Conservation agriculture reduces climate change impact of a popcorn and wheat crop rotation

generates LCI files ready to be imported into LCA software (here we used Simapro, see 2.1.3) to calculate impact indicators. Databases used for background processes (e.g. production of fertilizer, irrigation infrastructure), were AGRIBALYSE v3.0.1 and ecoinvent v3.5. Emissions to the air (ammonia (NH3), nitrous oxide (N2O), N oxides (NOx), CO2), water (NO3, phosphorus (P), phosphate (PO4), cadmium (Cd), chromium (Cr), copper (Cu), mer- cury (Hg), nickel (Ni), lead (Pb), zinc (Zn)) and soil (Cd, Cr, Cu, Hg, Ni, Pb, Zn, pesticides) were calculated using models recommended by the AGRIBALYSE methodology [22]. NH3 emissions from organic fertiliser application were modelled using EMEP/EEA 2016 Tier 2 [23]. N2O emissions were modelled using IPCC 2019 Tier 1 [24]. NOx emissions were mod- elled using EMEP/EEA 2009 Tier 1 [25]. NO3 leaching was modelled according to Tailleur et al. [26]. CO2 emissions from fuel combustion and emissions of active substances of pesti- cides were modelled according to ecoinvent1 v2 [27]. Composting of green waste consists of the following processes: shredding, stabilisation, mixing, composting and screening [28]. The compost of green waste LCI available in the AGRIBALYSE database [28] attributes all polluting emissions of the composting process to the compost, and thus to the crop on which the compost is applied. From a methodological point of view, this is questionable, as the composting process has two functions: waste treatment and compost production. According to the LCI for compost in the AGRIBALYSE database, pro-

ducing 1 t of compost requires 3 t of green waste. The cost of treating green waste by compost- ing was assumed to be 60 €/t [29] (i.e. 180 € per t of compost), and the price of the green waste compost delivered to the Nataïs farm was 14 €/t. Given a total cost of 1 t of compost of 194 €/t, 7.2% of it (i.e. 14/194×100) was for the compost and the rest for the waste treatment process. Consequently, following an approach proposed by Christensen et al. [30], we used economic allocation to allocate pollutant emissions and resource use between the two functions: 7.2% to compost production and 92.8% to waste treatment. 2.1.3 Impact assessment: LCA impact categories. Impacts were assessed using SimaPro software v9.0.0.35. We calculated climate change, terrestrial acidification, and marine eutro- phication impacts according to the ReCiPe 2016 Midpoint (H) characterisation method [31]. We also calculated land competition impact according to the CML-IA non-baseline method [32] and cumulative energy demand [33]. 2.2 Estimating changes in soil organic carbon stock Most LCA studies of agricultural production systems assume that farming practices remain constant over time, and thus that the SOC stock remains constant. This assumption is often reasonable, but not when considering a change in farming practices that increases the input of organic C to the soil, such as the introduction of cover crops and organic fertiliser. When such practices are implemented, the increase in SOC stock needs to be included in the LCI, as it rep-

resents an absorption of CO2 from the atmosphere (i.e. a negative emission). The change in SOC stock for the comparison of the two rotations was estimated using the AMG model [34], which considers three compartments: fresh exogenous organic C, active SOC, and stable SOC. The sources of fresh organic C are organic fertilisers and crop residues. The active SOC compartment is fed by fresh organic C inputs and influenced by annual miner- alisation, while stable SOC is considered to be completely inert over the short and mid-terms (i.e. its turnover time is millenary). The AMG model represents the generally accepted fact that the SOC pool is heterogeneous. The AMG model has three main parameters: the humifi- cation coefficient, which is the conversion factor from fresh organic C inputs into humified SOC; the annual rate of SOC mineralisation; and the initial proportion of stable C in the initial SOC stock. The humification coefficient depends only on the nature of fresh organic C inputs, PLOS ONE Conservation agriculture reduces climate change impact of a crop rotation PLOS ONE | https://doi.org/10.1371/journal.pone.0285586 May 23, 2023 6 / 18