Lessons learned from oxygen isotopes in modern precipitation applied to interpretation of speleothem records of paleoclimate from eastern Asia
Variability in oxygen isotope ratios collected from speleothems in Chinese caves is often interpreted as a proxy for variability of precipitation, summer precipitation, seasonality of precipitation, and/or the proportion of 18O to 16O of annual total rainfall that is related to a strengthening or weakening of the East Asian monsoon and, in some cases, to the Indian monsoon. We use modern reanalysis and station data to test whether precipitation and temperature variability over China can be related to changes in climate in these distant locales. We find that annual and rainy season precipitation totals in each of central China, south China, and east India have correlation length scales of ∼ 500 km, shorter than the distance between many speleothem records that share similar long-term time variations in δ18O values. Thus the short distances of correlation do not support, though by themselves cannot refute, the idea that apparently synchronous variations in δ18O values at widely spaced (> 500 km) caves in China are due to variations in annual precipitation amounts. We also evaluate connections between climate variables and δ18O values using available instrumental measurements of δ18O values in precipitation. These data, from stations in the Global Network of Isotopes in Precipitation (GNIP), show that monthly δ18O values generally do not correlate well with either local precipitation amount or local temperature, and the degree to which monthly δ18O values do correlate with them varies from station to station. For the few locations that do show significant correlations between δ18O values and precipitation amount, we estimate the differences in precipitation amount that would be required to account for peak-to-peak differences in δ18O values in the speleothems from Hulu and Dongge caves, assuming that δ18O scales with the monthly amount of precipitation or with seasonal differences in precipitation. Insofar as the present-day relationship between δ18O values and monthly precipitation amounts can be applied to past conditions, differences of at least 50% in mean annual precipitation would be required to explain the δ18O variations on orbital time scales, which are implausibly large and inconsistent with published GCM results. Similarly, plausible amplitudes of seasonal cycles in amounts or in seasonal variations in δ18O values can account for less than half of the 4-5‰ difference between glacial and interglacial δ18O values from speleothems in China. If seasonal cycles in precipitation account for the amplitudes of δ18O values on paleoclimate timescales, they might do so by extending or contracting the durations of seasons (a frequency modulation of the annual cycle), but not by simply varying the amplitudes of the monthly rainfall amounts or monthly average δ18O values (amplitude modulation). Allowing that several processes can affect seasonal variability in isotopic content, we explore the possibility that one or more of the following processes contribute to variations in δ18O values in Chinese cave speleothems: different source regions of the precipitation, which bring different values of δ18O in vapor; different pathways between the moisture source and the paleorecord site along which exchange of 18O between vapor, surface water, and condensate might differ; a different mix of processes involving condensation and evaporation within the atmosphere; or different types of precipitation. Each may account for part of the range of δ18O values revealed by speleothems, and each might contribute to seasonal differences between past and present that do not scale with monthly or even seasonal precipitation amounts.