Doubly labeled water (DLW) is an accurate, portable method for measuring free-living energy expenditure. However, under certain conditions shifts in baseline abundance of deuterium and oxygen-18 tracers used in the method may produce errors in derivation of both turnover (k) rates and calculated energy expenditure. Present objectives were to examine during what experimental situations baseline errors arise and to address means of correcting for such baseline shifts so that consequent errors in energy expenditure calculations are minimized. Under conditions where shifts in baseline abundance for deuterium and oxygen-18 parallel abundances corresponding to the natural meteoric water ratio, self-compensating changes in k values for both deuterium and oxygen will result in minimal error to the DLW energy expenditure calculations, provided that the dose ratio of isotopes also mimics the meteoric water line. However, in situations where relative shifts in abundance of each isotope across the measurement period are not in parallel relative to the natural meteoric water line, then the potential for larger DLW errors exists. Optimally, subjects should equilibrate with the new water source. Failing this, correction for shifting baseline can be accomplished by measuring isotopic abundance changes in a control group of subjects not given the DLW dose, but performing similar tasks and consuming the same diet as the group given DLW. Alternatively, theoretically based correction values can be calculated given knowledge of the abundances of the final drinking water and the interval time that subjects consumed the new fluid.