Dear Editor-in-Chief, We are grateful for the opportunity to respond to the insightful critiques provided by our colleagues and clarify our study’s rationale (1). Load carriage over complex terrain is a common scenario for dismounted military. However, it is challenging to study as it involves prolonged outdoor walking at a varying pace over a series of route segments of different distances, grades, and surfaces. Global positioning systems (GPS) offer a practical solution for obtaining data that can be input into walking equations to estimate energy expenditure (EE). We made estimates of EE using GPS-based modeling as a pragmatic means of providing needed solutions to our military customers for use in real-world, non–steady-state activities. Our study attempted to estimate EE contributed by aerobic metabolism in dynamic field conditions. When taken in the context of exercise with a discrete on and off phase, we see how our methodology has led to some confusion. Because of respiratory gas exchange kinetics, measures obtained in real time via gas exchange display a classic half-time response, both on and off; therefore, they do not directly compare with steady-state estimates. The approach we took sought to adjust for this half-time response. As part of our rationale, we included references describing delayed respiratory gas exchange kinetics (2,3) and excess postexercise oxygen consumption (EPOC) (4,5). Typically, estimates of EE are made using a square-wave prediction; however, using this approach avoids accounting for these half-time responses. To address these time offsets, we applied a simple exponential smoothing filter to better align walking equation estimates with aerobic EE calculated from measured oxygen uptake. This approach assumes the steady-state energy cost for a given work is correct but adjusts the time period during which the cost is expended to closer approximate how the energy cost is met by aerobic metabolism. For example, exponential smoothing a basic on–off square-wave results in a curve similar to the oxygen deficit/EPOC response to exercise. As with any method, there are limitations and room for improvement. On one hand, exponential smoothing does not immediately account for the contribution of all anaerobic energy stores when walking speed, grade, or terrain costs increase. On the other hand, instantaneous EE estimates are difficult to verify, and solely using them ignores order effects such as additional aerobic EE during trail segments that follow more metabolically taxing ones (e.g., downhill slope after a climb to the summit). To address other raised concerns in order: (i) GPS device details were listed in the original manuscript; (ii) we did state the need for downhill correction factor(s) but do note our colleagues have since developed a series of equations estimating EE during level, uphill, and downhill walking from GPS, heart rate, and accelerometry (6); (iii) the same conversion factor was applied to both measured and estimated V˙O2; and (iv) our colleagues’ equation estimates V˙O2 from combined mass of the individual plus all equipment. We appreciate the constructive feedback and look forward to future work in this area. Adam W. Potter David P. Looney William R. Santee U.S. Army Research Institute of Environmental MedicineNatick, MA