Species Sensitivity Distributions for Ecotoxicological Risk Assessment: elaboration of a shiny app to facilitate data analysis.

MARIA FLORENCIA D´ANDREA; JULIE C. BRODEUR

Congreso; 1er Congreso Latinoamericano de R ("LatinR").; 2018

The species have different sensitivity to toxics. This variability can be repre-sented by constructing a species sensitivity distribution (SSD), whereby a group of species is described by a statistical distribution estimated from the available toxicity data. Building the SSD allows calculating the HC5, that is, the concentra-tion at which 5% of the species would be affected, allowing the establishment of an environmental quality criterion and a tool for risk assessment.The objectives of the present work are (1) estimate the HC5 from data extracted ECOTOX database from the Environment Protection Agency of the United States (EPA), (2) to develop an HTML user interface using the Shiny package, (3) allow the user to apply quality and quantity criteria to the database.A database was generated for 20 pesticides, 5 groups of aquatic animals and the toxicological endpoints of LC50 and NOEC. The first tab is used for visualiza-tion of the dataset according to the number of species with toxicological data available by pesticide. The quantity of data is important since there is a minimum number of data recommended to build an SSD. After selecting the pesticide and animal group the user can apply different quality criterion, for example, excluding from the previous selection the data from studies that did not report the chemical analysis. The final SSD is fitted to a log normal, log logistic and weibull distribu-tion using the package fitdistrplus. The HC5 is estimated by the distribution with best goodness of fit. The uncertainty analysis is performed by parametric boot-strap analysis for the selected distribution. The goal of our work is to generate a tool that allows environmental managers to interpret how the values that are important at regulatory level are affected by the quantity and quality criteria selection of data as well making easier the data analysis of a complete database.