The relationship between diatoms and climate in a European mountain lake training set : implications for detecting the Little Ice Age in lake sediments from central Norway
This study evaluates the use of fresh water diatoms, from lacustrine sediments, to infer past changes in climate. The main aim of the thesis is to reconstruct changes in climate over the past 500 years, including the Little Ice Age event, using lake sediments from Central Southern Norway. The relationship between surficial high altitude lake sediment diatom species and measured environmental variables is explored in 80 lakes from Central Southern Norway and North West Scotland through the creation of a 'surface sample training set'. 40 of these lakes have ice-cover duration data and diatom assemblages from these are examined to determine their relationship with measured environmental parameters. Multivariate statistical techniques demonstrate that ice-cover duration accounts for a statistically significant proportion of the diatom variance, and an ice-cover transfer function using weighted- averaging- tolerance- down- weighted (WA(tol)) techniques is developed. Unfortunately, large errors are incurred with the resulting model. For this reason diatom assemblages from all 80 lakes are explored in conjunction with the available environmental parameters. Canonical correspondence analysis with forward selection and Monte Carlo permutation tests reveal that three environmental variables independently explain significant proportions of the diatom variance. These variables are January temperature, TOC and pH. A January temperature transfer function is produced using weighted- averaging- partial- least- squares (WA-PLS) techniques and a pH transfer function is produced using WA(tol) methods. These quantitative inference models are used to help identify changes in climate for two high altitude Norwegian lakes (Gavalivatnet and Hornsjoen). Few changes occur in the diatom inferred pH reconstructions and few linkages are, therefore, made between pH and climate. The January temperature transfer function was applied and its performance is directly compared with measured values of century- long meteorological records. It appears to perform well at Hornsjoen and infers successfully the recent climatic warming. The reconstruction at Gavalivatnet does not reflect the extent of recent climatic warming suggesting that the method is not so robust for this site. Inferences about the changes in the Norwegian climate during the LIA are made using a combination of diatom inferred reconstructed temperatures, lithostratigraphic changes, chrysophyte cyst production and diatom species shifts. Several trends are evident within the proxy records of the lakes which can be linked with the climatic changes associated with the Little Ice Age and recent climatic warming. It is concluded, however, that although changes in the lake biota could be related to climatic fluctuations the signal is complex and multi-faceted. It is proposed that the two lakes respond individually to climatic changes, due to differing lake chemistry and bathymetry, despite being exposed to similar driving macroclimate variables.