The validation of diatom-phosphorus transfer functions: an example from Mondsee, Austria

1. A diatom-phosphorus weighted averaging (WA) transfer function, derived from a training set of currently oligotrophic to mesotrophic European Alpine lakes, was applied to a high-resolution sediment core with annual laminae from Mondsee, an Austrian pre-alpine lake, in order to reconstruct the eurrophication history of the lake. 2. The water chemistry records of total phosphorus (TP) available for Mondsee were compared with the diatom-inferred TP from the model for the period 1975-93. The trend in TP values as inferred by the model paralleled the monitored trend in TP values closely, with matching peaks in 1979/80, a decrease in values from the early 1980s, a second smaller peak in 1986/7, and a further reduction in concentrations in the last 6 years. 3. However, there was a clear mismatch between the actual timing of the major TP peak, with the water chemistry records reporting its occurrence in 1979, and the diatom model indicating a small peak in 1980 and the highest concentrations in 1982. This can be attributed to the uncertainty of the sediment chronology for this section of the core, and possibly to the inconsistency between the core resolution and the resolution of the diatom model. 4. In terms of the actual concentrations of TP inferred by the model, they compared reasonably well with the measured data, although the model tends to underestimate for the lower core section owing largely to poor diatom assemblage analogues. In the upper part of the core, the diatom-inferred TP values were in extremely close agreement with the monitored chemical data. 5. This validation study indicates that diatom-phosphorus transfer functions are robust and are able reliably to infer past-TP concentrations from fossil diatom assemblages in sediment cores. Despite the natural intra- and interannual variability in diatom assemblages and epilimnetic water chemistry, the technique can provide accurate estimates of TP with an annual resolution. The model can be applied to selected sites with suitable sediment records to reconstruct lake TP histories, thus providing a pragmatic management tool for addressing lake eutrophication problems.