Isotope hydrology and residence times of the unimpounded Meramec River Basin, Missouri

Strong isotopic forcing of hydrologic systems in eastern Missouri, caused b y the large seasonal variations in the delta(18)O values of meteoric precip itation, can be used to determine numerous characteristics of hydrologic sy stems including the residence time of the water. The normal annual average delta(18)O value of meteoric precipitation in this region is about -6.9 par ts per thousand, but during the period June 1997-May 1998, which incorporat es an Fl Nino event, the average delta(18)O of precipitation was -9.4 parts per thousand. Monthly averages are highly variable, ranging from -2.8 part s per thousand in August 1996 to -15.1 parts per thousand in January 1998, and define a cycloid-like annual pattern. This meteoric forcing gives rise to similar patterns of isotopic variation in springs and rivers, but with g reatly reduced amplitudes. Thus the delta(18)O variations for the precipita tion have an amplitude exceeding 10 parts per thousand, yet the annual ampl itudes of the variations in the unimpounded Meramec and Big Rivers are only about 3 parts per thousand, and the amplitudes of several karst springs, i ncluding the 'first magnitude' Maramec Spring, are even smaller at about 1 parts per thousand. Most of the isotopic variation in streamflow can be exp lained by a simple exponential weighting of the preexisting rainfall events , such that the most recent precipitation more greatly influences the flow than earlier precipitation events, according to our formulation: delta(18)O(flow) = Sigma delta(i)P(i)e(-ti/tau)/Sigma P(i)e(-ti/tau) where delta(i) and P-i are the delta(18)O value and amount for a given rain event, t(i) is the time interval between the storm and the stream or sprin g sample, and tau is the residence time. For the Meramec and Big Rivers, ta u takes on a value of close to 100 days, whereas it is 1-2 years for severa l springs. Smaller contributions with tau on the order of 1-10 days are sup erimposed, representing the latest storm events. This method has a signific ant advantage over the standard mixing arguments for overland flow and base flow contributions to hydrologic systems, in that it not only demonstrates the dominant contributions of 'pre-event' water in the systems, but it impl icitly accounts for the variability of baseflow and also provides the appro ximate time scale for subsurface mixing. (C) 1999 Elsevier Science B.V. All rights reserved.