AM, when the temps ranged from 78°F to 97°F depending on whether the sun had decided to punish us yet. My pace was locked at 8:30 per mile. My watch was a Garmin Forerunner 965. Here is what the spreadsheet looked like:
| Temperature | Heart Rate (avg) | Watch Calories | Perceived Effort (1-10) |
|---|---|---|---|
| 78°F | 152 bpm | 487 | 5 |
| 85°F | 161 bpm | 534 | 6 |
| 92°F | 171 bpm | 589 | 7 |
| 97°F | 179 bpm | 621 | 8 |
The watch calories climbed linearly with heart rate. But my actual metabolic work — measured by pace, distance, and elevation — was identical. The calorie increase was pure heat artifact. If I had eaten back those "extra" 134 calories on the 97°F day, I would have been consuming surplus energy for work I never actually performed. Over a month, that is a pound of fat I did not earn.
The heart rate zones were even more distorted. My lactate threshold is around 168 bpm. At 78°F, my average of 152 bpm kept me comfortably in Zone 3. At 97°F, my average of 179 bpm had me spending 40% of the run in Zone 5 — anaerobic territory. But I was not anaerobic. I was thermoregulating. My muscles were not producing more lactate. My cardiovascular system was just struggling to keep my brain from cooking inside my skull. The training effect was completely different from what the data suggested.
So I built an adjustment formula. It is not peer-reviewed. It is not elegant. But it works better than trusting a watch that thinks sweat is a sign of effort. I use a simple heat multiplier based on wet-bulb globe temperature, which accounts for both heat and humidity — the two factors that actually matter in Austin summers.
For runs under 85°F, I trust the watch. Between 85°F and 95°F, I multiply the calorie estimate by 0.92. Above 95°F, I multiply by 0.85. For heart rate zones, I add a temperature-adjusted offset: above 90°F, I subtract 8 bpm from my zone thresholds. A 170 bpm reading at 95°F is treated like 162 bpm for training purposes. Is it perfect? No. Is it better than pretending my cardiovascular system is magically 15% fitter because the asphalt is melting? Absolutely.
The water weight issue is another outlier that drove me crazy. During a typical Austin summer week, my morning weight fluctuated by as much as four pounds day to day. Not fat. Water. I would finish a run, drink a liter of electrolyte fluid, and watch the scale jump two pounds by evening. My BMI moved from 24.1 to 24.8 and back again, not because my body composition changed, but because I was fighting dehydration like a character in a Mad Max movie. I started weighing myself only after waking, before coffee, and I stopped caring about any single data point. The moving average was the only number that mattered.
I talked to Dr. Elena Voss, an exercise physiologist at UT Austin, about my data. She was not surprised. "Consumer-grade calorie algorithms assume standard thermoneutral conditions," she said. "Above 30°C, the error rate increases exponentially." Thirty degrees Celsius is 86°F. In Austin, that is called "May." By July, we are so far outside the algorithm's design parameters that the numbers are basically decorative.
Here is my practical advice, distilled from three weeks of sweaty spreadsheets and one very concerned dog named Pixel who started refusing to join me on runs after the second 100°F day. First, run early or run indoors. The Butler Trail at 6:00 AM is tolerable. At 10:00 AM, it is a bad decision wrapped in spandex. Second, ignore calorie estimates above 90°F. They are wrong in the direction that makes you eat more, which is the worst kind of wrong. Third, use perceived exertion as your primary metric. If it feels like an 8, treat it like an 8, regardless of what your watch says. Fourth, drink electrolytes, not just water. Hyponatremia is real, and it is not fun.
The hardest part was admitting that my data was dirty. I like clean data. I like variables I can control. Heat is not a variable. It is a force of nature that does not care about my training plan. But once I accepted that summer in Austin requires a different data framework — one that treats heat as a confounding variable rather than ignoring it — my training actually improved. I stopped overeating. I stopped overreaching. I stopped treating thermoregulatory heart rate spikes like fitness gains.
Your fitness tracker is a tool, not a truth. And in 105°F Austin heat, it is a tool that needs a serious reality check. Adjust your numbers. Trust your body. And maybe just stay inside until October like a sane person.
