Knowledge of natural variability in aquatic ecosystems is vital for assessing the nature and amplitude of human-induced change, and for predicting future anthropogenic impacts. Distinguishing between naturally and anthropogenically caused variability in lake sediment records can be problematic, however, because both drivers can produce similar ecological effects. Standard sediment-based approaches for reconstructing past environmental changes tend to focus on qualitative and quantitative variations in palaeoenvironmental indicators, with little significance attached to their complete absence. We used multiple variables in radiometrically dated sediment cores collected from two sites in Lough Mask, a lake in western Ireland. Results suggest that the Lough Mask sediment record has been a sensitive recorder of past climate variability, especially changing precipitation, since the middle Holocene. Variations in the presence of aquatic siliceous microfossils and calcareous macrofossils, and changing sediment lithology and geochemistry, indicate a quasi-cyclic response to oscillations in climate conditions that correspond generally with palaeoclimate findings from elsewhere in NW Europe, including other sites in Ireland. We conclude that during much of the middle to late Holocene, prolonged periods of relatively high rainfall in the catchment reduced nutrient inputs to the lake, particularly silica and calcite. Diatom productivity consequently declined, whereas dissolution of frustules was enhanced. During relatively dry climate periods, availability of these nutrients increased, diatom productivity was higher, and dissolution was reduced. Relatively early human impacts are evident in the sediment record beginning ca. 1,000 BP. The results highlight the aquatic and taphonomic effects of complex interactions among past variations in catchment conditions, climate and water chemistry. The complexity of these interactions and their effects, mediated through the characteristics of Lough Mask and its catchment, pose problems for conventional interpretation of palaeolimnological data and their use in computer-based simulations of future changes in stresses on aquatic ecosystems and their consequent impacts.