Impact assessment of soybean yield and water productivity in Brazil due to climate change

Highlights

• We calibrated the DSSAT/CROPGRO for several soybean maturity groups.

• Crop model showed good performance under different tropical conditions.

• Water productivity increase can ensure the future growth of soybean yield.

• Mean soybean yields increase likely in the future climate due to CO₂ increase.

Abstract

In the next decades, the population is expected to rise by more than two billion people, and the projections of climate change have been considered as one of the greatest future challenges for world food security. Soybean represents more than 60 % of all plant protein produced in the world, and Brazil is the largest world exporter and the second-largest producer. In this paper, we simulated soybean yields for 16 strategically selected agroclimatic zones (CZs) to represent Brazilian production. Experiments conducted throughout the country were used to calibrate the CROPGRO-Soybean model for Brazilian conditions, for the main maturity groups used in Brazil, to simulate current and 40 future climate scenarios, provided by Coupled Model Intercomparison Project 5 (CMIP5) for 2050 in the both 4.5 and 8.5 representative concentration pathways (RCP). We found soybean yield varying by +1 to +32 % across 16 CZs in the average scenario of future climate when compared to the current yields. Yet, we found an increase of about 5% in the yield production risk for RCP 8.5. The main reason for such results was associated with the positive effect of increasing CO₂ on crop water productivity, which overcomes the negative effects of temperature and water stress increases on rainfed Brazilian soybeans.

Introduction

The global population has grown dramatically in the last decades. Between 2017 and 2050, the world’s population is projected to rise from 7.7 billion to 9.8 billion (United Nations, 2017). Similarly, there is a projection of income increase across the developing world, which would imply an increase in demand for meat and dairy products by rates ranging from 50 to 100 % in the next decades (Thornton, 2010; Foley et al., 2011; Tilman et al., 2011; Clark and Tilman, 2017; Searchinger et al., 2019). This may result in grain price increases (Marchand et al., 2016), which might worsen food insecurity in the world’s poorest countries (Sternberg, 2012; D’Amour et al., 2016).

Soybean (Glycine max L.) is currently the world's most important food protein source, and so it is crucial for food security. Soybean is main source of high-quality vegetable protein for the production of food of animal origin (Speedy, 2002; Thomson, 2019). In the past twenty years, Brazil had one of the most significant expansions of agricultural land use (Zalles et al., 2019), and has established itself as the world's largest soybean producer, with over 38 million hectares of soybean production (CONAB, 2017), being the main economic product of Brazil (Ministério da Economia, 2020).

Climate change is expected to affect agriculture worldwide in the next decade (Zabel et al., 2019; Baldos et al., 2020). General circulation models (GCMs) are central to climate change research (Field, 2014) and can be coupled to the cropping system model (CSM) to investigate the scientific hypotheses about the impacts of climate change on agriculture (Rosenzweig et al., 2013; He et al., 2017). The CSM-CROPGRO-Soybean (Boote et al., 1998) is one of the more robust and widely model used to simulate crop production systems. The model has a modular structure (Jones et al., 2001), which consider process of carbon and nitrogen balance (Godwin and Allan, 1991; Godwin and Singh, 1998), soil water balance (Ritchie, 1998; Silva et al., 2021); and it is able to simulate evapotranspiration (Boote et al., 2008; Cuadra et al., 2021), crop water productivity (Dias et al., 2020; Edreira et al., 2018; Er-Raki et al., 2020; Timsina et al., 2008), soybean growth and development (Boote et al., 1998; Bhatia et al., 2008), and crop production under climate change conditions (Adhikari et al., 2016; Antolin et al., 2021; Bao et al., 2015; Fava et al., 2020; Quansah et al., 2020; Souza et al., 2019).

Climate change effects on soybean have been globally studied in recent papers (Schauberger et al., 2017; Lee and Mccann, 2019; Mourtzinis et al., 2019). In Brazil, Rio et al. (2016) found a decrease in the yield of up to 35 % for the southern Brazilian region, while Justino et al. (2013) found an increase of up to 60 % for the Midwest and northern Brazilian regions. These researches, however, lack consistency in modeling frameworks, did not cover the whole country using a consistent protocol, and did not consider the regional variability in terms of genetics and farming systems.

In this paper, we used a large experimental dataset collected in the several key-producing regions of Brazil for calibrating a process-based crop model and then simulated 40 future climate scenarios to prospect the soybean yield and water productivity in Brazil, for 2050. For those simulations, we have considered the future atmospheric dioxide carbon concentration ([CO₂]), maturity group, sowing date window, and soil water content across environments to represent the climate change impact around Brazil. Our main goal was to evaluate the effects of future climate scenarios on rainfed soybean yield and water productivity and propose strategies to cope with possible future climate limitations.