Utilization of CERES Maize (DSSAT) Model For Estimation of the Impact of Climate Changes on Corn Yield in South-Eastern Romania

Cătălin Lazăr¹, Mihaela Caian², Zenaida Chițu², Antoanela Dobre², Daniela Horhocea¹, George Cizmaș¹

¹National Agricultural Research and Development Institute Fundulea

²Administrația Națională de Meteorologie București

Keywords: corn, climate changes, DSSAT, CERES maize model, RCP 4.5, RCP 8.5.

Abstract: The importance of early shaping of an ideotype adapted to local climatic conditions in the coming decades under climate change is also evident in the very long time (about ten years) needed to cover the period between shaping new breeding objectives and approving new maize hybrids. The present paper presents preliminary tests using a deterministic modeling system implemented here by coupling the outputs of regional climate models (CORDEX/CMIP) with a dynamic model for simulating plant growth and development (CERES-DSSAT) with the aim to identify new corn ideotypes. The numerical simulations performed with this system used daily meteorological data for the Fundulea region and included: a set of DSSAT simulations with observed climatic for 1976-2005 (MARS-JRC Ispra), set that was used to calibrate and validate the system; then four sets of DSSAT simulations using modeled climate data, respectively: current climate data (HIST, 1976-2005) and simulated data for two radiative forcing scenarios (RCP 4.5 and RCP 8.5) and 2 time horizons (2021-2051 and 2071-2100). These last 4 sets of simulations were run using 3 climate models: CNRM-CERFACS-CNRM-CM5, ICHEC-EC-EARTH, MPI-M-MPI-ESM-LR (regionalized at 11 km with SMHI-RCA4 model). For the dynamic agro-meteorological simulations, the latest available open-source version of the CERES code was used. In each set of experiments the CERES model was run for each year for 12 agro-technical scenarios (4 sowing dates and 3 levels of fertilization).
Climate data analysis confirmed the general warming trend: in Romania with 1-1.5°C in winter and up to 2°C in summer in CMIP5 data, while the latest scenarios (CMIP6) amplify this signal even more (up to 2°C in winter and 3°C in summer), the southern region of the country having an even more significant growth. Regarding the precipitations, increases of the total seasonal quantities for winter and especially spring are projected, especially in the near decades 2021-2050 (on country average even by up to 8-10%, and in extremes exceeding the values currently registered) but at the same time with a decrease in the number of rainy days which indicates a higher frequency of days with extreme rainfall. Decreases in rainfall are projected in the summer, and more significantly in RCP 8.5.
The implemented system proved stability in operations and the simulated productions for the historical period presented inter-annual variations that are in accordance with the measured data for Călărași County. The signal of climate change was analyzed on 10 agronomic parameters simulated and compared in scenarios versus Hist for each model and for the model’s ensemble.
The simulated date of flowering occurs earlier in both climatic scenarios, on the average of the ensemble models this being ahead by up to one week; the gap is slightly larger for early sowing data. Also, the simulated maturity date is earlier (on ensemble average up to 10 days but the decrease is slightly smaller for early sowing dates).
The work was financed by UEFISCDI through the PrepClim project “System for the identification of corn ideotypes, optimal sowing data and nitrogen fertilization in the context of climate change” (PED 464/2020).