Sustainable development goal 6 monitoring through statistical machine learning – Random Forest method

Global reports from the United Nations project significant deficits in achieving water and sanitation targets by 2030, emphasizing the need for advanced methodologies in ecosystem monitoring. This study examines the integration of the Random Forest machine learning algorithm with freely available sa...

Full description

Saved in:
Bibliographic Details
Main Authors: Murilo de Carvalho Marques, Abdoulaye Aboubacari Mohamed, Paulo Feitosa
Format: Article
Language:English
Published: Elsevier 2025-06-01
Series:Cleaner Production Letters
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2666791624000344
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849690376171945984
author Murilo de Carvalho Marques
Abdoulaye Aboubacari Mohamed
Paulo Feitosa
author_facet Murilo de Carvalho Marques
Abdoulaye Aboubacari Mohamed
Paulo Feitosa
author_sort Murilo de Carvalho Marques
collection DOAJ
description Global reports from the United Nations project significant deficits in achieving water and sanitation targets by 2030, emphasizing the need for advanced methodologies in ecosystem monitoring. This study examines the integration of the Random Forest machine learning algorithm with freely available satellite imagery and open-source tools to monitor Permanent Protected Areas (PPAs) in the Distrito Federal, Brazil, contributing to Sustainable Development Goal (SDG) 6, which prioritizes clean water and sanitation. The research adopts a methodological approach that classifies land use changes within PPAs, with a focus on riparian zones along riverbanks, utilizing high-resolution Sentinel-2 satellite data processed through the Google Earth Engine platform. The findings indicate a 6% increase in native vegetation within PPAs from 2015 to 2022, highlighting the utility of machine learning technologies in environmental monitoring. The Random Forest algorithm demonstrated robust performance, with classification accuracy rates ranging from 83% to 88% and Kappa coefficients between 0.73 and 0.84. These results underscore the method's ability to enhance data granularity and reliability, supporting informed decision-making in ecosystem management. This research contributes to advancements in environmental monitoring methodologies and aligns with international efforts to achieve SDG targets. Further studies should investigate the incorporation of additional machine learning models to improve monitoring accuracy and support sustainable development initiatives.
format Article
id doaj-art-e1b0c9e8b51f40d5980bbc919798861a
institution DOAJ
issn 2666-7916
language English
publishDate 2025-06-01
publisher Elsevier
record_format Article
series Cleaner Production Letters
spelling doaj-art-e1b0c9e8b51f40d5980bbc919798861a2025-08-20T03:21:19ZengElsevierCleaner Production Letters2666-79162025-06-01810008810.1016/j.clpl.2024.100088Sustainable development goal 6 monitoring through statistical machine learning – Random Forest methodMurilo de Carvalho Marques0Abdoulaye Aboubacari Mohamed1Paulo Feitosa2University of São Paulo, Av. Pádua Dias, 11, Piracicaba, SP, 13418-900, BrazilDepartment of Rural Economics, Universidade Federal de Viçosa, Avenida Purdue, s/nº, Campus Universitário, Edifício Edson Potsch Magalhães, 36570-900, Viçosa, Minas Gerais, BrazilUniversity of São Paulo, Av. Prof. Lúcio Martins Rodrigues, 443, Butantã, São Paulo, SP, Brazil; Corresponding author. University of São Paulo, Av. Prof. Lúcio Martins Rodrigues, 443, Butantã, São Paulo, SP, 05508-0203, Brazil.Global reports from the United Nations project significant deficits in achieving water and sanitation targets by 2030, emphasizing the need for advanced methodologies in ecosystem monitoring. This study examines the integration of the Random Forest machine learning algorithm with freely available satellite imagery and open-source tools to monitor Permanent Protected Areas (PPAs) in the Distrito Federal, Brazil, contributing to Sustainable Development Goal (SDG) 6, which prioritizes clean water and sanitation. The research adopts a methodological approach that classifies land use changes within PPAs, with a focus on riparian zones along riverbanks, utilizing high-resolution Sentinel-2 satellite data processed through the Google Earth Engine platform. The findings indicate a 6% increase in native vegetation within PPAs from 2015 to 2022, highlighting the utility of machine learning technologies in environmental monitoring. The Random Forest algorithm demonstrated robust performance, with classification accuracy rates ranging from 83% to 88% and Kappa coefficients between 0.73 and 0.84. These results underscore the method's ability to enhance data granularity and reliability, supporting informed decision-making in ecosystem management. This research contributes to advancements in environmental monitoring methodologies and aligns with international efforts to achieve SDG targets. Further studies should investigate the incorporation of additional machine learning models to improve monitoring accuracy and support sustainable development initiatives.http://www.sciencedirect.com/science/article/pii/S2666791624000344Machine learningSDG 6Ecosystem monitoringRandom forestSatellite imageryProtected areas
spellingShingle Murilo de Carvalho Marques
Abdoulaye Aboubacari Mohamed
Paulo Feitosa
Sustainable development goal 6 monitoring through statistical machine learning – Random Forest method
Cleaner Production Letters
Machine learning
SDG 6
Ecosystem monitoring
Random forest
Satellite imagery
Protected areas
title Sustainable development goal 6 monitoring through statistical machine learning – Random Forest method
title_full Sustainable development goal 6 monitoring through statistical machine learning – Random Forest method
title_fullStr Sustainable development goal 6 monitoring through statistical machine learning – Random Forest method
title_full_unstemmed Sustainable development goal 6 monitoring through statistical machine learning – Random Forest method
title_short Sustainable development goal 6 monitoring through statistical machine learning – Random Forest method
title_sort sustainable development goal 6 monitoring through statistical machine learning random forest method
topic Machine learning
SDG 6
Ecosystem monitoring
Random forest
Satellite imagery
Protected areas
url http://www.sciencedirect.com/science/article/pii/S2666791624000344
work_keys_str_mv AT murilodecarvalhomarques sustainabledevelopmentgoal6monitoringthroughstatisticalmachinelearningrandomforestmethod
AT abdoulayeaboubacarimohamed sustainabledevelopmentgoal6monitoringthroughstatisticalmachinelearningrandomforestmethod
AT paulofeitosa sustainabledevelopmentgoal6monitoringthroughstatisticalmachinelearningrandomforestmethod