Login / Signup

Sensitivity and requirement of improvements of four soybean crop simulation models for climate change studies in Southern Brazil.

Rafael BattistiP C SentelhasK J Boote
Published in: International journal of biometeorology (2017)
Crop growth models have many uncertainties that affect the yield response to climate change. Based on that, the aim of this study was to evaluate the sensitivity of crop models to systematic changes in climate for simulating soybean attainable yield in Southern Brazil. Four crop models were used to simulate yields: AQUACROP, MONICA, DSSAT, and APSIM, as well as their ensemble. The simulations were performed considering changes of air temperature (0, + 1.5, + 3.0, + 4.5, and + 6.0 °C), [CO2] (380, 480, 580, 680, and 780 ppm), rainfall (- 30, - 15, 0, + 15, and + 30%), and solar radiation (- 15, 0, + 15), applied to daily values. The baseline climate was from 1961 to 2014, totalizing 53 crop seasons. The crop models simulated a reduction of attainable yield with temperature increase, reaching 2000 kg ha-1 for the ensemble at + 6 °C, mainly due to shorter crop cycle. For rainfall, the yield had a higher rate of reduction when it was diminished than when rainfall was increased. The crop models increased yield variability when solar radiation was changed from - 15 to + 15%, whereas [CO2] rise resulted in yield gains, following an asymptotic response, with a mean increase of 31% from 380 to 680 ppm. The models used require further attention to improvements in optimal and maximum cardinal temperature for development rate; runoff, water infiltration, deep drainage, and dynamic of root growth; photosynthesis parameters related to soil water availability; and energy balance of soil-plant system to define leaf temperature under elevated CO2.
Keyphrases
  • climate change
  • working memory
  • machine learning
  • physical activity
  • radiation therapy
  • convolutional neural network
  • ultrasound guided