In-water testing is frequently used to simulate reduced gravity for quasi-static tasks. For dynamic motions, however, the assumption has been that drag effects invalidate any data, and in-water testing has been dismissed in favor of complex and restrictive techniques such as counterweight suspension and parabolic flight. In this study, motion-capture was used to estimate treadmill gait metrics for three environments: underwater and ballasted to 1 g and to 1/6th g, and on dry land at 1 g. Ballast was distributed anthropometrically. Motion-capture results were compared with those for a simulated dynamic walker/runner, and used to assess the effect of the in-water environment on simulation fidelity. For each test case, the model was tuned to the subject's anthropometry, and stride length, pendulum frequency, and hip displacement were computed. In-water environmental effects were found to be sufficiently quantifiable to justify using in-water testing, under certain conditions, to study partial-gravity gait dynamics.