Urban change detection: assessing biophysical drivers using machine learning and Google Earth Engine
| dc.contributor.author | Durowoju, Olufemi Sunday | |
| dc.contributor.author | Obateru, Rotimi Oluseyi | |
| dc.contributor.author | Adelabu, Samuel | |
| dc.contributor.author | Olusola, Adeyemi | |
| dc.date.accessioned | 2025-03-24T20:59:24Z | |
| dc.date.available | 2025-03-24T20:59:24Z | |
| dc.date.issued | 2025-03-20 | |
| dc.description.abstract | Urban areas are experiencing rapid transformations, driven by population growth, economic development, and policy changes. Understanding and monitoring these dynamic changes is crucial for sustainable urban planning and management. This study leverages machine learning and Google Earth Engine to investigate urban dynamics and its interactions with biophysical conditions in the Kaduna River Basin (KRB), Nigeria. This study utilized a dataset of 192 points, initially extracted from Google Earth Engine, to analyze urban transitions between 1987 and 2020, incorporating biophysical and environmental variables such as population density, precipitation, and surface temperature. The dataset was processed in R using the CARET package, with missing data imputed via K-nearest neighbors (KNN), categorical variables transformed using One-Hot Encoding, and numerical variables rescaled for consistency. A tenfold cross-validation approach was used to train and validate machine learning models, including random forest, support vector machine, KNN, and multivariate adaptive regression splines, ensuring optimal model performance. Model evaluation metrics such as overall accuracy, kappa, F1 score, and area under the curve confirmed the reliability of the models in identifying the biophysical factors influencing urban changes. The findings revealed overall accuracy of 0.80, 0.73, 0.71, and 0.72 and kappa statistics of 0.60, 0.46, 0.42, and 0.45 for random forest (RF), multivariate adaptive regression splines, support vector machine, and KNN, respectively, with RF emerging as the most accurate model (80%) for predicting urban change patterns in KRB. Land cover changes reveal rapid urban expansion (154.81%), declining water bodies (− 95.79%), and vegetation growth (174%). Machine learning models identify population density and water stress index as key urban change drivers, with climate factors like temperature and precipitation playing crucial roles. The results of this study offer valuable insights into the processes driving urban transformation and present a robust methodology for monitoring and predicting future urban development. This study aids in the creation of strategies for sustainable urban growth and the mitigation of adverse environmental impacts. | |
| dc.description.sponsorship | Open access funding provided by University of the Free State. | |
| dc.identifier.citation | Durowoju, O.S., Obateru, R.O., Adelabu, S., & Olusola, A. (2025). Urban change detection: assessing biophysical drivers using machine learning and Google Earth Engine. Environ Monit Assess 197, Article 441. https://doi.org/10.1007/s10661-025-13863-4 | |
| dc.identifier.issn | 1573-2959 | |
| dc.identifier.other | 441 | |
| dc.identifier.uri | https://doi.org/10.1007/s10661-025-13863-4 | |
| dc.identifier.uri | https://hdl.handle.net/10315/42689 | |
| dc.language.iso | en | |
| dc.publisher | Springer Science and Business Media LLC | |
| dc.rights | Attribution 4.0 International | en |
| dc.rights.uri | http://creativecommons.org/licenses/by/4.0/ | |
| dc.subject | Landscape changes | |
| dc.subject | Land use and land cover | |
| dc.subject | Random forest | |
| dc.subject | Support vector machines | |
| dc.subject | Urban ecosystem | |
| dc.symplectic.issue | 4 | |
| dc.symplectic.journal | Environmental Monitoring and Assessment | |
| dc.symplectic.subtype | Journal article | |
| dc.symplectic.volume | 197 | |
| dc.title | Urban change detection: assessing biophysical drivers using machine learning and Google Earth Engine | |
| dc.type | Article |