Models based on transformed Gaussian curves are described and used to summarise the seasonal variations in the surface temperature, dissolved reactive phosphorus concentration and phytoplankton biomass of three lakes in the English Lake District. The models are then used to quantify the effects of reducing the frequency of sampling on the detection of long-term (decadal-scale) change in the selected lakes. Three case studies are presented to demonstrate the effect of local changes in the catchment and regional-scale changes in the weather. The results demonstrate that the fitted models typically explained more than 70 percent of the recorded variation and can be used to quantify systematic changes in the timing and magnitude of the inter-annual variations. Tests where these time-series were systematically degraded showed that relatively little information was lost by fortnightly sampling but there was a marked reduction in the proportion of the variance explained when the sampling frequency was reduced from monthly to bi-monthly. Further tests on sub-sets of the acquired data showed that the capacity to detect change was not strongly influenced by the period of observation. When the length of record was reduced from 30 to 9 years the temperature and phosphorus models still explained a high proportion of the recorded variance but the results with the chlorophyll model were more time-dependent.

You do not currently have access to this content.