Cropping system productivity and evapotranspiration in the semiarid Loess Plateau of China under future temperature and precipitation changes: An APSIM-based analysis of rotational vs. continuous systems

Agriculture in the semiarid region is undergoing radical changes driven by global warming and increasing incidences of extreme weather events. Predicting and evaluating the responses of crop yield and water use patterns of rainfed cropping systems according to future temperature and precipitation changes could provide important information regarding adoption of climate-smart farming systems that could offer great resilience and sustainability. The objective of this study was to evaluate the potential changes in agronomic productivity, hydrological balance and economic profitability of cropping systems with perennial legumes and rotation on the Loess Plateau of China affected by different future temperature and precipitation scenarios using APSIM-based modeling. Five different cropping systems were investigated: (i) continuous maize (Zea mays) (M), (ii) continuous winter wheat (Triticum aestivum) (W), (iii) continuous lucerne (Medicago sativa) (L), (iv) maize-wheat-soybean (Gyleine max) rotation (MWS) and (v) lucerne (4-yr)-winter wheat (2-yr) rotation (LW) in Xifeng, Gansu, China. Five series of temperature and precipitation changes scenarios (no changes in atmospheric CO2 concentrations were considered) based on RCPs (representative concentration pathways) were integrated into scenario simulations, including the baseline scenario (1980-2010), mid-century scenarios of RCP4.5 (M45) and RCP8.5 (M85), and end-century scenarios of RCP4.5 (E45) and RCP8.5 (E85). The results showed that, compared with baseline simulations, mean maize yield under RCPs decreased by 6.7 0 /0-37.7 %, and the mean wheat yield decreased by 1.7 %-23.6 %. Lucerne yield consistently increased by 7.2 %-12.3 % under M45 and E45. Although different cropping systems affected the yield to a certain extent, only the difference between the wheat yields of W and MWS was significant (P < 0.05). On a cropping system level of every 6-yr (totally five phases for a scenario of 30-yr), the temperature and precipitation changes did not significantly affect plant transpiration (T-c) per phase. For all systems except L [which received its highest soil evaporation (E-s) per phase in baseline], the greatest E-s per phase was found in the E85. L tended to provide greater T-c per phase and evapotranspiration (ET) per phase (ranging from 1219 to 1332 mm and 2910-3034 mm, respectively), followed by LW. L and MWS presented the highest and lowest E-s per phase (1636-1724 and 1592-1677 mm), respectively. The gross profit (GP per 6-yr phase) and water productivity (WP per 6-yr phase) of L were the greatest among all cropping systems (11.5-13.7 thousand US$ ha(-1) per phase and 3.72-4.41 US$ ha(-1) mm(-1) per phase, respectively), followed by LW and MWS. For M, W and MWS, the GP per phase and the WP per phase were predicted to be greater under the baseline than under any of the RCPs, whereas M45 and E45 scenarios improved the GP per phase and WP per phase of L and LW compared with those of the baseline. In general, we advocate that LW-based systems have the greatest potential for producing acceptable yield and economic profit under future temperature and precipitation scenarios for this local environment.