δ18O and δD values in meteoric groundwaters reflect those in their source precipitation, which normally displays seasonal ranges of several to many permil. Normally there are substantial reductions in the groundwater isotopic ranges due to mixing in any vegetation, soil cover and the aquifer itself, plus the distortions that can be introduced by evaporation. Nevertheless, in carbonate karst regions (and others) many spring waters are found to preserve significant proportions of these seasonal signals. They are of current interest because they may also be detected in calcite speleothems precipitated from groundwater in caves, in calcretes and calcareous nodules in soils, etc. thus permitting detailed paleoclimate reconstructions.
This paper presents some baseline data on reduction of isotope seasonal signals in groundwater that are pertinent to such speleothem studies. They are from artificial tank experiments at Shawan Karst Test Site, Guizhou, China. The climate is humid subtropical with mean annual precipitation of 1340 mm, >80% falling between May and October. The concrete tanks measure 20 × 5 m, 3 m deep each, point-drained at one end to simulate a spring. They are filled with 2.5 m of local limestone quarry gravel, creating carbonate aquifers with primary porosities ∼ 50%. A first tank is left bare, simulating the rocky desertification found in parts of the region. The other tanks are topped with 0.5 m of local, clay-rich residual soil, one left bare, the other three sown with rainy season corn, permanent grass and permanent shrubs respectively. Measurements cover the three climatic years, November 2015-November 2018. > 90% of the rainfall events were sampled, and the drains every ten days.
δ18O and δD in the rainfall displayed strong and regular sinusoidal oscillations with the successive seasons. In the bare rock tank these were reduced (damped) by 56~74% in the different years, with phase (time) lags of 127~134 days at the drain. In the soil-topped tanks damping ranged from 61% to 93% and the time lags from 164 to 202 days: differences between them are not statistically significant but all do differ significantly from the bare rock. In all cases, the responses of the two isotopes were similar in direction and magnitude. The 0.5 m soil cover had the most important effect. The bare rock aquifer conserved the most information. Losses were similar in the four soil-topped tanks, responding to stronger evaporation and obstruction of flow in the soil. These findings are briefly compared to seasonal isotope variations reported in speleothem drip waters in limestone caves in China and elsewhere in the world.
