Impact of changes in analytical techniques for the measurement of polychlorinated biphenyls and organochlorine pesticides on temporal trends in herring gull eggs

Abstract

Changes in analytical approaches during the tenure of monitoring programs for organochlorine (OC) pesticides and polychlorinated biphenyls (PCBs) may affect estimates of temporal trends. We used an in-house reference material to create multiplication factors to adjust the estimates of OC pesticides and PCBs (Aroclor equivalents) in Great Lake herring gull eggs analyzed using electron capture detection (1987-1997) to be more equivalent to estimates using mass spectrometric detection (1998-2005) as well as accompanying differences in analytical procedures. We examined temporal trends in contaminant concentrations in herring gull eggs using change point regressions, to determine whether significant changes in long-term trends were associated with analytical methodology. The highest frequency of change point occurrences shifted from 1997 (when analytical methodology was altered) to 2003 after data adjustment. The explanatory power (r2) of the regressions was lower after adjustment, although only marginally so (mean r2 difference=0.04). The initial rates of decline before change points in contaminant concentrations were generally slower after the data adjustment, but after any change points the declines were not significantly different. The regression models did not change for 83.3% of the cases. The effects on the interpretation of long-term temporal trends in herring gull eggs, although not negligible, were minor relative to the magnitude of the temporal changes.