Occurrence of chemical substances in water supply systems of Brazil: a nonparametric approach for statistical analysis of Sisagua data
The objective of this work was to develop a methodology for statistical analysis of monitoring data of chemical compounds in drinking water supply systems in Brazil, using data from Sisagua (Drinking Water Quality Surveillance Information System). Initially, the inconsistencies in the database were...
| Autores: | , , , , , , |
|---|---|
| Formato: | artículo |
| Estado: | Versión publicada |
| Fecha de publicación: | 2022 |
| País: | Brasil |
| Recursos: | Universidade Federal de Santa Maria (UFSM) |
| Repositorio: | Revista Ciência e Natura (Online) |
| Idioma: | inglés |
| OAI Identifier: | oai:ojs.pkp.sfu.ca:article/63368 |
| Acesso em linha: | https://periodicos.ufsm.br/cienciaenatura/article/view/63368 |
| Access Level: | acceso abierto |
| Palavra-chave: | Censored data drinking water environmental data Kaplan-Meier estimators nondetects Dados ambientais Dados censurados Estimadores de Kaplan-Meier Não detectados Água potável |
| Resumo: | The objective of this work was to develop a methodology for statistical analysis of monitoring data of chemical compounds in drinking water supply systems in Brazil, using data from Sisagua (Drinking Water Quality Surveillance Information System). Initially, the inconsistencies in the database were identified and adjusted. Then, the descriptive statistics were estimated using the Kaplan-Meier (KM) method, evaluating its applicability in different censored data sets. The descriptive parameters were compared with the substitution method. The substitution method showed susceptibility to biased estimates, especially for groups of compounds containing high percentage of censored data and with high limits of quantification and detection, leading to higher descriptive parameters compared to KM method. This work reinforces the need to use appropriate methods for analyzing environmental data and evidences that the analysis of this type of data may be complex. The methods proposed here can help environmental scientists to deal with this issue, providing a systematic procedure to check and solve consistency problems, as well as presenting a nonparametric approach for computing descriptive statistics for environmental monitoring data. |
|---|