How well do sediment indicators record past climate? An evaluation using annually laminated sediments

摘要

The reliability of paleoclimatic inferences from lake-sediment records rests on the understanding of how various sediment indicators respond to environmental changes. Despite the recent proliferation of paleoclimatic records, only a limited number of studies have rigorously evaluated potential indicators by comparing lake-sediment records with instrumental weather data. We analyzed annually laminated sediments of the past 100 years from a lake in north-central Minnesota for a suite of variables commonly used for climatic reconstructions. Results were compared with time series of climatic or climate-derived variables, as well as with indices of climate modes thought to influence the regional climate of the midwestern United States. The oxygen-isotopic composition of calcite (δ~(18)O_c) shows trends similar to those of effective moisture (as measured by precipitation [P] minus actual evapotranspiration [AET], P-AET, and the Palmer Drought Severity Index [PDSI]), with high δ~(18)O_c values generally corresponding to low P-AET and low PDSI. δ~(18)O_c also exhibits striking correspondence with the index of Pacific Decadal Oscillation (PDO) but it lags behind PDO by 3 years. Elevated δ~(18)O_c values during the warm PDO phase probably reflect warm and dry climatic conditions in the midwestern US, especially during winter months. The carbon-isotopic composition of calcite (δ~(13)C_c) shows some similarity with δ~(18)O_c but also displays stratigraphic patterns resembling those of lake-productivity indicators, including biogenic silica, inverse of inorganic carbon, ratio of organic carbon to nitrogen, and to a lesser extent, organic carbon. δ~(13)C_c is correlated with mean annual air temperature (MAAT) and the index of Atlantic Multidecadal Oscillation (AMO), probably because elevated MAAT stimulates the rate of algal carbon uptake, lengthens growing season, and/or enhances CO_2 degassing. Varves are typically thicker during periods of lower δ~(18)O_c and higher P-AET (or PDSI) values, because wet climatic conditions probably increase nutrient availability and lake productivity. Gray-scale intensity does not co-vary with any of the above climatic variables or climate-mode indices. These results demonstrate the utility of lake-sediment analyses for reconstructing temperature, drought, and large-scale climatic modes at Steel Lake. However, application to down-core reconstructions may be compromised by a number of factors, including the site specificity and non-stationarity of such relationships.