Modelling inter-annual veriation in dry matter yield and precipitation use efficiency of prennial pastures and annual forge crops sequences

In livestock systems of the Argentinean Pampas, its forage production stability relies on the integration of two landcovers, annual forage crop sequences and perennial pastures. Despite the key role that these forage cropping systems have on current milk and beef production, it is unclear how year-by-year variability of precipitation affect forage dry matter (DM) yield and precipitation use efficiency (PUE, i.e. the quotient between forage DM yield and precipitation). The aims of this study were to analyze the impact (i) of year-by-year precipitation variability on DM yield and PUE of oats-maize (Avena sativa L. - Zea mays L.) double-crop and alfalfa (Medicago sativa L.) and (ii) of cumulative precipitation during the critical period of maize on DM yield and PUE of oats-maize double-crop. We used a modelling approach to estimate DM yield and PUE of oats-maize (sequence) and alfalfa in five locations of the Argentinean Pampas, which differed in annual precipitation (AP) and variability of it. Coefficient of variation (CV) was used as the main statistical variable to compare the variability of AP (CV_AP), DM yield (CV_DM), and PUE (CV_PUE). Mean DM yield of both landcovers was higher in locations with high AP (>800 mm) than with low AP (<800 mm). Although alfalfa had lower mean DM yield than sequence in all locations, it showed a lower CV_DM than sequence. In contrast, sequence showed lower and higher CV_DM than CV_AP, depending on location. Moreover, changes in DM yield due to variations of AP were higher in sequence than in alfalfa. On the other hand, mean PUE was higher for sequence (2.2 g DM m⁻² mm⁻¹) than that of alfalfa (1.6 g DM m⁻² mm⁻¹). The CV_PUE between locations, i.e. an index that reflects the spatial variability, ranged from 20 for the sequence to 68% for alfalfa, whereas CV_PUE between years, i.e. an index that reflects the temporal variability, ranged from 16 to 31 % for both landcovers. Precipitation use efficiency tended to be similar across locations in years with low AP (<800 mm) compared to years with high AP (>800 mm). Our results provided valuable knowledge for decision making in livestock systems of this region through the development of spatial and temporal models between DM yield and AP. In a broader sense, they also showed that shifts from perennial to seasonal forage covers increased yields but also its inter-annual variability, posing a risk for farmers.