RESUMEN
Water use and carbon exchange of a red oak-dominated (Quercus rubra L.) forest and an eastern hemlock-dominated (Tsuga canadensis L.) forest, each located within the Harvard Forest in north-central Massachusetts, were measured for 2 years by the eddy flux method. Water use by the red oak forest reached 4 mm day(-1), compared to a maximum of 2 mm day(-1) by the eastern hemlock forest. Maximal carbon (C) uptake rate was also higher in the red oak forest than in the eastern hemlock forest (about 25 versus 15 micromol m(-2) s(-1)). Sap flux measurements indicated that transpiration of red oak, and also of black birch (Betula lenta L.), which frequently replaces eastern hemlock killed by hemlock woolly adelgid (Adelges tsugae Annand.), were almost twice that of eastern hemlock. Despite the difference between species in maximum summertime C assimilation rate, annual C storage of the eastern hemlock forest almost equaled that of the red oak forest because of net C uptake by eastern hemlock during unusually warm fall and spring weather, and a near-zero C balance during the winter. Thus, the effect on C storage of replacing eastern hemlock forest with a forest dominated by deciduous species is unclear. Carbon storage by eastern hemlock forests during fall, winter and spring is likely to increase in the event of climate warming, although this may be offset by C loss during hotter summers. Our results indicate that, although forest water use will decrease immediately following eastern hemlock mortality due to the hemlock woolly adelgid, the replacement of eastern hemlock by deciduous species such as red oak will likely increase summertime water use over current rates in areas where hemlock is a major forest species.