High-yielding sugarcane in tropical Brazil – Integrating field experimentation and modelling approach for assessing variety performances

Highlights

• Outstanding yields by cane varieties were obtained under unlimited conditions in large field experiments in tropical Brazil.

• Guadalupe, PI State, is a more favourable production environment for sugarcane than São Romão, MG State.

• The radiation use efficiency was similar between varieties up to 6–8 months regardless of site.

• The radiation use efficiency declined differently among varieties after the onset of the growth slowdown.

• Traits such as stalk fraction and growth slowdown with ageing were derived for use in modelling with the APSIM-Sugar model.

Abstract

Aiming to gain an understanding of how the genotype × environment × management (G×E×M) interaction influences the yield accumulation by elite sugarcane varieties in Brazil, a large dataset from field plot experiments carried out in two tropical sites (Guadalupe, 6.8 °S; São Romão, 16.4 °S) involving distinct planting dates, varieties and harvest ages, was analysed with statistical techniques and with the APSIM-Sugar model.

Radiation use efficiency (RUE) was determined via a series of regressions and employed in an analysis of variance to investigate site, seasonal, developmental, and varietal differences. Outstanding yields were achieved at both sites. RUE declined as the crop progressed, confirming previous observations on declining RUE with age, known as the reduced growth phenomenon (RGP). RUE was always greater at Guadalupe than São Romão, evidencing that Guadalupe is a more suitable environment for sugarcane production, favoured by higher air temperatures during crop establishment and canopy formation. Varietal differences in RUE appeared only after the early developmental stage, and the observed growth slowdown with age was consistent across the two experimental sites, indicating that RGP is a varietal trait that should be considered for high-yielding environments.

The process-based APSIM-Sugar model was set up with recently determined canopy traits and a new RGP feature based on leaf appearance. RGP parameters were obtained for each variety and site through calibration. The calibrated model was accurate to account for yield accumulation by the varieties in both experiments. The new parameters were evaluated with independent datasets from other local experiments at each tropical site as well as from published rainfed experiments in sub-tropical Southeast Brazil. Independent verification of the RGP traits added confidence in the new way of dealing with RGP based on leaf stage. The G×E×M interaction on yield accumulation can now be explored more confidently with APSIM-Sugar for the purpose of optimising the choice of varieties, planting dates and harvest ages for sugarcane industries in favourable irrigated lands in tropical Brazil.

Introduction

Sugarcane is cultivated on about 26 M ha across the world mainly for sugar and bioenergy production, but also for molasses, alcoholic beverages, bioplastics and chemicals (FAO, 2020). Brazil is the world’s largest producer of sugarcane which is used mostly as a raw material for edible sugar, biofuel (ethanol) and bioelectricity production (Leal et al., 2013; Waclawovsky et al., 2010). The Brazilian government is adopting measures through the ‘RenovaBio’ program to stimulate the production and use of biofuels, including ethanol, aiming to reduce greenhouse gas emissions (Brazil, 2015; Grassi and Pereira, 2019; MME, 2017). The incentives for biofuels consumption will mainly require, from the agricultural perspective, increments in yields, but the increasing production in recent decades also came to some extent from the expansion of sugarcane on existing degraded lands, particularly extensive pastures (Adami et al., 2012; Oliveira et al., 2019). Therefore, understanding how sugarcane will perform in new environments to where the crop is expanding is a fundamental step to achieving sustainable production.

When the crop is harvested, all genotype × environment × crop management (G×E×M) interactions along the crop cycle result in the sugarcane biomass, which can be simply expressed in terms of radiation use efficiency (RUE). RUE can be defined as the aboveground biomass accumulated by a crop per MJ of global solar radiation or photosynthetically active radiation (PAR) intercepted or absorbed by the green leaf canopy (Bonhomme, 2000; Monteith, 1977, 1972; Sinclair and Muchow, 1999). Sugarcane is one of the most efficient crops concerning RUE (Sinclair and Muchow, 1999; Park et al., 2005), with values, based on global solar radiation, ranging between 1.4 g/MJ to 2.1 g/MJ (De Silva and De Costa, 2012; Ferreira Junior et al., 2015; Muchow et al., 1997; Robertson et al., 1996; da Silva, 2009; Singels and Smit, 2009). The high RUE of sugarcane could be associated with the high C₄-type photosynthesis (Sage et al., 2014), long growing season (Inman-Bamber, 2014) and low metabolic cost of plant organs (de Vries et al., 1989). Although being one of the most efficient crops, yields are constrained by a growth slowdown frequently observed across many traditional producing regions.

The growth slowdown as crop ages was recognised earlier by authors such as Rostron (1974); Lonsdale and Gosnell (1976); Thompson (1978); Inman-Bamber and Thompson (1989); Muchow et al. (1994), and Robertson et al. (1996). These last authors attributed the poor utilisation of intercepted radiation, hence reduced RUE, due to the loss of live millable stalks with crop age under high input conditions in a tropical Australian site. The term reduced growth phenomena (RGP) was first used by Park et al. (2005) in a comprehensive and evidence based study of yield accumulation in the Australian sugar industry. This phenomenon results in reduced RUE with age, usually under highly favourable environments where lodging is often noticed. Lodging reduces RUE, damages stalks and disrupts the canopy (impairing light interception), negatively affecting cane yield. It normally occurs in high-yielding areas with wet soil, where roots are kept in the upper layer, wet canopy (influencing the crop’s gravity centre) and windy conditions (>200 km/d) (Singh et al., 2002; van Heerden et al., 2015). Field experiments in Australia (Singh et al., 2002) and South Africa (van Heerden et al., 2015) showed that lodging reduces cane yields from 7.3 to 15% and sucrose yields from 8.8 to 35%, depending on the variety and weather conditions. Park et al. (2005) showed that RGP was also observed in some sugarcane crops which remained erect, raising the hypothesis that other factors such as an irreversible decline in leaf nitrogen content with age, higher maintenance respiration rate when crops were large, and feedback inhibition of leaf photosynthesis by high sugar content in mature internodes, could be involved (Park et al., 2005; van Heerden et al., 2010).

RGP factors are not fully accounted for any of the common sugarcane simulation models so far. Amongst the available sugarcane models, APSIM-Sugar (Keating et al., 1999) and DSSAT/CANEGRO (Jones and Singels, 2018; Singels et al., 2008) are those that simulate to some extent four of the six contributing RGP factors. The local effect of negative feedback of sucrose accumulation on leaf-level photosynthesis is not taken into account by the models listed, perhaps because of the lack of a complete understanding of the factors underpinning the sucrose dynamics.

In the widely used APSIM-Sugar model, maximum RUE (maximum ‘aboveground’ dry biomass produced per unit of canopy-intercepted radiation at an optimal temperature, crop water and nutrient status for a young, healthy crop, in g/MJ; Jones et al., 2019) is a parameter assumed to be constant over the growing season and has historically been used in this way, with few exceptions where site-specific lodging rules were applied to better simulate irrigated yields in Australia (Biggs et al., 2013; Inman-Bamber, 2004; Meier and Thorburn, 2016; Thorburn et al., 2017). Aiming to deal with this phenomenon through crop modelling, Dias et al. (2019) introduced a new feature in the APSIM-Sugar model (version 7.9 r4404 and later), which allows RUE to be modified by leaf stage as a catchall for all RGP factors aiming to avoid excessive prediction of high yields. The authors suggested that this feature in APSIM-Sugar should be used in situations where simulated yields are likely to exceed 150 t/ha (∼ 40 t/ha in dry mass basis). In Dias et al. (2019)’s study, cane yield started to differentiate between varieties from 7 to 8 months-old onwards, which was attributed to varietal differences in the RGP. These authors acknowledged that such varietal differences needed further investigation and inclusion into modelling capability.

The large G×E×M experiment at Guadalupe, PI, Brazil reported by Dias et al. (2019), was similar to another large concurrent experiment at Sao Romão, MG, Brazil for which yield results have yet to be published. These experiments were designed to inform the planning of greenfield sugarcane projects in the Brazilian tropics where choice of variety, planting dates and the length of harvest season were yet to be decided. While each experiment included six varieties, six planting dates and 3–5 harvest ages from 7 to 16 months, it was acknowledged that these combinations included only a fraction of all combinations of variety, planting and harvest seasons possible. Measurements were conducted so that models could be tested and calibrated to allow many more G×E×M treatments to be analysed than those represented in the experiments. Simulations were also required to assess the impact of medium-term climate on sugarcane yields and quality. This paper reports on how such simulations are now possible for sugarcane in tropical Brazil with a focus on RUE and its RGP, which are crucial physiological parameters for such simulations. The specific objectives were to:

• Assess the impact of environment, planting date, variety, and crop developmental stage on RUE.

• Test the hypothesis that RGP differs between varieties and is, therefore, a varietal trait.

• Evaluate the capability of APSIM-Sugar to represent the G×E×M interaction on sugarcane yields in highly productive environments where RGP plays an important role.