ABSTRACT
The objectives of this study were to: (1) analyze colonization of 12 new, unmanaged, untreated pond mesocosms to determine if they acquired a biota typical of shallow lentic environments within 1 year and (2) analyze the precision of various biometrics in terms of the percent change that must occur to distinguish treatment effects from natural variability. We analyzed colonization of phytoplankton, zooplankton, and macroinvertebrate communities in conjunction with various environmental characteristics. Phytoplankton, zooplankton, and macroinvertebrate metrics were analyzed by a statistical procedure that indicates the percent change that must occur to detect true differences between two means. Biometrics evaluated included various measures of community structure and function. For the overall mesocosm facility, communities followed expected seasonal patterns of succession and resembled the communities that would be expected in shallow lentic environments; however, there were considerable differences among individual mesocosms. With some biological management (introduction of certain macroinvertebrates and macrophytes), the mesocosms might be ready for conducting pesticide registration tests within 1 year. Biometrics showed 106a wide range of detection limits with number of taxa metrics ranging from 2 to 40%, proportion of numerical abundance metrics ranging from 2 to 60%, and density metrics ranging from 10 to 100%. Of the 20 metrics evaluated, 7 allowed detection of impact with a 20% change from the mean. Four of these were number of taxa metrics and the other three were proportion of numerical abundance metrics. Density metrics would not allow detection of treatment effects at the 20% level. For all metrics, using six replicate mesocosms, instead of three, would usually improve the ability to detect differences between treatments. With a management program (mixing water and sediments, planting macrophytes), phytoplankton and zooplankton might be more evenly distributed among the mesocosms. Macroinvertebrates were more evenly distributed among the mesocosms and might only require seeding of a few nonflying taxa. Based on the results of this study, statistically realistic detection limits would be <20% for metrics involving number of taxa, <40% for metrics involving proportions of numerical abundance, and <60% for density and biomass metrics. These criteria should provide adequate environmental protection because changes of this magnitude occur naturally in unperturbed ecosystems. We recommend analyzing pretreatment data by this approach to establish biological criteria that are realistic for a particular study.
