PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia

This first pass implementation of the AgWise Fertilizer Recommendation Module for rice in Cambodia assembles fertilizer response trial dataset, calibrates both mechanistic and machine learning models, and generates preliminary maps of indigenous nutrient supply along with prototype fertilizer recomm...

Descripción completa

Detalles Bibliográficos
Autores principales: Mabilangan, Abigail, Mkuhlani, Siyabusa, Saito, Kazuki
Formato: Informe técnico
Lenguaje:Inglés
Publicado: International Rice Research Institute 2025
Materias:
Acceso en línea:https://hdl.handle.net/10568/179040
_version_ 1855517924742135808
author Mabilangan, Abigail
Mkuhlani, Siyabusa
Saito, Kazuki
author_browse Mabilangan, Abigail
Mkuhlani, Siyabusa
Saito, Kazuki
author_facet Mabilangan, Abigail
Mkuhlani, Siyabusa
Saito, Kazuki
author_sort Mabilangan, Abigail
collection Repository of Agricultural Research Outputs (CGSpace)
description This first pass implementation of the AgWise Fertilizer Recommendation Module for rice in Cambodia assembles fertilizer response trial dataset, calibrates both mechanistic and machine learning models, and generates preliminary maps of indigenous nutrient supply along with prototype fertilizer recommendations. A total of 1,909 georeferenced fertilizer response trials across 17 provinces were harmonized with soil, climate, and terrain covariates following the AgWise framework. Analyses reveal higher and more stable yields in the dry season than in the wet season; strong yield responses to nitrogen, a narrow -optimum for phosphorus, and potassium responses that peak at intermediate rates. Reverse and forward -QUEFTS, together with -gradient boosting and random forest models, reproduce observed yields with moderate to good predictive skill and provide -apparent soil N, P, and K supply estimates that can be mapped across the rice domain. These maps show uniformly low P, moderate but heterogeneous N, and high, highly variable K, providing clear evidence for a transition from current fertilizer recommendation toward site- and season--specific nutrient management.- These are explicitly treated as preliminary, pending refinement through incorporation of additional datasets and model re calibration.
format Informe técnico
id CGSpace179040
institution CGIAR Consortium
language Inglés
publishDate 2025
publishDateRange 2025
publishDateSort 2025
publisher International Rice Research Institute
publisherStr International Rice Research Institute
record_format dspace
spelling CGSpace1790402025-12-20T02:03:23Z PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia Mabilangan, Abigail Mkuhlani, Siyabusa Saito, Kazuki machine learning fertilizer technology frameworks innovation adoption nutrient management crop management agronomic practices sustainable agriculture This first pass implementation of the AgWise Fertilizer Recommendation Module for rice in Cambodia assembles fertilizer response trial dataset, calibrates both mechanistic and machine learning models, and generates preliminary maps of indigenous nutrient supply along with prototype fertilizer recommendations. A total of 1,909 georeferenced fertilizer response trials across 17 provinces were harmonized with soil, climate, and terrain covariates following the AgWise framework. Analyses reveal higher and more stable yields in the dry season than in the wet season; strong yield responses to nitrogen, a narrow -optimum for phosphorus, and potassium responses that peak at intermediate rates. Reverse and forward -QUEFTS, together with -gradient boosting and random forest models, reproduce observed yields with moderate to good predictive skill and provide -apparent soil N, P, and K supply estimates that can be mapped across the rice domain. These maps show uniformly low P, moderate but heterogeneous N, and high, highly variable K, providing clear evidence for a transition from current fertilizer recommendation toward site- and season--specific nutrient management.- These are explicitly treated as preliminary, pending refinement through incorporation of additional datasets and model re calibration. 2025-11-28 2025-12-19T04:21:41Z 2025-12-19T04:21:41Z Report https://hdl.handle.net/10568/179040 en Open Access application/pdf International Rice Research Institute Mabilangan, Abigail, Siyabusa Mkuhlani, and Kazuki Saito (2025). PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia. Los Baños, Philippines: International Rice Research Institute. 18 p.
spellingShingle machine learning
fertilizer technology
frameworks
innovation adoption
nutrient management
crop management
agronomic practices
sustainable agriculture
Mabilangan, Abigail
Mkuhlani, Siyabusa
Saito, Kazuki
PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia
title PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia
title_full PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia
title_fullStr PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia
title_full_unstemmed PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia
title_short PROGRESS REPORT: Developing Data‑Driven Agronomic Recommendations for Site‑Specific Nutrient Management on Rice in Cambodia
title_sort progress report developing data driven agronomic recommendations for site specific nutrient management on rice in cambodia
topic machine learning
fertilizer technology
frameworks
innovation adoption
nutrient management
crop management
agronomic practices
sustainable agriculture
url https://hdl.handle.net/10568/179040
work_keys_str_mv AT mabilanganabigail progressreportdevelopingdatadrivenagronomicrecommendationsforsitespecificnutrientmanagementonriceincambodia
AT mkuhlanisiyabusa progressreportdevelopingdatadrivenagronomicrecommendationsforsitespecificnutrientmanagementonriceincambodia
AT saitokazuki progressreportdevelopingdatadrivenagronomicrecommendationsforsitespecificnutrientmanagementonriceincambodia