Performance of the DeNitrification-DeComposition model in simulating agronomic properties and N2O emissions under smallholder climate-smart farming systems

Publication Type

Journal Article

Journal Name

Climate Smart Agriculture

Name of Author

• Esphorn Kibet, Jaramogi Oginga Odinga University of Science and Technology
• Collins M. Musafiri, Cortile Scientific Limited
• Milka Kiboi, Forschungsinstitut für biologischen Landbau
• Onesmus K. Ng'etich, University of Embu
• David K. Kosgei, Moi University
• Abdirahman Zeila, The World Bank, USA
• Franklin Mairura, University of Embu
• Felix K. Ngetich, Cortile Scientific Limited

Publication Date

8-1-2026

Abstract

The DeNitrification-DeComposition (DNDC) model is a crucial tool for estimating soil greenhouse gas fluxes and understanding soil-plant interactions. This study evaluates the performance of the DNDC model in simulating soil temperature, moisture, crop yield, and nitrous oxide (N2O) fluxes across different land utilization types in Western Kenya. The land utilization types included: i) agroforestry M (agroforestry with Markhamia lutea, ii) sole sorghum, iii) agroforestry L (agroforestry with Leucaena leucocephala), iv) sole maize, and v) grazing Land, each replicated thrice. Fertilizer and manure were applied as part of the management practices, with manure applied at 2t ha⁻¹. Soil greenhouse gas samples were collected using vented static chambers and analyzed via gas chromatography. Sensitivity analysis was conducted by varying soil pH, soil organic carbon, clay content, bulk density, and nitrates. The model was calibrated and validated using input data from grazing land. Model performance was assessed using mean error, root mean square error, normalized root mean square error, Nash-Sutcliffe efficiency, and the index of agreement (d). Across land utilization types, agroforestry systems showed higher soil organic carbon (2.07–2.23%) compared to sole cropping systems (0.61–1.50%), while daily field-measured N2O fluxes were also elevated under agroforestry (1.75–2.55 g N2O–N ha⁻¹ day⁻¹) compared to sole maize (0.56 g N2O–N ha⁻¹ day⁻¹). The DNDC model demonstrated high sensitivity to soil pH and bulk density, with good fit in simulating daily soil temperature and moisture. However, the model showed variable performance in simulating N2O emissions and crop yields. The model performed well in predicting soil temperature (0.98 ≤ R² ≤ 0.99, 0.91 ≤ d > 0.99), moderately well for soil moisture (0.88 ≤ R² ≤ 0.99, 0.56 ≤ d ≤ 0.94), and had mixed results for crop yields (0.4 ≤ R² ≤ 0.9). The model showed poor to moderate performance for simulating N2O fluxes (0.4 ≤ R² ≤ 0.66, 0.87 ≤ d ≤ 0.97), suggesting it can be a useful tool for estimating N2O emissions and improving reporting of Nationally Determined Contributions. The inclusion of manure and fertilizer application as management practices highlights the importance of nutrient management in influencing soil greenhouse gas fluxes and crop productivity.

Keywords

Climate change, Net sink, Soil biogeochemical processes, Soil-atmosphere exchange

Recommended Citation

Kibet, E., Musafiri, C., Kiboi, M., Ng'etich, O., Kosgei, D., Zeila, A., Mairura, F., & Ngetich, F. (2026). Performance of the DeNitrification-DeComposition model in simulating agronomic properties and N2O emissions under smallholder climate-smart farming systems. Climate Smart Agriculture, 3 (3) https://doi.org/10.1016/j.csag.2026.100126