Long-term analysis from a cropping system perspective: Yield stability, environmental adaptability, and production risk of winter barley

Highlights

• Higher crop diversity and organic matter inputs increased yield stability.

• Moderate N fertilization (70 kg N ha⁻¹) benefitted yield stability.

• Organic matter improved environmental adaptability of barley in cereal rotations.

• Production risk was lower with higher crop diversity and organic matter inputs.

Abstract

In the face of climate change, the potential impacts of adverse weather conditions on the productivity and vulnerability of cropping systems (CS) worldwide constitute a key agronomic issue upon which global food security depends. To date, little information regarding how the diversity of CS or agricultural practices affect the long-term yield responses of winter barley is available. The aim of this study was to investigate the impact of CS diversity in terms of cropping sequences, organic matter supplies, and nitrogen (N) fertilization on the yield vulnerability of winter barley. Yields were evaluated in terms of their stability, environmental adaptability, and production risk (probability of yield loss). Data from a 27-year experiment were statistically analyzed using three mixed models giving residual maximum likelihood estimates including the Shukla stability variance index, the Finlay-Wilkinson environmental regression, and the Eskridge risk assessment approach. The results showed that winter barley grown in cropping sequences dominated by cereals had lower yield stability and environmental adaptability and greater production risks compared with winter barley grown in CS with higher crop diversity and additional organic matter inputs. When barley yields were compared at three doses of mineral N, the highest yield stability was achieved with the medium N dose (70 kg N ha⁻¹), followed by the higher level of N (140 kg N ha⁻¹). The most unstable yields with the highest production risks were observed when barley was grown without any mineral N fertilization. Overall, a higher level of crop diversity with organic matter inputs and intermediate N fertilization resulted in lower yield vulnerability in winter barley.

Introduction

Yield vulnerability can be defined as the susceptibility of a crop to extreme environmental change that negatively impacts yield or the resilience of crops (Urruty et al., 2016; Reidsma et al., 2010). More specifically, yield vulnerability can be used as a measure of (1) yield stability, expressed as the annual variability in crop yield, (2) environmental adaptability, indicating how well a crop responds to improved environmental conditions, and (3) production risk, which can be described as the probability of yield losses resulting from detrimental environmental change. These parameters include the effects of several environmental conditions, such as annual weather and soil conditions or pressure from pathogens (Manns and Martin, 2018; Dijkman et al., 2017; Wiebe et al., 2015; Cook, 2001) as well as agronomic factors, such as the cropping sequence or fertilization (Nielsen and Vigil, 2018; Sieling et al., 2015; Brisson et al., 2010; Helmers et al., 2001). For example, Peltonen-Sainio et al. (2010) reported that a higher frequency of early summer droughts resulted in yield instabilities, especially in cereal-dominated cropping sequences with limited crop diversity. Other studies reported positive effects on the yield stability of cereal crops and a reduced production risk when these crops were grown with favorable cropping sequences that included greater crop diversity and organic matter inputs compared to those of cereal-dominated cropping sequences without organic matter inputs (Babulicova and Dyulgerova, 2019; Gaudin et al., 2015; Hejcman et al., 2012; Buraczyňska et al., 2011; Stanger et al., 2008; Christen, 2001; Berzesenyi et al., 2000). A recent French study focused on the impact of climate change on the vulnerability of winter barley yields (Schauberger et al., 2019). This report and others have indicated that the vulnerability of cropping systems (CS) has increased in recent decades due to the presence of environmental stress caused by climate change (Ray et al., 2015; IPCC, 2014; Rosenzweig et al., 2014; Peltonen-Sainio et al., 2010). Climate change models predict that the risk of crop losses will continue to increase due to the higher frequency of extreme weather events (Wiebe et al., 2015).

An evaluation of crop performance as a function of CS and agronomic practices is important to ensure sustainable agricultural production in the future (Dijkman et al., 2017; Ishag, 2015; Olesen et al., 2011). In particular, an analysis of crop yield stability requires data from long-term experiments (LTE) that take into consideration variable weather conditions to evaluate the effects of cropping sequences, growing seasons, and management practices on crop production over a sufficiently long period (Piepho, 1998). Several approaches are available for estimating yield stability, such as Taylor’s power law regression with the related power law residuals stability measure (Döring et al., 2015), the adjusted coefficient of variation method (Döring and Reckling, 2019), the coefficient of variation method (Knapp and van der Heijden, 2018), and the use of classic parameters like ecovalence (Wricke, 1962).

The use of mixed models and residual maximum likelihood (REML) estimation are recommended in LTE stability analysis because they can accommodate LTE-specific factors, including the autocorrelation of residuals over time, non-orthogonal designs, and appropriate variance-covariance structures (Onofri et al., 2016). In most LTEs, the variance is not constant over multiple years and compound symmetry correlation structures with variance heterogeneity between years must be considered in models to ensure statistical accuracy (Onofri et al., 2016; Littell et al., 2006).

The Shukla stability model, which is used for estimating yield stability (Shukla, 1972), and the Finlay-Wilkinson regression approach, which gives information about environmental adaptability (Finlay and Wilkinson, 1963), are both appropriate ways of using mixed models for yield vulnerability analysis. For both, the REML estimates of the variance parameters serve as measures of stability (van Eeuwijk et al., 2016; Raman et al., 2011; Piepho, 1998). A useful addition to the evaluation of stability is a risk analysis that takes into consideration both absolute yield performance and stability (Piepho, 2000). The production risk provides further information regarding treatment resilience in response to environmental limitations and can be defined as the probability that the crop yield falls a given percentage below a critical pre-determined yield level (Eskridge, 1990). As such, the calculation of risk should be based on a mixed model approach that includes REML estimations so that the trial design and all LTE characteristics can be taken into account and thus provide valid results.

To date, only a few LTEs analysis have focused on the vulnerability of winter barley with regard to CS diversity or agronomic practices. For example, a cropping sequence study was carried out in Syria by Singh and Jones (2002), and a soil cultivation study was performed in Mexico by Fuhrer and Chervet (2015). However, few studies are relevant for temperate climate conditions (e.g., St-Martin et al., 2017; Christen, 2001). In view of the current challenging agronomic conditions that are expected to persist in the future, improved knowledge regarding how winter barley yields may be maintained and improved is needed. In particular, little evidence regarding the long-term impact of CS diversity or agricultural practices is available. The aim of this study was to investigate the impact of CS diversity on the yield vulnerability of winter barley in relation to the following three properties: (a) yield stability, (b) environmental adaptability, and (c) production risk. This study was based on data from an LTE covering a period of 27 years (1993–2019), which included different cropping sequences, organic matter inputs (straw and green manure), and three nitrogen (N) fertilization treatments. The data were used to test the following hypotheses: (H 1) CS with higher crop diversity and organic matter inputs help maintain winter barley yields and decrease yield vulnerability; (H 2) in cereal CS, organic matter inputs decrease the yield vulnerability of winter barley; and (H 3) mineral N fertilization reduces the yield vulnerability of winter barley.