Organic amendments for jointly managing greenhouse-gas emissions and crop productivity in dryland croplands of the yellow river basin: evidence from meta-analysis and machine learning

Le résumé fourni par la source

Replacing mineral nitrogen (N) with organic materials is increasingly promoted, but its effects on greenhouse gas mitigation and crop production in dryland croplands of the Yellow River Basin have not been resolved. We synthesized 62 field studies by meta-analysis and used boosted regression tree (BRT) models to assess organic fertilizer, biochar, and straw return effects on CO 2 , N 2 O, CH 4 emissions, and maize and wheat yields. Relative to mineral N alone, organic fertilizer raised CO 2 by 8.72% but decreased N 2 O and CH 4 by 19.08% and 13.78%. Biochar decreased CO 2 and N 2 O by 10.08% and 16.57%, respectively, but increased CH 4 by 57.76%. Straw return elevated CO 2 and N 2 O by 22.58% and 26.53%, respectively; its apparent 4.87% CH 4 reduction was not robust. All amendments increased yield: straw return produced the largest maize gain (9.44%), whereas organic fertilizer produced the largest wheat gain (9.13%). Subgroup responses varied with site and management conditions. BRT models identified organic amendment-derived N input (OrN) and application duration as the main management controls, with precipitation and temperature especially important for crop-specific outcomes. Predicted yields under organic fertilizer were relatively high at about 120–170 kg − ha −1 OrN for maize and 100–160 kg·ha −1 for wheat. Under straw return, predicted CO 2 rose rapidly during the first 4 years and maize yield peaked near year 4. Biochar provided the clearest combination of yield enhancement and CO 2 and N 2 O mitigation. Organic fertilizer was most favorable for wheat, whereas straw return maximized maize yield but imposed the largest CO 2 and N 2 O penalties. These findings support amendment-, crop-, and region-specific management.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.