AbstractQuantitative ice-core paleoclimatology must account for post-depositional processes, such as vapor-phase diffusion in the firn. After pore close-off, diffusion continues to smooth the stable-isotope records δ18O and δD that are eventually recovered from the ice, leading to the loss of high-frequency information. Johnsen and others (1997) found much higher rates of diffusive smoothing in the Greenland Icecore Project (GRIP) Holocene ice than would be predicted by diffusion through solid ice alone, and Nye (1998) argued that transport through liquid veins might explain this apparent excess diffusion. However, the analysis of Johnsen and others (2000) indicates that the required vein dimensions may be unrealistically large. Here, we model the diffusion of stable isotopes in polycrystalline ice and show that the predictions of Nye (1998) and those of Johnsen and others (2000) actually represent two end-members in a range of potential behavior. Our model determines which of these asymptotic regimes more closely resembles the prevailing conditions and quantifies the role of pre-melted liquid in the smoothing of isotopic signals. The procedure thereby ties together the two approaches and provides a rostrum for accurate analysis of isotope records and paleotemperature reconstructions.
Rainfall and conduit drainage combine to accelerate nitrate loss from a karst agroecosystem: Insights from stable isotope tracing and high-frequency nitrate sensing
