Seleção de atributos de assembleias ictias e sua aplicação na avaliação da qualidade dos recursos hídricos

Abstract

The development of multimetric indices (IMM) based on biological assemblages to evaluate the environmental quality originated in 80’s with the Biotic Integrity Index (IBI) based on fish assemblages proposed by James Karr. Since then, several changes were implemented for different biological groups (macro invertebrates, plankton, plants, birds, etc.) and environments (lakes, estuaries, terrestrial environment, etc.) giving rise to new IMMs used in environmental monitoring of various countries. Improvements related to the development process of these indices have been proposed, and one of them deals with subjectivity in the selection of metrics. In this sense, it has been suggested that selection based on a wide list of metrics and evaluation of these, by means of statistical methods, provides a greater robustness to the selected metrics. This work aimed to select metrics tested and validated with fish assemblages in studies that has developed and applied IMMs to evaluate the environmental quality of rivers and streams in different regions around the world over the last 35 years. From a review of 117 studies, 41 were selected for identification of metrics considered robust for assessing the quality of rivers, resulting in seven selected metrics: % of lithophilics individuals; % of individuals sick injuries or abnormalities; number of intolerant species; number of native species; % of omnivores individuals; % of tolerant individuals; and number of species. The evaluation identified a set of metrics considered to be robust for future development of multimetric indices or other assessment tool for these types of environment. Rivers are among the ecosystems most affected by human activity and monitoring of these resources is critical to its management. Thus, the use of appropriate tools for assessment of their quality can lead to the adoption of measures to control and reduce degradation or promote recovery, contributing to the maintenance of ecosystem services assigned to these resources, and assist in the management process. Fuzzy logic is a field of artificial intelligence considered to deal with complex issues and Environmental Measures to development of environmental indices. This article aims to present a automated valuation model of environmental quality of rivers, based on fuzzy logic, using data from fish assemblages, which act as a support system for taking management decisions. The main focus of this work was the construction of a general framework that, in turn, can be applied in different types of environments. The constructed model was demonstrated with simulated data of an assembly of fish. Some strategies related to the solution of problems of high number of inference rules demanded by the model, were also fundamental, such as the hierarchical approach, grouping metrics, and rules grouping. To the problem of subjectivity involved in the definition of membership functions, a criterion based on the distribution of values expected from the reference sites was proposed. Although it has not been validated from a real database, the results of the literature simulations and reports regarding the efficiency of the application of fuzzy logic in this context, give the model proposed great potential as a tool for assessing the quality of rivers and support the management of these resources.