MEASURING AND MODELLING COVER CROP GROWTH AND AGRONOMIC EFFECTS

Cover crops are cultivated during the bare soil period between the harvest of a cash crop and the sowing of the next one. Their cultivation puts into effect the permanent soil organic cover principle of conservation agriculture and exerts several agro-ecological services, among which the most relevant are nitrate leaching reduction, weed growth control, soil organic matter increase, soil structure and water infiltration improvement. In temperate climates when crop rotations include summer cash crops (such as maize or soybean), autumn-winter cover crops are sown between late July and October and terminated from March to April of the following year. When sown in autumn, frost-sensitive cover crops may also be terminated efficiently by frost damage: this termination method is frequently called ‘winterkill’. Black oat (Avena strigosa Schreb.) and white mustard (Sinapis alba L.) are two of the most interesting and widespread frost-sensitive cover crops due to their adaptability to various environmental conditions and cropping systems. Even if these species are widely adopted as cover crops, there is a lack of information concerning both crop management and agronomic effects, as well as winterkill termination occurrence frequency and efficiency in temperate climates. Dynamic cropping systems simulation models can be used to determine crop management scenarios convenience for a wide range of weather and soil conditions, while the field trial assessments require large resource investments. However, the application of a simulation model to white mustard and black oat cover crops presents several knowledge gaps, as the limited number of studies focused on winterkilled cover crops growth and agronomic effects carried out in northern Italy. Furthermore, an integrated simulation model dealing both with cover crop growth and winterkill termination, and its consequent effect on the crop-soil system, including cover crop residue degradation on soil surface, is lacking. This work aimed at representing, within the ARMOSA cropping system model framework, frost sensitive cover crop species growth, development, and agronomic effects by enriching the simulation model with two additional modules, dealing respectively with winterkill events and cover crop superficial residue decomposition. The model was calibrated for white mustard and black oat cover crops using experimental data deriving both from a three-year field trial, from a commercial field monitoring campaign and from previous experiments, carried out in the region of interest. During the three-year field trial, white mustard, black oat and their mixture with purple vetch (Vicia benghalensis L.) have demonstrated a good aboveground biomass production potential (2-3 t DM ha-1), particularly when planted before the first half of September. Their nitrogen uptake (45 kg N ha-1 on average, up to 148 kg N ha-1) follows the biomass accumulation patterns, while their weed species control ability has proven to be consistently high. Overall, the improved ARMOSA model correctly simulated these species development (RRMSE equal to 27.3 and 29.5% respectively for black oat and white mustard), as well as soil water content and temperature (RRMSE equal to 8.4% and 19.2%). The employment of the new ARMOSA version to simulate black oat and white mustard cultivation, generally improved significantly both aboveground biomass simulation (RRMSE was decreased by 56.3% in comparison to the use of the original model version), leaf area index (RRMSE reduction of 31.6%) and C:N ratio simulations (RRMSE reduction of 8.8%). The convenience of the new model version employment was assessed in a wide range of sites (six sites of several provinces of Lombardy region in northern Italy), pedological conditions (soil textures from sandy-loam to silty-clay), weather conditions (calibration seasons ranged from 2019/2020 to 2021/2022) and management practices (minimum and no till seed bed preparation, slurry application, early and late sowing dates). To summarize, the new model version was able to successfully capture the main crop-related variables trends over time, as well as to correctly reproduce soil water content and temperature dynamic.