Are Your Paleoecological Indicators Telling the Truth? Diatoms in Great Lakes Sediment Cores As an Example

Many existing models to infer paleoecological conditions from species assemblages are based on weak species-environmental relationships that are confounded by secondary variables, resulting in poor predictive power. We evaluated diatoms from Laurentian Great Lakes phytoplankton samples in the development of a total phosphorus (TP) transfer function based on weighted-average species abundance. Several tests were applied to validate our new indicator. Of the 118 common diatom taxa 76% had non-monotonous responses along the Great Lakes TP gradient. However, there was substantial autocorrelation among samples, justifying the need for further validation. A randomization procedure indicated that the actual transfer function consistently performed better than simulated functions. Further, TP was minimally confounded by other environmental variables. We then hindcasted TP using fossil diatom assemblages in a Lake Superior sediment core. Passive, multivariate analysis of the fossil samples against the training set data indicated that TP was a major determinant of historical diatoms, verifying that the function was well suited to reconstruct past TP. Collectively, these results show that diatom coefficients for water quality are robust indicators of Great Lakes condition. We recommend use in management when retrospective data are needed for tracking long-term degradation, remediation and trajectories.