Work

Design, analysis, and interpretation of higher tier risk assessment of chemicals in aquatic microcosms

Year: 2002
Type: other
Abstract: Microcosms have been used in ecological and ecotoxicological science and research for decades.During the 1980s and 1990s, the methodology was gradually evaluated for implementation within the framework of higher tier risk assessment of pesticides both in the United States of America and within the European Union.However, in 1992 the United States Environmental Protection Agency decided to discontinue the requirements for microcosm testing for registration of pesticides and other chemicals primarily due to paramount inherent uncertainties related to the design, conduct, analysis, and interpretation of microcosm experiments.The scientific problem related to microcosm experiments is still the determination of the relative ecological (accuracy) versus statistical (precision) significance of the microcosm design.The aim of the present doctoral dissertation is thus to elucidate the inherent ecological and statistical significance trade-offs in three spatially different microcosm experiments.A further aim is to quantify the uncertainty of the three spatially different designs by using statistical power analysis..inally, it will analyse a scientific application of precautionary approaches to manage uncertainty from an environmental regulator and policy perspective.The exordium paper in this dissertation provides a state-of-the-art literature review (1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000) on the replicability of microcosm studies for risk assessment of pesticides.The review focuses on the calculation of coefficients of variation (CVs) within the studies as a measure of replication and thus the ability to detect subtle effects.Subsequent to this, a series of microcosm studies were initiated at three independent levels of ecological significance and relevance tentatively set at 25m 2 earthen outdoor ponds > 12 m 3 PVC outdoor ponds > 30L transparent PVC indoor tanks.The statistical power of each level of ecological significance was determined, bearing in mind that a direct comparison of the power is context dependent and impeded due to the unknown magnitude of the extrapolation between the systems and between the compounds used in the tests..inally,I discuss the scientific application of a reflexive science based on power analysis, analysis of needs and post-normal science in light of precautionary approaches for the environmental risk management of uncertainty and lacking knowledge.The dissertation presents the following new information.A review illustrates that a majority of microcosm studies suffer from insufficient replication and high CVs, which hampers the detectability of the study.Hence, if the natural background variation amplitude among replicates (s 2 ) are great, then will the amplitude due to toxic effect size (∆) will need to be even larger to significantly (p ≤0.05) break-through the noise of natural variation.This could again lead to low statistical power (1-β < 0.8) and an unacceptable or unknown high risk of a type II error (false negative) (paper I).A similar trend was observed in the three subsequent microcosm experiments I conducted or reviewed, the general pattern in terms of statistical significance is inverted to the ecological significance.The applicability of aquatic microcosm studies and plankton for risk assessment of Roundup is questionable due to rapid removal of the compound from the water column (paper II).The dissertation reports the first published higher tier risk assessments of two dominating perfluoro surfactants (paper III & IV).Paper V is the first published review and ranking of ten different phyto-toxicological endpoints according to ASTM guideline 1913-97 E in microcosm experiments..inally, the dissertation discusses a scientific application of precautionary approaches via power analysis for handling uncertainty in light of the current EU position, and reflexive modernisation of science (paper VI-VIII).Overall, the dissertation as a whole presents new and quantified aspects of uncertainty associated to higher tier risk assessment of chemicals in aquatic microcosms, plus guidance/recommendations for design, analysis and interpretation of higher tier risk assessment of chemicals in aquatic microcosms.The dissertation concludes that due to economic and logistical constrains it is generally easier and more feasible to increase power in small scaled microcosms by increased replication.There is a paradox between an increased ecological significance and the need for high effect sizes and thus unrealistic treatment concentrations in microcosms as opposed to realistic exposure under less ecological relevant single species laboratory conditions.Power analysis could be implemented as a short term scientific application of reflexive and precautionary approaches with an extended community peer-review system where the acceptable (∆) and risk of type I and II errors (α & β) are a priori defined based on sound ecological expert judgement for protection of system integrity, societal cost-benefit analysis, and ethics.Hence, after a pilot study a successful microcosm design could be designed with high power.The design of a microcosm experiment is a context dependent evaluation of the relative importance of statistical versus ecological significance and protection aims..inally, the thesis reflects upon the future role of microcosm experiments in environmental protection.If, the environmental agencies pay attention to the statistical power of the studies and a priori define the entities of the power equation, then microcosm studies are a good higher tier risk assessment test.If they fail to do this environmental protection could be jeopardised due to high risk of type II errors..orresearch and education they seem invaluable in pursuing better knowledge of manmade chemicals effects on a suite of species, populations and ecosystems, which can only be tested and validated in model ecosystems like microcosms.Moreover, they can be used to develop ecology and ecotoxicology and increase our understanding of extrapolation over time, in space among species, different levels of biological organisation and ecosystems.
Institution: Aarhus University
Language: English
Open Access status: green
OpenAlex ID: W2890868406
Cites: 124
Cited by: 4
Related to: 9