Machine learning predictions of climate change effects on nearly threatened bird species (Crithagra xantholaema) habitat in Ethiopia for conservation strategies Tadele Bedo Gelete, Diriba Tulu, Kalid Hassen Yasin, Erana Kebede Scientific Reports, 2025 Endemic and endangered bird species, such as Salvadori serin ( C. xantholaema ), are vulnerable to environmental and anthropogenic changes. Understanding the impact of climate change on ecological niches is essential for effective conservation. This study employed advanced ML algorithms to model the current and future suitability of C. xantholaema under two scenarios (SSP245 and SSP585) for the years 2050 and 2070. The four machine learning models, namely, Maximum Entropy (MaxEnt), Random Forest (RF), Support Vector Machine (SVM), and Extreme Gradient Boost (XGBoost), predicted habitat suitability using 188 presence occurrence data and 15 environmental factors. Model performance was assessed using AUC-ROC, accuracy, precision, sensitivity, specificity, kappa, and F1 score, with ensemble modeling techniques enhancing reliability. The current analysis indicated high predictive accuracy, with XGBoost achieving the highest AUC (0.99), followed by RF (0.98), SVM (0.97), and MaxEnt (0.92). Regarding habitat suitability, 75.3% of Ethiopia’s land was unsuitable for C. xantholaema , with only 3.9% classified as highly suitable. By 2050, 61.82% and 57.14% of areas were projected to be unsuitable under SSP245 and SSP585, respectively. By 2070, unsuitable habitats may increase to 65.24% (SSP245) and 60.17% (SSP585), further decreasing habitat suitability. High-suitability habitats are expected to decline by 80.8% in 2050, covering approximately 8,259.95 km 2 , and by 73.2% in 2070, covering about 11,584.6 km 2 . Precipitation during the driest month (Bio14) was the most crucial predictor of habitat suitability, with importance values ranging from 32.5% (XGBoost) to 100% (SVM and RF), while temperature-related factors, particularly annual mean temperature (Bio1), contributed differently across ML models. According to this study, climate factors impact habitat changes. The findings emphasize the urgent need for conservation strategies to mitigate C. xantholaema habitat loss. Future research should include local data and other human-related factors to enhance the effectiveness of conservation efforts and improve predictions.
Advanced geospatial and machine learning models identify groundwater potential and reveal storage dynamics in Ethiopia's abbay river basin Kalid Hassen Yasin, Tadele Bedo Gelete, Erana Kebede, Anteneh Derribew Iguala, Mohammed Yusuf Abdo Journal of Hydrology Regional Studies, 2025 The Abbay (Blue Nile) River Basin in Ethiopia is a critical sub-basin of the Nile, facing mounting groundwater management challenges due to its complex hydrogeology, which is compounded by climate change, population growth, and agricultural intensification. We developed a hydrologically validated groundwater potential zone (GWPZ) map using four machine learning algorithms—random forest (RF), extremely randomized trees (EXT), support vector machines (SVM), and extreme gradient boosting (XGBoost)–to capture spatial nonlinearity and hydrogeological complexity. Models were trained on 18 environmental predictors and 7100 well/spring locations, balanced with pseudoabsences generated via target-group background sampling in low-potential geomorphological units > 5 km from known water points. To reduce spatial autocorrelation bias, a 5-fold spatial cross-validation was employed. Model performance was evaluated using accuracy, F1 score, log loss, and AUC, with RF achieving the highest predictive accuracy (91 %) and rainfall as the dominant predictor. Spatial patterns revealed high-potential zones in the northeast and low-potential zones in the northwest and south. The ML-delineated high-potential zones demonstrated remarkable congruence with GRACE/GLDAS-derived groundwater storage trends, revealing significant recharge (+4.41 mm· yr⁻¹, 2003–2023) without dataset integration. This independent validation, emerging from methodologically distinct approaches, robustly confirmed the active recharge dynamics of the basin. By leveraging ML alongside satellite hydrology, we established a scalable framework for resolving hydrogeological complexities in data-scarce regions, with direct implications for evidence-based groundwater governance and regional water security. • Random Forest algorithms achieve 91 % accuracy in groundwater potential mapping. • Rainfall is identified as the key predictor in the model. • GRACE/GLDAS data integration reveals a 2.3 cm/year groundwater storage increase post-2015 infrastructure changes. • Machine learning algorithm enhances prediction accuracy in complex hydrogeological systems.
Methodological Integration of Machine Learning and Geospatial Analysis for PM10 Pollution Mapping Kalid Hassen Yasin, Muaz Ismael Yasin, Anteneh Derribew Iguala, Tadele Bedo Gelete, Erana Kebede Methodsx, 2025 Air pollution mitigation necessitates accurate spatial modelling to inform public health interventions. Traditional approaches inadequately capture complex predictor-pollutant interactions, whereas machine learning (ML) offers a superior capacity for modelling nonlinear relationships. This study compares three ML Random Forest (RF), K-Nearest Neighbors (KNN), and Naïve Bayes (NB) algorithms using annual PM 10 data from 11 monitoring stations alongside atmospheric, urban, and terrain covariates. The methodological framework employed rigorous preprocessing and cross-validation to classify pollution into three categorical levels. Results demonstrate RF superior performance, achieving 94% balanced accuracy and 97% specificity, significantly outperforming KNN (92%) and NB (89%). RF excelled in capturing spatial heterogeneity and complex variable interactions, while KNN and NB exhibited limitations in managing feature dependencies and localized variability. Despite computational demands, findings substantiate RF reliability for robust air quality monitoring applications. The study contributes valuable insights for implementing scalable pollution prediction systems in resource-constrained urban environments while acknowledging interpretability challenges inherent to complex ML models. • Preprocessing of spatial data from various sources, incorporating the handling of missing/abnormal data, analysis, and normalization • Implementation of the three ML algorithms with rigorous hyperparameter tuning, model validation, and performance assessment • Mapping PM 10 Hotspots on the Gradient Direction and Distance from the City Center
Predictive machine learning and geospatial modeling reveal PM10 hotspots and guide targeted air pollution interventions in Addis Ababa, Ethiopia Kalid Hassen Yasin, Muaz Ismael Yasin, Anteneh Derribew Iguala, Tadele Bedo Gelete, Diriba Tulu, Erana Kebede Discover Applied Sciences, 2025 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.
Optimum plant density and inorganic fertilizer application improved selected soil chemical properties and common bean productivity in southern Ethiopia Demissie Alemayehu, Deressa Shumi, Erana Kebede, Nano Alemu Daba, Nigussie Dechassa Agrosystems Geosciences and Environment, 2025 Poor soil fertility and inappropriate plant density are the major factors that constrain the productivity of common bean (Phaseolus vulgaris L.) in tropical Africa, including Ethiopia. This problem necessitates improving soil fertility and optimizing agronomic practices. Therefore, we conducted field experiments from 2019 to 2021, integrating plant density and multinutrient fertilizer application to improve soil properties and common bean productivity in southern Ethiopia. The treatments included four plant densities (333,300 plants ha−1, 250,000 plants ha−1, 200,000 plants ha−1, and 166,600 plants ha−1) and five fertilizer rates (0, 50, 100, 150, and 200 kg NPS ha−1). The application of NPS fertilizer reduced the soil pH while increasing the soil organic carbon, total nitrogen, and available sulfur and phosphorus contents but did not affect the cation exchange capacity. Similarly, at the lowest plant density, the available soil sulfur and cation exchange capacity improved. Increasing the NPS application increased common bean growth and yield components, particularly when the plant density was the lowest. An optimum grain yield of 3056.28 kg ha−1 was obtained with the application of 150 kg NPS ha−1 and a plant density of 200,000 plants ha−1, with a net return of 80,132.56 ETB ha−1 and a marginal return rate of 4169.10%. It was concluded that applying 150 kg of NPS at a common bean plant density of 200,000 ha−1 resulted in an optimum grain yield. Using the stated amount of NPS and optimizing the density in the study area, smallholder farmers can improve common bean productivity and soil organic carbon, total nitrogen, sulfur, and phosphorus availability.
Green manure substitution for chemical nitrogen reduces greenhouse gas emissions and enhances yield and nitrogen uptake in rice[sbnd]rice cropping systems Nano Alemu Daba, Jing Huang, Zhe Shen, Tianfu Han, Md Ashraful Alam, Jiwen Li, Kiya Adare Tadesse, Ntagisanimana Gilbert, Erana Kebede, Tsegaye Gemechu Legesse, Shujun Liu, Lisheng Liu, Kailou Liu, Huimin Zhang Field Crops Research, 2025 Although nitrogen (N) is important for rice growth, its excessive use can have negative environmental effects, such as greenhouse gas (GHG) emissions. Thus, sustainable and eco-friendly rice production demands precise N management strategies. This includes the use of milk-vetch (MV) as a green manure (GM) for substitution. However, how GM substitution for chemical N fertilizer (NF) affects yield, uptake, methane (CH 4 ) emissions, nitrous oxide (N 2 O) emissions and related microbial mechanisms in rice rice cropping systems remains poorly understood. The present study aimed to (i) investigate the effects of MV substitution for NF on grain yield, N uptake, and emissions of CH 4 , N 2 O, and GHG intensity; (ii) comparatively analyze the mechanistic effects of major microbial associated with CH 4 and N 2 O emissions under MV substitution for NF; and (iii) identify the optimal substitution level of NF by MV for mitigating GHG emission intensity while improving crop N uptake and yield in rice rice cropping systems. To address the aforementioned knowledge gap, we conducted a two-year field experiment based on a long-term study established in 2008. Six treatments, i.e., no fertilizer (N0), farmers’ N practice (N100), N100 and MV (N100MV), 80 % N100 and MV (N80MV), 60 % N100 and MV (N60MV) and only MV, were set up in a randomized complete block design in triplicate. Compared with the other treatments, N80MV significantly increased early and late rice yields, with its average N uptake exceeding that of N100, N100MV, N60MV, and MV by 126.3 %, 88.3 %, 54.2 %, and 31.5 %, respectively. The relative yield was strongly related to the N nutrition index (NNI), with the highest mean NNI values of 1.08 and 1.01 observed in N80MV during the early and late rice seasons, respectively. These findings identify N80MV as the optimal fertilization strategy for increasing both N nutrition and productivity. The balance between the mcr A and pmo A genes as well as between carbon (C) and N played a major role in explaining the variation in CH 4 emissions, whereas ammonia oxidation , the C:N ratio, available N, and the nir K gene played key roles in controlling N 2 O emissions. The moderate GWP and relatively high grain yield resulting from N80MV led to the mitigation of GHG emission intensity. The effectiveness of MV substitution for NF in mitigating GHG emissions while improving yield and N uptake in rice rice cropping systems can vary considerably on the basis of the NF levels substituted by MV. We suggest that substituting MV for 20 % N100 is a viable fertilization strategy not only for mitigating the GHG intensity but also for simultaneously improving yield and N uptake in rice rice cropping systems. Our findings have direct implications for extending our understanding of the dynamics of CH 4 and N 2 O emissions, along with their associated drivers, when GM substitutes for NF in rice rice cropping systems. • CH 4 emissions most driven by C/N and mcr A/ pmo A ratios. • N 2 O emissions were primarily controlled by Ammonia oxidation, available N, and nir K gene. • Full milk-vetch substitution raised CH 4 and lowered N 2 O; farmers’ N practice did opposite. • Milk-vetch substituted for 20 % chemical N-fertilizer decreased greenhouse gas intensity, while increasing yield and N-uptake.
Random Forest-Based Species Distribution Modeling Reveals Intensifying Multi-Species Invasion Risks of Alien Plants in Ethiopia Under Climate Change KH Yasin, D Tulu, TB Gelete, BA Yuya, AD Iguala, KA Tadesse, E Kebede Remote Sensing Applications: Society and Environment, 101869 , 2026 2026 Citations: 1
Green manure and rice straw recycling: A triple-win for productivity, environmental sustainability and net ecosystem economic benefit NA Daba, J Huang, MA Alam, N Gilbert, KA Tadesse, I Ahmed, ... Journal of Environmental Management 397, 128381 , 2026 2026
Production practices and agronomic approaches of Khat (Catha edulis Forsk) in eastern Ethiopia A Hassen, Z Bekeko, A Mohammed, M Goftishu, T Fite, E Kebede, ... 2025
Machine learning predictions of climate change effects on nearly threatened bird species ( Crithagra xantholaema) habitat in Ethiopia for conservation strategies TB Gelete, D Tulu, KH Yasin, E Kebede Scientific Reports 15 (1), 36972 , 2025 2025 Citations: 2
Advanced geospatial and machine learning models identify groundwater potential and reveal storage dynamics in Ethiopia’s Abbay River basin KH Yasin, TB Gelete, E Kebede, AD Iguala, MY Abdo Journal of Hydrology: Regional Studies 61, 102762 , 2025 2025 Citations: 2
Green manure substitution reduces carbon and nitrogen footprints and improves net ecosystem economic benefits in double rice systems NA Daba, J Huang, MA Alam, J Li, Z Shen, KA Tadesse, N Gilbert, T Han, ... Journal of Cleaner Production 521, 146266 , 2025 2025 Citations: 7
Associations of angular leaf spot ( Pseudocercospora griseola ) epidemics and yield losses in common bean as influenced by integration of variety and fungicide … GG Mengesha, H Terefe, G Yitayih, E Kebede, A Abera, A Jambo, ... Journal of Plant Pathology 107 (3), 1343-1361 , 2025 2025 Citations: 2
Methodological integration of machine learning and Geospatial analysis for PM10 pollution mapping KH Yasin, MI Yasin, AD Iguala, TB Gelete, E Kebede MethodsX 14, 103322 , 2025 2025 Citations: 5
Predictive machine learning and geospatial modeling reveal PM 10 hotspots and guide targeted air pollution interventions in Addis Ababa, Ethiopia KH Yasin, MI Yasin, AD Iguala, TB Gelete, D Tulu, E Kebede Discover Applied Sciences 7 (4), 263 , 2025 2025 Citations: 11
Optimum plant density and inorganic fertilizer application improved selected soil chemical properties and common bean productivity in southern Ethiopia D Alemayehu, D Shumi, E Kebede, NA Daba, N Dechassa Agrosystems, Geosciences & Environment 8 (1), e70079 , 2025 2025 Citations: 2
Green manure substitution for chemical nitrogen reduces greenhouse gas emissions and enhances yield and nitrogen uptake in rice-rice cropping systems NA Daba, J Huang, Z Shen, T Han, MA Alam, J Li, KA Tadesse, N Gilbert, ... Field Crops Research 322, 109715 , 2025 2025 Citations: 22
Optimal interpolation approach for groundwater depth estimation KH Yasin, TB Gelete, AD Iguala, E Kebede MethodsX 13, 102916 , 2024 2024 Citations: 20
Biostimulants for climate-smart and sustainable agriculture M Baslam, M Anli, JS Patel, DL Smith Frontiers in Sustainable Food Systems, 102 , 2024 2024
Vermicompost and bactericide application minimized common bacterial blight development and enhanced nodulation and agronomic performances of bean varieties in Southern Ethiopia H Terefe, GG Mengesha, A Abera, E Kebede, G Yitayih Agrosystems, Geosciences & Environment 7 (1), e20465 , 2024 2024 Citations: 6
Integrated Machine Learning and Geospatial Analysis Enhanced Gully Erosion Susceptibility Modeling in the Erer Watershed in Eastern Ethiopia TB Gelete, P Pasala, NG Abay, GW Woldemariam, KH Yasin, E Kebede, ... Frontiers in Environmental Science 12, 1410741 , 2024 2024 Citations: 27
Nitrogen use to improve sustainable yields in agricultural systems E Kebede Nitrogen use to improve sustainable yields in agricultural systems, 16 , 2023 2023 Citations: 2
Integrating multiple soil management practices: A system‐wide approach for restoring degraded soil and improving Brachiaria productivity T Gutema, E Kebede, H Legesse, T Fite Agrosystems, Geosciences & Environment 6 (2), e20360 , 2023 2023 Citations: 7
Sugarcane productivity and sugar yield improvement: Selecting variety, nitrogen fertilizer rate, and bioregulator as a first-line treatment B Desalegn, E Kebede, H Legesse, T Fite Heliyon 9 (4) , 2023 2023 Citations: 67
Endophytic fungi: versatile partners for pest biocontrol, growth promotion, and climate change resilience in plants T Fite, E Kebede, T Tefera, Z Bekeko Frontiers in Sustainable Food Systems 7, 1322861 , 2023 2023 Citations: 49
Contribution of climate‐smart forage and fodder production for sustainable livestock production and environment: Lessons and challenges from Ethiopia D Tulu, S Gadissa, F Hundessa, E Kebede Advances in agriculture 2023 (1), 8067776 , 2023 2023 Citations: 74
MOST CITED SCHOLAR PUBLICATIONS
Contribution, Utilization, and Improvement of Legumes-Driven Biological Nitrogen Fixation in Agricultural Systems E Kebede Frontiers in Sustainable Food Systems 5 (767998), 1-18 , 2021 2021 Citations: 484
Grain legumes production and productivity in Ethiopian smallholder agricultural system, contribution to livelihoods and the way forward E Kebede Cogent Food & Agriculture 6 (1), 1722353 , 2020 2020 Citations: 191
Expounding the production and importance of cowpea ( Vigna unguiculata (L.) Walp.) in Ethiopia E Kebede, Z Bekeko Cogent Food & Agriculture 6 (1), 1769805 , 2020 2020 Citations: 188
Contribution of climate‐smart forage and fodder production for sustainable livestock production and environment: Lessons and challenges from Ethiopia D Tulu, S Gadissa, F Hundessa, E Kebede Advances in agriculture 2023 (1), 8067776 , 2023 2023 Citations: 74
Competency of rhizobial inoculation in sustainable agricultural production and biocontrol of plant diseases E Kebede Frontiers in Sustainable Food Systems 5, 728014 , 2021 2021 Citations: 70
Sugarcane productivity and sugar yield improvement: Selecting variety, nitrogen fertilizer rate, and bioregulator as a first-line treatment B Desalegn, E Kebede, H Legesse, T Fite Heliyon 9 (4) , 2023 2023 Citations: 67
Endophytic fungi: versatile partners for pest biocontrol, growth promotion, and climate change resilience in plants T Fite, E Kebede, T Tefera, Z Bekeko Frontiers in Sustainable Food Systems 7, 1322861 , 2023 2023 Citations: 49
Grain legumes production in Ethiopia: A review of adoption, opportunities, constraints and emphases for future interventions EK Neda Turkish Journal of Agriculture-Food Science and Technology 8 (4), 977-989 , 2020 2020 Citations: 46
Contribution, utilization, and improvement of legumes-driven biological nitrogen fixation in agricultural systems. Front Sustain Food Syst 5: 767998 E Kebede Frontiers in Sustainable Food Systems , 2021 2021 Citations: 35
Integrated Machine Learning and Geospatial Analysis Enhanced Gully Erosion Susceptibility Modeling in the Erer Watershed in Eastern Ethiopia TB Gelete, P Pasala, NG Abay, GW Woldemariam, KH Yasin, E Kebede, ... Frontiers in Environmental Science 12, 1410741 , 2024 2024 Citations: 27
Symbiotic effectiveness of cowpea ( Vigna unguiculata (L.) Walp.) nodulating rhizobia isolated from soils of major cowpea producing areas in Ethiopia E Kebede, B Amsalu, A Argaw, S Tamiru Cogent Food & Agriculture 6 (1), 1763648 , 2020 2020 Citations: 27
Green manure substitution for chemical nitrogen reduces greenhouse gas emissions and enhances yield and nitrogen uptake in rice-rice cropping systems NA Daba, J Huang, Z Shen, T Han, MA Alam, J Li, KA Tadesse, N Gilbert, ... Field Crops Research 322, 109715 , 2025 2025 Citations: 22
Optimal interpolation approach for groundwater depth estimation KH Yasin, TB Gelete, AD Iguala, E Kebede MethodsX 13, 102916 , 2024 2024 Citations: 20
Eco-physiological and physiological characterization of cowpea nodulating native rhizobia isolated from major production areas of Ethiopia E Kebede, B Amsalu, A Argaw, S Tamiru Cogent Biology 6 (1), 1875672 , 2020 2020 Citations: 13
Predictive machine learning and geospatial modeling reveal PM 10 hotspots and guide targeted air pollution interventions in Addis Ababa, Ethiopia KH Yasin, MI Yasin, AD Iguala, TB Gelete, D Tulu, E Kebede Discover Applied Sciences 7 (4), 263 , 2025 2025 Citations: 11
Abundance of native rhizobia nodulating cowpea in major production areas of Ethiopia as influenced by cropping history and soil properties E Kebede, B Amsalu, A Argaw, S Tamiru Sustainable Environment 7 (1), 1889084 , 2021 2021 Citations: 10
Green manure substitution reduces carbon and nitrogen footprints and improves net ecosystem economic benefits in double rice systems NA Daba, J Huang, MA Alam, J Li, Z Shen, KA Tadesse, N Gilbert, T Han, ... Journal of Cleaner Production 521, 146266 , 2025 2025 Citations: 7
Integrating multiple soil management practices: A system‐wide approach for restoring degraded soil and improving Brachiaria productivity T Gutema, E Kebede, H Legesse, T Fite Agrosystems, Geosciences & Environment 6 (2), e20360 , 2023 2023 Citations: 7
Nodulation potential and phenotypic diversity of rhizobia nodulating cowpea isolated from major growing areas of Ethiopia E Kebede, B Amsalu, A Argaw, S Tamiru Agrosystems, Geosciences & Environment 5 (4), e20308 , 2022 2022 Citations: 7
Vermicompost and bactericide application minimized common bacterial blight development and enhanced nodulation and agronomic performances of bean varieties in Southern Ethiopia H Terefe, GG Mengesha, A Abera, E Kebede, G Yitayih Agrosystems, Geosciences & Environment 7 (1), e20465 , 2024 2024 Citations: 6
