Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia
This study evaluates the broader potential of a Next-Generation Decision Support Tool (DST) for delivering data-driven, season-tailored, site-specific fertilizer recommendations (SSRs) to enhance maize performance in Ethiopia. The DST integrates a national maize yield response to fertilizer database...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Informe técnico |
| Language: | Inglés |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10568/178470 |
| _version_ | 1855525931254284288 |
|---|---|
| author | Mesfin, Tewodros Abera, Wuletawu Ebrahim, Mohammed Tilaye, Amsalu Tibebe, Degefie Tamene, Lulseged Adimassu, Zenebe Liben, Feyera Bekere, Jemal Nazib, Nibras |
| author_browse | Abera, Wuletawu Adimassu, Zenebe Bekere, Jemal Ebrahim, Mohammed Liben, Feyera Mesfin, Tewodros Nazib, Nibras Tamene, Lulseged Tibebe, Degefie Tilaye, Amsalu |
| author_facet | Mesfin, Tewodros Abera, Wuletawu Ebrahim, Mohammed Tilaye, Amsalu Tibebe, Degefie Tamene, Lulseged Adimassu, Zenebe Liben, Feyera Bekere, Jemal Nazib, Nibras |
| author_sort | Mesfin, Tewodros |
| collection | Repository of Agricultural Research Outputs (CGSpace) |
| description | This study evaluates the broader potential of a Next-Generation Decision Support Tool (DST) for delivering data-driven, season-tailored, site-specific fertilizer recommendations (SSRs) to enhance maize performance in Ethiopia. The DST integrates a national maize yield response to fertilizer database, geo-referenced soil profiles, spatial covariates, and a Random Forest algorithm to generate optimized nitrogen–phosphorus (N–P) rates at 250 m resolution. During the 2024 season, its performance was validated through on-farm trials on 59 fields across Meskan, Mareko, and Shala districts, comparing three fertilization strategies: DST-based SSR, district-level local recommendations (LR), and farmers’ current practices (FP). A linear mixed-effects model was used to assess treatment effects on grain and biomass yield, while partial fertilizer profit (PFP) and empirical distribution analyses quantified economic outcomes and risk profiles.
Results showed that grain yield was significantly affected by both location and fertilizer strategy, with SSR consistently outperforming LR and frequently FP in Mareko and Shala, where maize was highly responsive to tailored nutrient management. In Meskan, yield responses were limited across all treatments, highlighting the influence of non-nutrient constraints. Biomass yield differences among treatments were not statistically significant, suggesting that the primary agronomic gains from SSR relate to improved grain production rather than total biomass. Economic analyses demonstrated clear profitability advantages of SSR over LR in all districts and over FP in Mareko and Shala, albeit with heterogeneous outcomes at field level, especially in Shala. Farmer testimonies corroborated these findings, emphasizing improved yields, input efficiency, and perceived relevance of field-specific advice. Collectively, the results demonstrate that ML-enabled SSRs, delivered via a robust DST, offer a scalable, climate-smart pathway to increase maize productivity and fertilizer profitability in Ethiopia, provided they are accompanied by context-specific agronomic support and enabling policies for digital, data-driven advisory systems. |
| format | Informe técnico |
| id | CGSpace178470 |
| institution | CGIAR Consortium |
| language | Inglés |
| publishDate | 2025 |
| publishDateRange | 2025 |
| publishDateSort | 2025 |
| record_format | dspace |
| spelling | CGSpace1784702025-12-17T02:08:18Z Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia Mesfin, Tewodros Abera, Wuletawu Ebrahim, Mohammed Tilaye, Amsalu Tibebe, Degefie Tamene, Lulseged Adimassu, Zenebe Liben, Feyera Bekere, Jemal Nazib, Nibras machine learning digital agriculture site-specific nutrient management economic benefits decision-support systems-decision support tools This study evaluates the broader potential of a Next-Generation Decision Support Tool (DST) for delivering data-driven, season-tailored, site-specific fertilizer recommendations (SSRs) to enhance maize performance in Ethiopia. The DST integrates a national maize yield response to fertilizer database, geo-referenced soil profiles, spatial covariates, and a Random Forest algorithm to generate optimized nitrogen–phosphorus (N–P) rates at 250 m resolution. During the 2024 season, its performance was validated through on-farm trials on 59 fields across Meskan, Mareko, and Shala districts, comparing three fertilization strategies: DST-based SSR, district-level local recommendations (LR), and farmers’ current practices (FP). A linear mixed-effects model was used to assess treatment effects on grain and biomass yield, while partial fertilizer profit (PFP) and empirical distribution analyses quantified economic outcomes and risk profiles. Results showed that grain yield was significantly affected by both location and fertilizer strategy, with SSR consistently outperforming LR and frequently FP in Mareko and Shala, where maize was highly responsive to tailored nutrient management. In Meskan, yield responses were limited across all treatments, highlighting the influence of non-nutrient constraints. Biomass yield differences among treatments were not statistically significant, suggesting that the primary agronomic gains from SSR relate to improved grain production rather than total biomass. Economic analyses demonstrated clear profitability advantages of SSR over LR in all districts and over FP in Mareko and Shala, albeit with heterogeneous outcomes at field level, especially in Shala. Farmer testimonies corroborated these findings, emphasizing improved yields, input efficiency, and perceived relevance of field-specific advice. Collectively, the results demonstrate that ML-enabled SSRs, delivered via a robust DST, offer a scalable, climate-smart pathway to increase maize productivity and fertilizer profitability in Ethiopia, provided they are accompanied by context-specific agronomic support and enabling policies for digital, data-driven advisory systems. 2025-11-17 2025-12-03T11:04:50Z 2025-12-03T11:04:50Z Report https://hdl.handle.net/10568/178470 en Open Access application/pdf Mesfin, T.; Abera, W.; Ebrahim, M.; Tilaye, A.; Tibebe, D.; Tamene, L.; Adimassu, Z.; Liben, F.; Bekere, J.; Nazib, N. (2025) Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia. 21 p. |
| spellingShingle | machine learning digital agriculture site-specific nutrient management economic benefits decision-support systems-decision support tools Mesfin, Tewodros Abera, Wuletawu Ebrahim, Mohammed Tilaye, Amsalu Tibebe, Degefie Tamene, Lulseged Adimassu, Zenebe Liben, Feyera Bekere, Jemal Nazib, Nibras Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia |
| title | Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia |
| title_full | Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia |
| title_fullStr | Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia |
| title_full_unstemmed | Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia |
| title_short | Technical report: Next-Generation decision support tool for data-driven fertilizer recommendations to enhance maize performance in Ethiopia |
| title_sort | technical report next generation decision support tool for data driven fertilizer recommendations to enhance maize performance in ethiopia |
| topic | machine learning digital agriculture site-specific nutrient management economic benefits decision-support systems-decision support tools |
| url | https://hdl.handle.net/10568/178470 |
| work_keys_str_mv | AT mesfintewodros technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT aberawuletawu technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT ebrahimmohammed technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT tilayeamsalu technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT tibebedegefie technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT tamenelulseged technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT adimassuzenebe technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT libenfeyera technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT bekerejemal technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia AT nazibnibras technicalreportnextgenerationdecisionsupporttoolfordatadrivenfertilizerrecommendationstoenhancemaizeperformanceinethiopia |