Applied Ecosystem Services, LLC

The Clean Water Act (CWA) declares it is the national goal of water quality to provide for the protection and propagation of fish, shellfish, and wildlife and provide for recreation in and on the water. The US Environmental Protection Agency’s(EPA) historic approach is developing a national chemical criterion for chemical elements. There are so many differences in aquatic ecosystems across the US that a more pragmatic approach uses the aquatic biota – benthic macroinvertebrates – to quantify local community’s variability and response to anthropogenic activities.

The Clean Water Act (CWA) declares it is the national goal of water quality to provide for the protection and propagation of fish, shellfish, and wildlife and provide for recreation in and on the water. The US Environmental Protection Agency’s(EPA) historic approach is developing a national chemical criterion for chemical elements. There are so many differences in aquatic ecosystems across the US that a more pragmatic approach uses the aquatic biota – benthic macroinvertebrates – to quantify local community’s variability and response to anthropogenic activities.

Environmental laws, statutes, and implementing regulations have two specific pur- poses: forecasting future conditions and determining cause and effect. The analytical models applied to existing environmental data sets do not provide results that explic- itly address the two purposes. They leave decision-makers confused, uncertain how to justify a decision, and suffering paralysis by analysis as they ask for more results that prove to be ineffective. This document explains how environmental data are analyzed (wrongly and correctly), why the analytical approaches differ, and why knowing the details makes you more effective in managing environmental data for your operations.

Most people are familiar with statistical hypothesis tests such as the t-test and ANOVA to analyze whether two or more samples (from a parametric distribution) came from the same population. The nonparametric equivalents (Wilcoxon and Kruskal-Wallis tests) are less familiar but equally robust. What is not always clear is that these models are applied to one or more response variables; e.g., chemical concentrations that result from natural or anthropogenic causes. They do not answer the question of why these values were observed.

To make informed regulatory decisions it is necessary to understand differences between ecological and environmental data. Analyses of environmental data historically use models adapted from engineering models by numerical ecologists for ecological data collected by academic and research agency scientists. These numeric and statistic models require well-structured data collected to fit assumptions and requirements of the models. This works for researchers who identify a question to be answered and work forward from that to determine when, where, and how much data need collecting to answer that question.

Toxic metals and organics commonly occur in very low concentrations in water, sediments, soils, and rocks. These concentrations are so low they cannot be quantified by analytical chemists and today’s instruments. Censored data are commonly mis-analyzed with potential costly, unnecessary, or harmful results. EPA regulations and guidelines often tell data analysts to ignore (drop) censored data or substitute an arbitrary value. The results of dropping or substituting arbitrary values are wrong.