Evaluating the DNDC Model for Predicting N₂O Emissions in Integrated Crop-Livestock Systems: Insights from Inoculant and Nitrogen Fertilizer Management in Brazil

Integrated crop-livestock systems (ICL) hold significant potential as greenhouse gas sinks in Brazil, offering a promising avenue for mitigating climate change impacts. The DNDC (DeNitrification-DeComposition) model, a robust tool for simulating biogeochemical processes, provides an advanced framework for modelling nitrous oxide (N₂O) emissions. This capability is crucial for assessing the effects of nitrogen (N) management within ICL systems, enabling the optimization of agricultural sustainability by balancing productivity with environmental stewardship.

Field data were obtained from an ICL experiment conducted at the ‘Capivara Experimental Farm’ by Embrapa Rice and Beans, located in Santo Antônio de Goiás, GO, Brazil (16°28´S; 49°17´W; 823 a.s.l.). The ICL experiment was evaluated over four years (2019–2022) using the following crop rotation sequence: common beans (Phaseolus vulgaris) - aerobic rice (Oryza sativa) - forage grass (Urochloa spp). The soil was classified as clayey Ferralsol with 2% organic matter content. All crop phases were conducted under zero tillage.

N₂O emissions were measured using manual static chambers during the bean phase. The experiment included four treatments: Control (No N), Inoculated (No N + Ino), Urea (UR), and Inoculated + Urea (Ino + UR), with four replicates each. N₂O emissions were recorded during 30 sampling events over nearly 70 days throughout the bean cycle. Nitrogen was applied at a rate of 119 kg/ha.

The above treatments were used to parameterize the DNDCv.CAN model, which demonstrated satisfactory performance in predicting N₂O emissions in the ICL system, showing a significant correlation with observed data (r = 0.57, p < 0.001), a MAE of 0.011, and a RMSE of 0.016. The average daily observed N-N₂O fluxes were 0.017 kg ha⁻¹ day⁻¹, compared to 0.012 kg ha⁻¹ day⁻¹ simulated by the DNDC model.

Accumulated N₂O emissions were 0.770, 0.399, 0.808, and 0.991 kg ha⁻¹ for Control, No N+Ino, UR, and Ino+UR, respectively. Simulations by DNDC for these treatments were 0.636 (UR and Ino+UR) and 0.237 kg ha⁻¹ (No N+Ino). In general, the model showed a good fit with daily N₂O fluxes but tended to underestimate accumulated emissions. Moreover, the model requires improvements to more accurately capture the influence of using inoculants. Further model parameterization and calibration is currently in progress to improve predictions. Using inoculants to substitute N significantly reduces N₂O emissions in bean production, enhances soil health, and lowers costs for farmers, contributing to food security. This practice aligns with Brazil’s environmental policy and strengthens its leadership in sustainable agriculture.