The East Hararghe Zone (EHZ) is one of the eastern Ethiopian zones most endowed with diverse landscapes and abundant resources to promote and use for ecotourism development. Potential ecotourism sites have, however, hardly ever been explored so far. The objective of this study was to model and identify potential ecotourism sites by combining Geographic Information System (GIS)-based Multi-Criteria Decision-Making (MCDM) and Analytical Hierarchy Process (AHP) methods. Six criteria, including landscape and naturalness, wildlife, topography, accessibility, geology, and climate, were established based on experts' preferences and literature, and the thematic factors for suitability modeling were derived from freely accessible satellite imagery and existing geospatial data and combined using a weighted linear combination (WLC) method. The results reveal that about 26.19% were highly suitable, 35.34% were moderately suitable, 25.28% were marginally suitable, and 13.17% were not suitable. Most areas with high to marginal suitability were found in the southeast, southwest, and uplands in the northern part, while the southernmost extent had the highest proportion of areas unsuitable for ecotourism development. The area under the curve (AUC) assessment has verified the model's performance, resulting in an overall AUC of 74.96%. This suggests that a model-driven map is a reliable spatial support tool for sustainable ecotourism development in areas with diverse landscapes and resources.

In arid and semi-arid regions where surface water resources are scarce, groundwater is crucial. Accurate mapping of groundwater depth is vital for sustainable management practices. This study evaluated the performance of three spatial interpolation techniques - inverse distance weighting (IDW), ordinary kriging (OK), and radial basis functions (RBF) - in predicting groundwater depth distribution across Dire Dawa City, Ethiopia. The results demonstrated the superiority of the RBF method, exhibiting the lowest RMSE (3.21 m), MAE (0.16 m), and the highest R² (0.99) compared to IDW and OK. The IDW method emerged as the next best performer (RMSE = 4.68 m, MAE = 0.16 m, R²= 0.97), followed by OK (RMSE = 5.32 m, MAE = 0.42 m, R²= 0.95). The RBF's superior accuracy aligns with findings from other semi-arid regions, underscoring its suitability for data-scarce areas like Dire Dawa. This comparative evaluation provides valuable insights for selecting the optimal interpolation method for groundwater depth mapping, supporting informed decision-making in local water resource management. The methodological approach comprised:•Implementation of three interpolation techniques, namely, inverse distance weighting (IDW), ordinary kriging (OK), and radial basis functions (RBF), utilizing 56 groundwater depth measurements from locations dispersed throughout the study area.•Cross-validation through randomly withholding 20 % of the data for validation purposes.•Comparison of the techniques based on statistical measures of accuracy, including root mean square error (RMSE), mean absolute error (MAE), and the coefficient of determination (R²).

Urban growth in sub-Saharan Africa presents significant challenges to sustainable development, food security, and environmental conservation. The rapid urban expansion and impact on agricultural land reduction in central Ethiopian metropolitan areas (Addis Ababa and Sheger city) exemplify these issues while simultaneously offering opportunities for sustainable development. This study aims to quantify and characterize the spatiotemporal dynamics of urban expansion in Addis Ababa and the surrounding Sheger city, explicitly focusing on understanding the impact of urban expansion on farmlands. The supervised random forest (RF) classification in the Google Earth Engine platform was used to prepare land use and land cover (LULC) for 1990, 2000, 2010, and 2023. The study employed an analytical framework incorporating multiple methodologies: intensity analysis at interval, categorical, and transitional levels to quantify urban growth trajectories; gradient direction and distance analyses to examine spatial expansion patterns; and Land Expansion Index (LEI) and Landscape Dynamic Typology (LDT) metrics to characterize the urban morphology and spatial dynamics of the study area. The results revealed that edge expansion is the predominant mode of urban development, primarily affecting farmlands in the eastern section. Built-up areas quadrupled between 1990 and 2023, whereas arable land declined. Intensity analysis revealed significant changes, particularly affecting farmlands. Our LDT analysis showed reduction in stable areas and increased in LULC changes from 1990 to 2023. The findings highlight the need for revised urban development strategies in Ethiopia to focus on compact and efficient growth while safeguarding agricultural lands, aligning with SDGs 2, 11, and 15 to promote balanced development that ensures urban and agricultural sustainability.

Land degradation from gully erosion poses a significant threat to the Erer watershed in Eastern Ethiopia, particularly due to agricultural activities and resource exploitation. Identifying erosion-prone areas and underlying factors using advanced machine learning algorithms (MLAs) and geospatial analysis is crucial for addressing this problem and prioritizing adaptive and mitigating strategies. However, previous studies have not leveraged machine learning (ML) and GIS-based approaches to generate susceptibility maps identifying these areas and conditioning factors, hindering sustainable watershed management solutions. This study aimed to predict gully erosion susceptibility (GES) and identify underlying areas and factors in the Erer watershed. Four ML models, namely, XGBoost, random forest (RF), support vector machine (SVM), and artificial neural network (ANN), were integrated with geospatial analysis using 22 geoenvironmental predictors and 1,200 inventory points (70% used for training and 30% for testing). Model performance and robustness were validated through the area under the curve (AUC), accuracy, precision, sensitivity, specificity, kappa coefficient, F1 score, and logarithmic loss. The relative slope position is most influential, with 100% importance in SVM and RF and 95% importance in XGBoost, while annual rainfall (AR) dominated ANN (100% importance). Notably, XGBoost demonstrated robustness and superior prediction/mapping, achieving an AUC of 0.97, 91% accuracy, 92% precision, and 81% kappa while maintaining a low logloss (0.0394). However, SVM excelled in classifying gully resistant/susceptible areas (97% sensitivity, 98% specificity, and 91% F1 score). The ANN model predicted the most areas with very high gully susceptibility (13.74%), followed by the SVM (11.69%), XGBoost (10.65%), and RF (7.85%) models, while XGBoost identified the most areas with very low susceptibility (70.19%). The ensemble technique was employed to further enhance GES modeling, and it outperformed the individual models, achieving an AUC of 0.99, 93.5% accuracy, 92.5% precision, 97.5% sensitivity, 95.4% specificity, 85.8% kappa, and 94.9% F1 score. This technique also classified the GES of the watershed as 36.48% very low, 26.51% low, 16.24% moderate, 11.55% high, and 9.22% very high. Furthermore, district-level analyses revealed the most susceptible areas, including the Babile, Fedis, Harar, and Meyumuluke districts, with high GES areas of 32.4%, 21.3%, 14.3%, and 13.6%, respectively. This study offers robust and flexible ML models with comprehensive validation metrics to enhance GES modeling and identify gully prone areas and factors, thereby supporting decision-making for sustainable watershed conservation and land degradation prevention.

Water erosion is accelerating soil loss rates in the East Hararghe Zone due to inappropriate human activities and their complex and intertwined interactions with natural factors, particularly in sensitive agroecosystems that lack soil and water conservation (SWC) measures. Although these dynamic processes cause prolonged impacts, a comprehensive assessment of the risk of soil erosion has not yet been undertaken at the zonal level. To bridge this gap, we employed the revised universal soil loss equation (RUSLE) prediction model, along with remote sensing and geographic information systems (GIS), to estimate annual soil erosion rates, analyze the temporal-spatial patterns of erosion risk, and evaluate the potential of standard conservation practices to reduce soil loss in croplands. Total soil erosion (in millions of tonnes/year; Mt yr⁻¹) was estimated to be 9 in 1990, 14 in 2000, 12 in 2010, and 11 in 2020, with average rates of 33, 50, 44, and 39 t ha⁻¹ yr⁻¹, respectively. This suggests an overall 18% increase in soil erosion from 1990 to 2020. Over 75% of the area showed a tolerable soil loss rate (120 t ha⁻¹ yr⁻¹), which accounts for more than 80% of soil loss, making SWC planning a priority. Analysis of land-use land-cover change (LULCC) confirmed a higher increase in soil loss for LULCC that involved conversion to croplands, with average rates of 36.4 t ha⁻¹ yr⁻¹ (1990–2000), 70 t ha⁻¹ yr⁻¹ (2000–2010), and 36 t ha⁻¹ yr⁻¹ (2010–2020). The results have further revealed that implementing supportive practices such as terracing, stripping, and contouring could reduce average soil erosion by approximately 87%, 65%, and 29%, respectively, compared to the baseline model’s prediction. Therefore, a rigorous cost–benefit analysis is essential to design and implement optimal location-specific practices that maximize investment returns in SWC efforts and ecological restoration. However, we acknowledge the limitations of this study, associated with an empirical model that does not account for all forms of erosion, as well as reliance mainly on secondary data, which may affect the accuracy of the predicted outcomes.

Air pollution is a critical twenty-first century environmental and public health challenge that is linked to millions of deaths and ecological harm. Accurate prediction of pollutants such as PM₁₀ is vital for mitigation and urban sustainability. This study combines geospatial modeling with three machine learning algorithms (MLAs), Random Forest (RF), K-Nearest Neighbor (KNN), and Naïve Bayes (NB), to identify PM₁₀ hotspots in Addis Ababa, Ethiopia. PM₁₀ data from 11 stations (August 2021–August 2023) were analyzed alongside 25 atmospheric, climatic, anthropogenic, and pollution source predictors. A concentric zonal approach was used to assess spatial variability across radial distances and directional sectors and was supported by 30 m-resolution satellite imagery, climate data, and local geospatial repositories. The model accuracies were 0.95 (KNN), 0.93 (RF), and 0.88 (NB), with distinct performance trade-offs: RF predicted the largest “Good” PM₁₀ zones (78.98 km²), KNN highlighted the most “UnHealSen” areas (279 km²), and NB predict “Moderate” coverage (311 km²). High PM₁₀ concentrations clustered in eastern and northwestern sectors, aligning with industrial zones and traffic density. The results demonstrate the efficacy of MLAs and geospatial integration in producing high-resolution pollution maps. We advocate for targeted emission controls in hotspots, expanding public transit to reduce vehicular emissions, and incorporating air quality metrics into urban planning. This study advances air quality assessment methods for rapidly urbanizing regions, providing data-driven strategies to combat pollution and enhance ecological resilience in African cities.