Seasonal climate inferences from high-resolution modern diatom data along a climate gradient: a case study

Abstract

This study represents a step towards developing seasonal climate inferences by using high-resolution modern data sets. The importance of seasonal climate changes is highlighted by the instrumental record of a meteorological station close to our study site (lac du Sommet in the Laurentian Mountains, Quebec, Canada): Between 1966 and 2001, May temperatures increased significantly by 3.1°C (r = 0.41, n = 35, p 0.05). Comparison of this instrumental record with fossil diatom assemblages in a sediment core from lac du Sommet showed that axis one of a principal component analysis (PCA) of the fossil diatoms was best correlated with wind velocity in June (r = 0.62, n = 19, p < 0.005) and that past diatom production was significantly enhanced in periods with colder July temperatures (r = −0.77, n = 19, p < 0.0005) and higher wind velocity in June (r = 77, n = 19, p < 0.0005). The strong impact of the spring and summer conditions on overall diatom composition and productivity suggests that seasonal lake responses to climate are more important than annual mean temperatures. However, the seasonal dynamics of diatom communities are not well understood, and seasonality is rarely inferred effectively from lake sediment studies. Our research presents a pilot study to answer a twofold question: Is it possible to identify diatom communities which are typical for warmer or colder seasonal climate using sediment traps, and if it is, can this knowledge be used to infer seasonal climate conditions from fossil diatom assemblages? To address these questions, the seasonal dynamics of diatom communities and water chemistry were studied using sediment traps and water samples at biweekly intervals in four lakes distributed along an altitudinal gradient in the Laurentian Mountains from May through October 2002. Date of ice break-up was significantly related to the diatom assemblages taken in spring and uncorrelated to other significant environmental variables. Summer water temperature, circulation of the water column and pH explained a significant part of the biological variance in summer, and total nitrogen (TN) explained most of the biological variance in autumn. To infer these variables, weighted averaging partial least squares models were applied to the seasonal data sets. Inferred ice break-up dates were significantly correlated with number of days below 0°C in April (r = 0.52, n = 19, p < 0.025), inferred circulation of the water column was significantly related to measured wind velocity in June (r = 0.64, n = 19, p < 0.005), inferred summer water temperature and inferred pH was significantly related to measured July air temperature (r = 0.50, r = −53, n = 19, p < 0.025) and inferred TN autumn concentrations had an inverse relationship to August temperatures (r = −0.53, n = 19, p < 0.01). This comparison of the historical record with diatom-inferred seasonal climate signals, based on the comparison of fossil diatom assemblages with modern sediment trap data of high temporal resolution, provides a promising new approach for the reconstruction of seasonal climate aspects in paleolimnological studies.