Whole-Body Sweat Rate Prediction: Outdoor Running and Cycling Exercise

Our aim was to develop and validate separate whole-body sweat rate prediction equations for moderate to high intensity outdoor cycling and running, using simple measured or estimated activity and environmental inputs. Across two collection sites in Australia, 182 outdoor running trials, and 158 outdoor cycling trials were completed at a wet-bulb globe temperature ranging from ~15 to ~29˚C, with ~60-min whole-body sweat rates measured in each trial. Data were randomly separated into model development (running: 120; cycling: 100 trials) and validation groups (running: 62; cycling: 58 trials), enabling proprietary prediction models to be developed and then validated. Running and cycling models were also developed and tested when locally measured environmental conditions were substituted with participant subjective ratings for black globe temperature, wind speed, and humidity. The mean absolute error for predicted sweating rate was 0.03 and 0.02 L·h -1 for running and cycling models, respectively. The 95% confidence intervals for running (+0.44 and -0.38 L·h -1) and cycling (+0.45 and -0.42 L·h -1) were within acceptable limits for an equivalent change in total body mass over 3 h of ±2%. The individual variance in observed sweating described by the predictive models was 77% and 60% for running and cycling, respectively. Substituting measured environmental variables with subjective assessments of climatic characteristics reduced the variation in observed sweating described by the running model by up to ~25%, but only by ~2% for the cycling model. These prediction models are publicly accessible (https://sweatratecalculator.com) and can guide individualized hydration management in advance of outdoor running and cycling.