Lentil enhances the productivity and stability of oilseed-cereal cropping systems across different environments

Lentil enhances the productivity and stability of oilseed-cereal cropping systems across different environments

European Journal of Agronomy 105 (2019) 24–31 Contents lists available at ScienceDirect European Journal of Agronomy journal homepage: www.elsevier...

936KB Sizes 0 Downloads 39 Views

European Journal of Agronomy 105 (2019) 24–31

Contents lists available at ScienceDirect

European Journal of Agronomy journal homepage: www.elsevier.com/locate/eja

Lentil enhances the productivity and stability of oilseed-cereal cropping systems across different environments

T

Kui Liua, Robert E. Blackshawb, Eric N. Johnsonc, Zakir Hossaina, Chantal Hameld, ⁎ Marc St-Arnaude, Yantai Gana, a

Agriculture and Agri-Food Canada, Swift Current Research and Development Centre, Swift Current, Saskatchewan, S9H 3X2, Canada Agriculture and Agri-Food Canada, Lethbridge Research and Development Centre, Lethbridge, Alberta, T1J 4B1, Canada c Department of Plant Science, University of Saskatchewan, Campus Road, Saskatoon, Saskatchewan, S9N 3A8, Canada d Agriculture and Agri-Food Canada, Québec Research and Development Centre, 2560 Blvd. Hochelaga, Québec, G1V 2J3, Canada e Institut de recherche en biologie végétale, Université de Montréal and Jardin botanique de Montréal, 4101 East, Sherbrooke St., Montréal, Québec, H1X 2B2, Canada b

A R T I C LE I N FO

A B S T R A C T

Keywords: Cropping system Durum wheat Oilseed crops Pulse crops Stability System productivity

Enhancing the stability of crop production is vital for agriculture under climate uncertainty. Conventional fallow and shallow-rooting pulse crops such as lentil (Lens culinaris Medic.) have been incorporated in oilseed-cereal cropping systems to cope with dry conditions on the semiarid Canadian prairies. However, the static and dynamic stability of these adaptation practices at a cropping system level is unclear due to the complexity of interactions. This study assessed the effects of diversified rotation systems as drought adaptation practices on the productivity and stability of oilseed-cereal cropping systems. Nine 3-year crop rotations were tested for two cycles at three sites from 2013 to 2016. The 3-year crop sequences included fallow, lentil, and spring wheat (Triticum aestivum L.) in rotation phase 1, followed by canola (Brassica napus L.), oriental mustard (Brassica juncea L.), and camelina (Camelina sativa L. Crantz) in phase 2, and the rotation phase 3 was durum wheat (Triticum turgidum L. subsp. durum (Desf.) Husn.) in all nine rotation systems. On average, the lentil system increased system productivity, expressed by annualized durum wheat equivalent yield, by 24% and 78% compared with the spring wheat and fallow systems, respectively. Stability analysis revealed that the lentil – B. juncea – durum wheat and lentil – B. napus – durum wheat systems had the least variation across the environments and were well adapted to high-yielding sites. The fully-phased rotations across various environments showed that the drought-induced reductions in system productivity ranged from 3 to 47% compared with yields under normal weather, with the lentil – oilseed – durum systems having least reduction. Quantitative assessments revealed that about 36% of the variation in system productivity was associated with rotation systems and additional 30% was due to weather-related factors. In conclusion, the inclusion of lentil in rotation increases systems productivity and reduces yield variation in oilseed-cereal cropping systems in changing environments.

1. Introduction Agroecosystems face enormous challenges with increased inter-annual weather variation due to climate change (Lobell et al., 2011). Higher weather variation is associated with higher frequency of abnormal heat, flood, and drought events (Hatfield et al., 2011). These abiotic stresses likely trigger biotic stresses in crops, and the interaction of biotic and abiotic stresses in agroecosystems result in pest infestation (Shennan, 2008), low nutrient use efficiency (Médiène et al., 2011), and yield reduction (Pandey et al., 2017). Globally, weather variation explains approximately one-third of the variation observed in crop yield

(Ray et al., 2015). Therefore, enhancing the stability of crop production under a changing climate becomes vital for agricultural sustainability. Recently, a variety of concepts such as stability, robustness, vulnerability, and resilience have been proposed to assess the sustainability of agricultural production systems (Lin, 2011; Urruty et al., 2016). Crop production is the central component of agricultural sustainability. Yield stability reflects the capacity of crops to maintain a constant production under changing environments (Urruty et al., 2016). The analysis of yield stability aims to assess the interaction between crop and the environment by determining the level of yield sensitivity to environmental factors in agroecosystems (Berzsenyi et al.,

Abbreviations:CV, coefficient of variation ⁎ Corresponding author. E-mail address: [email protected] (Y. Gan). https://doi.org/10.1016/j.eja.2019.02.005 Received 28 September 2018; Received in revised form 21 November 2018; Accepted 5 February 2019 1161-0301/ Crown Copyright © 2019 Published by Elsevier B.V. All rights reserved.

European Journal of Agronomy 105 (2019) 24–31

K. Liu, et al.

following crops (Niu et al., 2017), thus improving farm profitability. A cropping system is a living ecosystem (Gliessman, 2004), where the biotic and abiotic factors interact and affect the stability of production. Due to the complexity of cropping systems, the production stability of cropping systems integrating fallow and drought-tolerant crops has not been assessed on the semiarid Canadian prairies from a systems perspective. The objectives of this study were to: 1) determine the effects of crop stubble management practices and oilseed species on system productivity, 2) assess the yield stability of cropping systems across environmental gradients, and 3) estimate the contribution of cropping system and weather-related factors to system variation across multiple environ-sites.

2000). The relationship between the coefficient of variation (CV) and average yield is commonly used to describe the static stability of cropping systems across multiple environmental sites (Raseduzzaman and Jensen, 2017). The Finlay-Wilkinson regression model is often used to assess the dynamic stability (called agronomic stability) or responsiveness of yield across multiple environmental gradients (Finlay and Wilkinson, 1963; Berzsenyi et al., 2000). Static stability reflects the overall variation across environments with less emphasis on the cropenvironment interaction, while dynamic stability stresses the crop-environment interaction and the adaptability of crops to changing environments (Becker and Léon, 1988). A stable cropping system is characterized by a minimal variance and crop-environment interaction (Becker, 1981). In the context of crop-environment interaction, variance partitioning between crop and environment is typically used to determine the driving factors for system production (Lychuk et al., 2017; Niu et al., 2017). In the scientific literature, both static and dynamic yield stabilities are determined mostly at an individual crop level and rarely at the cropping system level due to the complexity of a cropping system. However, there are growing needs to assess cropping system stability at a regional scale to improve agricultural sustainability and overall profitability. The Canadian prairies account for approximately 83% of Canada’s arable land, producing more than 90% of overall wheat and oilseed crops in Canada, with a farm gate value of more than $20.6 billion annually (Statistics Canada, 2017). On the semiarid Canadian prairies, drought has been reported to be one of the main limiting factors of crop production (Gan et al., 2015). This is why fallow was widely practiced in the 1970s to conserve water for the production of a high yielding crop the following year (Campbell et al., 2002). However, fallowing causes yield penalty as there is zero production in the fallow year, which reduces the production of the whole system (Gan et al., 2015). Also, a high frequency of fallowing in rotation deteriorates soil quality in the long-term due to low crop residue input (Lemke et al., 2012). In recent years, fallow has largely been replaced by direct seeding into standing stubble, a more sustainable soil moisture conservation practice, on the semiarid Canadian prairies (Campbell et al., 2002; Bueckert and Clarke, 2013). Replacement of fallow with direct seeding of oilseed and pulse crops increased planted areas on an annual basis, making oilseed – cereal or pulse – cereal more profitable systems than the traditional fallow-based systems on the Canadian prairie. The production of canola, a cool-season crop, is often constrained by drought and heat stresses, especially when they occur during flowering (Gan et al., 2004). The projected climate shifts in temperature and precipitation threaten the stability of canola production (Meng et al., 2017). In comparison, mustard (B. juncea) has a better tolerance to drought and heat (Gan et al., 2004; Blackshaw et al., 2011). Similarly, camelina is reported to be suitable for dryland environments across the North American Great Plains (Obour et al., 2018). Diversifying canola production with alternative oilseed crops such as mustard and camelina might stabilize the production of oilseed for edible oil and biofuel (Blackshaw et al., 2011). In the oilseed-cereal cropping system, a dominant cropping system on the semiarid Canadian prairies (Martens et al., 2015), durum wheat has been increasingly used (Statistics Canada, 2017). A shift from spring wheat to durum wheat might be expected under the foreseen climate change on temperature and precipitation, because durum wheat is more tolerant to drought and heat stresses than spring wheat (Bueckert and Clarke, 2013). With the increasing uncertainties of a changing climate, biodiversity has been proposed as a key solution to adapt and enhance systems stability (Lin, 2011; Gan et al., 2015; Urruty et al., 2016). Pulse crops play an important role in diversifying the current cropping systems in dryland regions (Cutforth et al., 2007). Lentils require less water than cereal or oilseed crops, and its shallow rooting system leaves behind residual soil water for the following crops (Gan et al., 2017). In addition, the biological nitrogen fixation of lentil crops reduce the dependence of cropping system on N fertilizer and increase the yield of the

2. Materials and methods 2.1. Site description A 3-year crop rotation study was conducted at three sites in western Canada: Swift Current, Saskatchewan (50°25ʹ N, 107°44ʹ W); Scott, Saskatchewan (52°21ʹ N, 108°50ʹ W); and Lethbridge, Alberta (49°41ʹ N, 112°45ʹ W), in 2013, 2014, 2015, and 2016. These sites were selected to represent three different eco-zones on the Canadian prairies. Swift Current and Lethbridge are located in semiarid regions, but the Swift Current region is cooler than Lethbridge and characterized by the highest evaporation rate among the three sites, while the Lethbridge site was under irrigation. In comparison, Scott is located in a cool, subhumid region at a higher latitude with limited growing degree days. The soil type at Swift Current is silty loam Brown Chernozem, while it is loam Dark Brown Chernozem at Scott and clay loam Dark Brown Chernozem at Lethbridge. All three experimental sites were located at the research farms of Agriculture and Agri-Food Research Center. The long-term weather data were recorded at each site and the long-term (1981–2010) annual precipitation is 405.5, 351.2, and 372.1 mm at Lethbridge, Scott, and Swift Current, respectively. The long-term precipitation during the typical growing season (01 May to 31 August) is 229.2, 209.4, and 206.4 mm at Lethbridge, Scott, and Swift Current, respectively. 2.2. Experimental design The 3-year crop rotation started with three crop stubble management practices including fallow, lentil (cv. Maxim), and spring wheat (cv. AC Lillian) in rotation phase 1 with each separate rotation cycle starting in 2013, 2014, and 2015 at each of three sites (Fig. 1). Prior to phase 1, a wheat crop was seeded at all nine environ-sites, ensuring the same experimental background. In rotation phase 2, three oilseed crops, including canola (cv. L252LL), oriental mustard (cv. Cutlass), and camelina (cv. Midas), were grown into each of the three crop stubble management practices, forming nine cropping systems. In phase 3, durum wheat (cv. Brigade) was seeded into all plots. The experiment was arranged in a split-plot design, with three crop stubble management practices (in phase 1) being randomly assigned to the main plots and three oilseed crops (in phase 2) being randomly assigned to the subplots. There were four blocks, totaling 36 plots for each individual rotation phase at each environ-site. The rotation was repeated for three cycles; the first cycle started in 2013 and completed in 2015, the second cycle started in 2014 and completed in 2016, and the third cycle started in 2015 and ended in 2016 at the oilseed rotation phase (no durum wheat was seeded in the last year of the third cycle due to project management errors). With the experimental design, a rotation started each year from 2013 to 2015 formed a fully-phased 3-year rotation in 2015, where crops experienced severe drought at all sites. 2.3. Crop management In rotation Phase 1, lentil and wheat were directly seeded into 25

European Journal of Agronomy 105 (2019) 24–31

K. Liu, et al.

Fig. 1. Field plot layout showing crop sequence within and between rotation cycles at one experimental site, 2013–2016. This study was conducted at three sites (Lethbridge, Scott, and Swift Current) with 4 blocks per site.

harvested using a plot combine at plant full maturity stage. Harvested seeds were cleaned, air-dried, and adjusted to a moisture content of 13.0% for lentil, 14.5% for spring and durum wheat, and 10.0% for oilseeds for yield reporting. To compare the performance of different crop rotation systems, we used durum wheat, the most common crop in this study, as a denominator, to calculate durum wheat equivalent yield (DWEY) as:

wheat stubble from late April to mid May. The seeding rates were 150 seeds m−2 for both lentil and wheat. In rotation Phase 2 oilseeds were seeded in May each year. The seeding rates were 500 seeds m−2 for camelina, 200 seeds m−2 for B. juncea, and 150 seeds m−2 for B. napus. In rotation Phase 3, durum wheat was seeded from late April to midMay with a seeding rate of 250 seeds m−2. The seeding rate in this study followed the recommendation of the best local management practices. Lentil seeds were treated with rhizobial inoculant (TagTeam at 3.7 kg ha-1 in granular). Fertilization rates for all the oilseed and cereal crops were based on soil test recommendation on a yearly basis. Lentil crops received no synthetic chemical nitrogen fertilizer. For weed control, a pre-seed ‘burn off’ treatment using glyphosate (900 g a.e. ha1 ) was applied to all plots. Glyphosate (900 g a.e. ha-1), Assure II (36 g a.i. ha-1), and Buctril M (560 g a.i. ha-1) were applied to the fallow, lentil and oilseed, and wheat plots for weed control, respectively. Fungicides were applied when necessary following site-specific recommended practices. No irrigation was applied at Swift Current or Scott, while, at Lethbridge, supplementary irrigation was applied to crops during periods when below average precipitation was received.

Where Pnon − durum wheat is the price of non- durum wheat crops, Pdurum wheat is the price of durum wheat, and Ynon − durum wheat is the yield of nondurum wheat crops. The average prices during the study period were $657.4, $219.1, $747.6, $463.0, $391.3, and $283.6 per tonne for lentil, wheat, B. juncea, B. napus, camelina, and durum wheat, respectively (Government of Saskatchewan, 2017). System productivity of each rotation was reported as the annualized DWEY, which was calculated as the average of DWEY across the three rotation phases for each rotation.

2.4. Data collection and processing

2.5. Statistical analysis

Durum wheat equivalent yield =

Pnon − durum wheat × Ynon − durum wheat Pdurum wheat

The yield of durum wheat was analyzed using the PROC MIXED

A six to eight rows of plants from the centre of each plot were 26

European Journal of Agronomy 105 (2019) 24–31

K. Liu, et al.

temperature for warm season crops than for cool season crops (Luo, 2011).

model of SAS, where crop stubble management practices (fallow, lentil, and wheat), oilseed crop species (B. napus, B. juncea, and C. sativa), and sites (Lethbridge, Scott and Swift Current) were considered fixed factors, and the 4 blocks and 2 complete rotation cycles at each site were combined and formed 8 new blocks, which were considered a random factor. A similar analysis was conducted on the annualized DWEY to assess system performance. Given the volatility of crop prices, sensitivity analysis on annualized DWEY was conducted on selected crops which had relatively large price variation. Among the study crops, the prices of lentil and B. juncea fluctuated the most. The price of lentil ranged from $462 to $828 per tonne, while the price of B. juncea ranged from $634 to 893 per tonne. Taking price variability and crop yield into account, lentil and B. juncea crops resulted in the largest fluctuation of durum wheat equivalent yield; therefore, lentil and B. juncea were selected as example crops for separate sensitivity analysis for the crop stubble management and oilseed crop factors, respectively. For the sensitivity analysis of lentil, the DWEY was separately calculated three times using the low, average, and high lentil prices, while holding the prices of all other crops constant as their respective averages during the study period. The low, average, and high lentil prices correspond to the lowest, average, and highest prices during the study period. The annualized DWEY was analyzed using PROC MIXED model of SAS, where the four blocks and two separate rotation cycles at each site were combined as eight new blocks, which were considered a random factor; and crop stubble management, site, and price were fixed factors. A preliminary analysis indicated that there were no significant site and price interaction, and three sites were combined with eight new blocks and were considered a random factor. The same approach was used for the sensitivity analysis for the B. juncea crop. The normality and homogeneity of variance were tested on the residuals. For significant treatment effects, the mean comparisons were performed at the probability level of 0.05 and the least squares means were reported. Yield stability was determined for each 3-yr rotation system across the nine environ-sites, including six (3 sites × 2 separate cycles) complete rotation cycles and 3 fully-phased rotations in 2015. The static yield stability of cropping systems was estimated using the coefficient of variation (CV), with a lower CV meaning a higher stability. A CV vs mean system production was plotted to visualize the static stability, and a cropping system with lower CV but higher system production is ideal. The dynamic yield stability for each cropping system was assessed using the Finlay-Wilkinson stability regression coefficient, bi (Finlay and Wilkinson, 1963). The regression coefficient was obtained from the linear regression between system production for each cropping system and the environmental index across nine environsites. The environmental index was estimated as the mean system production of all cropping systems at each environ-site. A cropping system with a bi close to 1 indicates that this cropping system adapts to all study environments. A bi < 1 indicates that a cropping system has above-average stability but adapts to low-yielding environments. A bi > 1 indicates that a cropping system has below-average stability but is suitable for the high-yielding environments. Variance in system production was partitioned into cropping system and weather factors using the vegan package in R Statistical Software (Oksanen et al., 2017). The cropping system factor includes nine levels resulting from the three crop stubble management practices (fallow, lentil, and wheat) and three oilseed crops (B. juncea, B. napus, and C. sativa). Weather factors include the annualized growing degree-days during the actual growing season; and annualized precipitation accumulated one month prior to seeding, during the actual growing season, and during the typical growing season from May to August. Irrigation amounts when applicable were counted as part of growing season precipitation. For simplicity, the base temperature for the growing degree-days was set at 5 °C for all study crops according to a local study (Qian et al., 2013), although the base temperature, under which no crop growth or development occurs, varies by crops with higher base

3. Results 3.1. Yield of durum wheat The yield of durum wheat was significantly influenced by experimental site (P < 0.01) and crop stubble management practices (P = 0.03), and there was no significant two- or three-way interactions. Among the three sites, the durum wheat yield was the highest at Lethbridge (3617 kg ha−1), followed by the Swift Current site (3502 kg ha−1), and the lowest at Scott (1860 kg ha−1). Across the six site-years, the average durum wheat yield was 6 and 17% higher in the lentil cropping system (3117 kg ha−1) than in the spring wheat (2944 kg ha−1) and fallow (2654 kg ha−1) systems, respectively. 3.2. Production of cropping systems System production, as expressed by annualized durum wheat equivalent yield, was significantly affected by the site × crop stubble management practices (P < 0.01) and site × oilseed crop species (P < 0.01). Among the three crop stubble management practices, the fallow system resulted in significantly lower system production than the lentil and wheat systems at all study sites (Fig. 2A). Across the three sites, the lentil and spring wheat systems increased system production by 44–110 and 20–82% compared with the fallow system, respectively. The DWEY responded to lentil and wheat systems differently across the three sites; the DWEY in the lentil system was not different than in the spring wheat system at Lethbridge, but was 16 and 66% higher at Scott and Swift Current, respectively.

Fig. 2. Annualized durum wheat equivalent yield of the 3-year cropping systems as related to: (A) the two-way interaction of site (Lethbridge, Scott and Swift Current) and stubble management practices (Fallow, Lentil and Spring wheat), and (B) the two-way interaction of site (Lethbridge, Scott and Swift Current) and oilseed species (B. juncea, B. napus and C. sativa), 2013–2016. Bars marked by the same letters within each interaction are not different at the significance level of 0.05. 27

European Journal of Agronomy 105 (2019) 24–31

K. Liu, et al.

Fig. 4. Static stability of system production for the nine 3-year crop rotations conducted at three sites, 2013-2016. The coefficient of variation of each crop rotation was calculated across the nine environ-sites. A crop rotation with a lower coefficient of variation means a higher stability.

3.4. Stability and variance partitioning of system production Fig. 3. Annualized durum wheat equivalent yield of 3-year cropping systems in response to (A) the price change (High, Average and Low) of lentil and (B) the price change of B. juncea, 2013–2016. Bars marked by the same letters within each interaction are not different at the significance level of 0.05.

For static stability, the lentil-based cropping systems including B. juncea and B. napus were most stable, while the fallow-based cropping systems involving B. juncea or camelina were least stable (Fig. 4). For dynamic stability, the wheat-B. napus system was most stable among all cropping systems as indicated by the associated bi , which was closest to 1 (Fig. 5). The lentil-based B. juncea and B. napus cropping systems were most responsive to changing environments, while all three fallow-based systems were least responsive, as revealed from the regression coefficients. Variance partitioning analysis indicated that both cropping system and weather factors had significant effects on system production

Among the three oilseed types, the C. sativa cropping system resulted in significantly lower system production than the B. juncea and B. napus cropping systems at all study sites (Fig. 2B). Across the three sites, the B. juncea and B. napus systems increased system production by 18–57 and 11–39% compared with the C. sativa system, respectively. The response of DWEY to B. juncea and B. napus varied across the three sites. There was no difference in DWEY between the B. juncea and B. napus systems at either Scott or Swift Current, but, at Lethbridge, the DWEY for the B. juncea system was 13% higher than for the B. napus systems.

3.3. Sensitivity to lentil and B. juncea prices The annualized durum wheat equivalent yield was significantly affected by the price × crop stubble management (P < 0.01) (Fig. 3A) and price × oilseed crop species (P < 0.01) (Fig. 3B). At the full range of lentil price during the study period, the lentil system significantly increased DWEY by 57–98% compared with the fallow system. As the lentil price decreased to the lowest level, the lentil system produced DWEY similar to the spring wheat system. On the other hand, the lentil system increased DWEY significantly by 37 and 24% compared with the spring wheat system at high and average lentil prices, respectively. In all three selected B. juncea price scenarios, ranging from the lowest to highest during the study period, the B. juncea system significantly increased DWEY by 26–49% compared with the C. sativa system. As the B. juncea price decreased to the lowest level, the B. juncea system produced slightly lower DWEY than the B. napus system, while the B. juncea system increased DWEY significantly by 8 and 18% compared with the B. napus system at average and high B. juncea prices, respectively.

Fig. 5. Dynamic stability of system production for the nine 3-year crop rotations conducted at three sites, 2013–2016. Regression coefficients were obtained from a linear relationship of system production in each rotation between a specific site and the environmental index. A crop rotation with a regression coefficient closer to 1 is more stable across the study environments. 28

European Journal of Agronomy 105 (2019) 24–31

K. Liu, et al.

the sensitivity analysis indicated that even at the lowest lentil price during the study period, the lentil system production maintained the highest benefits among the three crop stubble management practices, due mainly to high yields of the crops involved in the lentil systems. The results of sensitivity analysis indicate that even though the grain price fluctuates widely, the lentil system is the most profitable system in the semiarid region, where inadequate water supply limits crop production.

(P < 0.01). Approximately 36% of the overall variance in system production was attributed to the influence of cropping system, while an additional 30% of the variance was explained by weather factors such as precipitation and temperature. 4. Discussion 4.1. Crop stubble management practices effects At the Year 3 durum wheat rotation phase, the lentil system significantly increased the durum wheat yield compared with the spring wheat and fallow systems. On the semiarid Canadian prairies, water deficit is the major concern for sustainable agricultural production. Fallow was practiced to conserve water for the following crops and, as expected, fallow practices increased the following oilseed crop yield (2772 kg ha−1) by 4 and 5% compared with the systems involving lentil (2663 kg ha−1) and spring wheat (2646 kg ha−1), respectively. However, fallow’s benefits for crop yield did not extend to the following second year, as evidenced by the low durum wheat yield. The relatively high yield of oilseed crops grown on the fallow practice might have used more nutrients and water, leaving relatively fewer resources to the following durum wheat, explaining in part the significantly low durum wheat yield for the fallow system. In the lentil system, the shallow root system of lentil crops utilized the water only in the shallow soil layer, conserving deep soil water for the following crops, acting as a partial fallow in term of water conservation as demonstrated by Gan et al. (2017). Our results clearly demonstrate that the positive rotational effect of lentils persisted for at least two years, while the fallow benefited only the following crop and only when precipitation was low. At the cropping system level, there was a significant site × crop stubble management interaction. Lentil increased system production (annualized DWEY) by 0–66 and 44–110% across different sites compared to spring wheat and fallow, respectively. The poor system production with fallow was attributed to (i) no yield in the fallow year (phase 1) and (ii) the significantly lower durum wheat yield in phase 3. The high system production in the lentil system was attributable to the positive lentil effects on oilseed crops in Year 2 and durum wheat in Year 3. As an annual legume crop, lentil supplies nitrogen to the following oilseed and cereal crops through positive rotational effects. As Beckie and Brandt (1997) reported, increased soil nitrogen is one of the main positive rotational effects of pulses on the following cereal crops. Niu et al. (2017) reported that annual pulse crops increased the following spring wheat yield by 18–26% in the study region. Similarly, a previous study found that replacing fallow with pulse crops increased system production by 35% (Gan et al., 2015). In addition, diversifying lentil-oilseed-durum wheat cropping systems was shown to improve biological processes, suppress pests, and enhance soil microbial diversity (Ellouze et al., 2014), which is the key driver for nutrient cycling in cropping systems (van Der Heijden et al., 2008). The enhanced ecosystem services of nitrogen cycling, pest control, and water conservation provided by lentil probably explain the yield benefits of systems involving lentil, at the cropping system level. The system involving fallow produced the highest oilseed yield but resulted in the lowest production at the cropping system level. In the real world, a farmer will determine the farm profitability at a system level in multiple years rather than just one year. This underlines the need for the assessment of crop production using a systems approach. In addition, the inclusion of lentil rather than fallow in cropping systems resulted in a greater yield advantage (44–110% greater) at the system level than at the durum wheat rotation phase (17% greater). The larger yield benefits obtained at the system level highlights the value of a systems approach that takes plant-soil feedback into account. The system production calculated using crop yields and prices might vary as a result of crop price fluctuation. Lentil price varied the most among all study crops and higher lentil price directly resulted in higher system production involving lentils. Most importantly, the results from

4.2. Oilseed crop species effects At the Year 3 durum wheat rotation phase, the oilseed crop species had similar effects on the following durum wheat yield. Alternating broad-leaf oilseeds crops with cereals is widely practiced to reduce pest pressure (Kirkegaard et al., 2004; Harker et al., 2014). However, the rotational effects of oilseed crop species on the following durum wheat were quite similar (Lenssen et al., 2012). Therefore, it is expected that different types of oilseed crops may have a comparable effect on the following cereal crops. At the cropping system level, however, the B. juncea produced either significantly higher or equal system production compared with the B. napus and C. sativa system across different sites. The yield of C. sativa (1823 kg ha−1) was 17% lower than B. juncea yield (2192 kg ha−1) and 41% lower than B. napus yield (3092 kg ha−1), explaining the significantly lower system production for the C. sativa system among the three oilseeds cropping systems. Across all sites, the B. juncea yield was 29% lower than the B. napus yield, but the price of B. juncea ($748 per tonne) was 62% higher than the price of B. napus ($463 per tonne), resulting in higher system production for the B. juncea crop. The price of B. juncea varied dramatically ranging from $639 to $893 per tonne during the study period. Sensitivity analysis indicated that at the lowest price of B. juncea, the system production was comparable between the B. juncea and B. napus systems, but significantly higher system production was obtained from the systems involving B. juncea when applying the high or average B. juncea prices. Based on the system production from the sensitivity analysis, it appears promising to grow B. juncea to diversify the existing oilseed cropping system, which was dominated by B. napus, and to increase the stability of the cropping system. 4.3. Environmental effects on system production Environmental factors such as temperature and precipitation interact with soil microbes, crops, and pests, thus affecting nutrient availability, plant health, and crop yield. Temperature is projected to increase 0.2 °C per decade at a global scale (Lobell et al., 2011), which likely increases evapotranspiration and reduces water availability for crops. Durum wheat has a higher water use efficiency than bread wheat as the crop can extract water from deep soil layers during grain-filling, thus enhancing drought tolerance (Wang et al., 2007), making durum wheat a good crop choice for dryland agriculture. As climate warms, the increased frequency of hot days and heat waves exacerbate the risk of drought, posing a higher risk of yield losses. Ray et al. (2015) reported that climate variability explained 34–45% of the wheat yield variation in the United States and Canada. To increase the stability of cropping system production under the climate uncertainty, adoption of drought tolerant crop species, such as durum wheat, pulse crops, and mustard, can mitigate the negative effects of climate change on the dominant cereal-oilseed cropping systems on the semiarid Canadian prairies. Crops in a multiple year rotation system experience different weather patterns, making it difficult to classify weather types at the cropping system level and to determine how weather affects the rotational outcomes. In the present study, the unique crop rotation design started a 3-year crop rotation each year from 2013 to 2015 at each site, forming a fully-phased crop rotation in 2015. The year of 2015 was 29

European Journal of Agronomy 105 (2019) 24–31

K. Liu, et al.

production in the lentil systems was related to large variations in N supplies from lentil crops under varying environmental conditions, as a result of the sensitivity of biological N2 fixation to weather (Hossain et al., 2016) and its impact on potential N mineralization. The highest yield potentials and high responsiveness of lentil to favorable weather indicated that these two lentil systems were suitable for high-yielding environments. The spring wheat-B. napus- durum wheat system, with a regression coefficient close to 1, was most stable and can be adapted to all the regions on the Canadian prairie. This is well supported by the fact that the cereal-oilseed crop rotation such as the wheat system in this study has been the dominant cropping system in the study region (Bueckert and Clarke, 2013; Martens et al., 2015), but this simplified cropping system might cause pest problems and the depletion of certain nutrients, putting the stability of oilseed-cereal production at risk. It is worth noting that this most stable system (e.g. wheat-B. napus) according to the dynamic stability also implied that the production of this system well represents the average production of all the cropping systems studied. However, the most stable system might not be an ideal system for all study regions, depending on growth environments. Diversifying cereal-oilseed cropping system with the consideration of local climate environments is necessary to maximize agricultural production without jepodizing yield stability. Unlike the dynamic stability described above, the two most responsive lentil systems showed the highest static stability as indicated by the lowest CV. The two fallow systems (fallow-B. juncea, fallow-C. sativa) had the lowest static stability as indicated by their highest CV. The static and dynamic stabilities complement each other with different focuses on stability. Static stability assesses the overall variation across different environments with less consideration of crop-environment interaction, regardless of environmental differences (Becker, 1981), while dynamic stability assesses the interaction between the crop and the environment across multiple sites, given that crop production varies in different environments in response to biotic and abiotic stresses (Dia et al., 2016). According to dynamic stability, the ideal cropping systems highly depends on specific environments, and the regression coefficient itself might not provide essential stability information without considering local environments. Caution is also needed when assessing stability using CV alone, since CV is not independent of mean as indicated by the Taylor’s power law (Taylor et al., 1999). A lower mean inflates CV but a higher mean deflates CV (Taylor et al., 1999; Döring et al., 2015). This might partially explain a relatively higher CV for the low-yielding systems such as fallow-B. juncea and fallow-C. sativa systems and a relatively lower CV for the high-yielding systems such as the lentil-B. juncea and lentil-B. napus systems. Overall, the lentil-B. juncea and lentil-B. napus systems showed the highest static stability across all study sites and were also the most productive under high-yielding environments. In this study, we have extended the concept of dynamic and static stability from individual crops to cropping systems, with the goal of assessing the production of cropping systems rather than individual crops. To better understand yield stability, it is recommended to use multiple stability indicators for a thorough assessment. This is especially true for cropping system stability assessment across environmental gradients, because of the complicated ecological interactions occurring at the cropping system level. Production stability assessment revealed the positive influence of lentil crops in stabilizing crop production for pulse-oilseed-wheat cropping systems under changing environments.

abnormally dry as evidenced by the fact that the growing-season precipitation in 2015 was 21–49% lower than their respective long-term average across the three study sites. We assumed that it is reasonable to classify the weather across temporal rotation cycles as a normalweather type at the cropping system level relative to any given extreme weather occurring in a fully-phased rotation year. With this assumption, we quantified the drought effect in 2015 on system production by comparing system production under normal weather. At the cropping system scale, the normal weather increased the average system production by 23% compared with the drought weather, with the lowest increase in the fallow system (3%) and the highest increase in the lentil system (47%). The fallow practice conserved more water for the following crops than a crop of lentil or wheat, providing early-growing season water supply, thus explaining the smallest production difference for the fallow practices between the normal and drought weather types. A previous study reported that nitrogen fixation by lentils varied from 23 kg ha−1 in the dry year to 87 kg ha−1 in the wet year on the Canadian prairies (Hossain et al., 2016). Also, the decomposition of nitrogen-rich lentil residues, including roots, and the associated soil nitrogen mineralization process are largely controlled by soil moisture, soil temperature, and soil microbes (Dessureault-Rompré et al., 2012). The temporal variation in nitrogen fixation and associated mineralization explained the large production difference for the lentil systems between weather types. In addition, drought resulted in a larger yield decrease at the durum wheat crop phase level (45%) than at the cropping system level (18%), supporting the conclusion that diversification at a cropping system level increases the stability of crop production. The present study further demonstrates that a fully-phased rotation system provides useful tools to assess rotational benefits under various weather conditions. Variance partitioning analysis showed that cropping systems explained 36% of the overall variation in system production and weather factors explained an additional 30% of the variation. Weather components such as temperature and precipitation are known as the major cause of inter-annual yield variation (Matiu et al., 2017). Iizumi and Ramankutty (2016) reported that more than 21% of the yield variation was related to the agro-climatic index. On the Canadian prairies, the pre- and early- growing season precipitations were reported to be critical for crop production (Meng et al., 2017), precipitation in April explained 22% of yield variation (Lychuk et al., 2017), and crop sequences accounted for up to 35% of yield variation (Niu et al., 2017). Water conservation practices such as fallowing and use of droughttolerant crop species might require less water to fulfill crop yield potentials, explaining the importance of climate adaptation in stabilizing crop production. Adaptation is considered a key strategy for mitigating the impact of changing environments on crop production (Altieri et al., 2015). Our variance partitioning analysis suggests that managing water is the key to enhancing production stability on the Canadian prairies under a changing climate. 4.4. Stability of cropping systems Our analysis showed that the fallow-based systems had aboveaverage stability among all systems evaluated. However, the lowest regression coefficients related to the environmental index suggested that these fallow-based systems were the least responsive to changing environments such as variations of precipitation, since the fallow practices conserved water for the following crops (Lafond et al., 1992), reducing the yield reliance on precipitation. Along with the low system production, the above-average stability suggested that the fallow systems were a good option for low-yielding environments. The lentil - B. juncea - durum wheat and lentil - B. napus - durum wheat systems had the largest regression coefficients, indicating that these two lentil systems had below-average stability, but the higher regression coefficients suggested that these two lentil systems were most responsive to the changing environments. Highly responsive

5. Conclusions The positive rotational effect of lentil crops on productivity persisted for at least two years in the lentil-oilseed-durum wheat cropping system compared with the conventional cereal-oilseeds cropping system. The higher system productivity, expressed by annualized 30

European Journal of Agronomy 105 (2019) 24–31

K. Liu, et al.

durum wheat equivalent yield, in the cropping system involving B. juncea made B. juncea a good substitute for the dominant oilseed crop of B. napus to diversify crop rotations and increase yield stability. The lentil-B. juncea-durum wheat and lentil-B. napus-durum wheat systems were most stable across all study environments as indicated by CV, and both were most responsive to soil-climatic factors and suitable for highyielding environments as indicated by regression coefficients. The static and dynamic stability analyses complement each other, and an integration of the two stability assessments is proposed for a thorough evaluation of cropping system stability. Variance partitioning indicated that 36% of system production variation was attributed to crop rotations and an additional 30% was attributed to environmental factors. We conclude that lentils improve the productivity and stability of the existing oilseed-cereal cropping systems under changing environments.

95, 9–20. Hatfield, J.L., Boote, K.J., Kimball, B.A., Ziska, L.H., Izaurralde, R.C., Ort, D., Thomson, A.M., Wolfe, D., 2011. Climate impacts on agriculture: implications for crop production. Agron. J. 103, 351–370. Hossain, Z., Wang, X., Hamel, C., Knight, J.D., Morrison, M.J., Gan, Y., 2016. Biological nitrogen fixation by pulse crops on semiarid Canadian prairies. Can. J. Plant Sci. 97, 119–131. Iizumi, T., Ramankutty, N., 2016. Changes in yield variability of major crops for 19812010 explained by climate change. Environ. Res. Lett. 11, 034003. Kirkegaard, J.A., Simpfendorfer, S., Holland, J., Bambach, R., Moore, K.J., Rebetzke, G.J., 2004. Effect of previous crops on crown rot and yield of durum and bread wheat in northern NSW. Aust. J. Agric. Res. 55, 321–334. Lafond, G.P., Loeppky, H., Derksen, D.A., 1992. The effects of tillage systems and crop rotations on soil water conservation, seedling establishment and crop yield. Can. J. Plant Sci. 72, 103–115. Lemke, R.L., Vandenbygaart, A.J., Campbell, C.A., Lafond, G.P., McConkey, B.G., Grant, B., 2012. Long-term effects of crop rotations and fertilization on soil C and N in a thin Black Chernozem in southeastern Saskatchewan. Can. J. Soil Sci. 92, 449–461. Lenssen, A.W., Iversen, W.M., Sainju, U.M., Caesar-TonThat, T.C., Blodgett, S.L., Allen, B.L., Evans, R.G., 2012. Yield, pests, and water use of durum and selected crucifer oilseeds in two-year rotations. Agron. J. 104, 1295–1304. Lin, B.B., 2011. Resilience in agriculture through crop diversification: adaptive management for environmental change. Bioscience 61, 183–193. Lobell, D.B., Schlenker, W., Costa-Roberts, J., 2011. Climate trends and global crop production since 1980. Science 333, 616–620. Luo, Q., 2011. Temperature thresholds and crop production: a review. Clim. Change 109, 583–598. Lychuk, T.E., Moulin, A.P., Lemke, R.L., Gossen, B.D., Leeson, J.Y., Kirk, A., Johnson, E.N., Olfert, O.O., Brandt, S.A., Thomas, A.G., 2017. Effects of crop inputs, diversity, environment, and terrain on yield in an 18-yr study in the semi-arid Canadian Prairies. Can. J. Plant Sci. 97, 715–730. Martens, J.R.T., Entz, M.H., Wonneck, M.D., 2015. Review: redesigning Canadian prairie cropping systems for profitability, sustainability, and resilience. Can. J. Plant Sci. 95, 1049–1072. Matiu, M., Ankerst, D.P., Menzel, A., 2017. Interactions between temperature and drought in global and regional crop yield variability during 1961–2014. PLoS One 12, e0178339. Médiène, S., Valantin-Morison, M., Sarthou, J.P., De Tourdonnet, S., Gosme, M., Bertrand, M., Roger-Estrade, J., Aubertot, J.N., Rusch, A., Motisi, N., Pelosi, C., Doré, T., 2011. Agroecosystem management and biotic interactions: a review. Agron. Sustain. Dev. 31, 491–514. Meng, T., Carewa, R., Florkowski, W.J., Klepacka, A.M., 2017. Analyzing temperature and precipitation influences on yield distributions of canola and spring wheat in Saskatchewan. J. Appl. Meteorol. Climatol. 56, 897–913. Niu, Y., Bainard, L.D., Bandara, M., Hamel, C., Gan, Y., 2017. Soil residual water and nutrients explain about 30% of the rotational effect in 4-yr pulse-intensified rotation systems. Can. J. Plant Sci. 97, 852–864. Obour, A.K., Chen, C., Sintim, H.Y., McVay, K., Lamb, P., Obeng, E., Mohammed, Y.A., Khan, Q., Afshar, R.K., Zheljazkov, V.D., 2018. Camelina sativa as a fallow replacement crop in wheat-based crop production systems in the US Great Plains. Ind. Crops Prod. 111, 22–29. Oksanen, J., Blanchet, F.G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., Minchin, P.R., O’Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H., Szoecs, E., Wagner, H., 2017. Vegan: Community Ecology Package. R Package Version 2.4-4. https:// CRAN.R-proejct.org/package=vegan. Pandey, P., Irulappan, V., Bagavathiannan, M.V., Senthil-Kumar, M., 2017. Impact of combined abiotic and biotic stresses on plant growth and avenues for crop improvement by exploiting physio-morphological traits. Front. Plant Sci. 8, 537. Qian, B., De Jong, R., Gameda, S., Huffman, T., Neilsen, D., Desjardins, R., Wang, H., McConkey, B., 2013. Impact of climate change scenarios on Canadian agroclimatic indices. Can. J. Soil Sci. 93, 243–259. Raseduzzaman, M., Jensen, E.S., 2017. Does intercropping enhance yield stability in arable crop production? A meta-analysis. Eur. J. Agron. 91, 25–33. Ray, D.K., Gerber, J.S., Macdonald, G.K., West, P.C., 2015. Climate variation explains a third of global crop yield variability. Nat. Commun. 6. https://doi.org/10.1038/ ncomms6989. Shennan, C., 2008. Biotic interactions, ecological knowledge and agriculture. Philos. Trans. Biol. Sci. 363, 717–739. Statistics Canada, 2017. Cansim. http://www5.statcan.gc.ca/cansim/a29?lang=eng& groupid=001&p2=17. Taylor, S.L., Payton, M.E., Raun, W.R., 1999. Relationship between mean yield, coefficient of variation, mean square error, and plot size in wheat field experiments. Commun. Soil Sci. Plant Anal. 30, 1439–1447. Urruty, N., Tailliez-Lefebvre, D., Huyghe, C., 2016. Stability, robustness, vulnerability and resilience of agricultural systems. A review. Agron. Sustain. Dev. 36, 15. van Der Heijden, M.G.A., Bardgett, R.D., Van Straalen, N.M., 2008. The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecol. Lett. 11, 296–310. Wang, H., McCaig, T.N., DePauw, R.M., Clarke, J.M., Lemke, R., 2007. Water use of some recent bread and durum wheat cultivars in western Canada. Can. J. Plant Sci. 87, 289–292.

Acknowledgments The project was funded by Mustard Innovation Canadian Advantage, Agri-Innovation Program Stream B of ‘Growing Forward II’ of Agriculture and Agri-Food Canada. The authors appreciate the excellent technical support from A. Kapiniak, C. Gampe, L. Molnar, L. Luan, and L. Poppy. References Altieri, M.A., Nicholls, C.I., Henao, A., Lana, M.A., 2015. Agroecology and the design of climate change-resilient farming systems. Agron. Sustain. Dev. 35, 869–890. Becker, H.C., 1981. Correlations among some statistical measures of phenotypic stability. Euphytica 30, 835–840. Becker, H.C., Léon, J., 1988. Stability analysis in plant breeding. Plant Breed. 101, 1–23. Beckie, H.J., Brandt, S.A., 1997. Nitrogen contribution of field pea in annual cropping systems. 1. Nitrogen residual effect. Can. J. Plant Sci. 77, 311–322. Berzsenyi, Z., Győrffy, B., Lap, D., 2000. Effect of crop rotation and fertilisation on maize and wheat yields and yield stability in a long-term experiment. Eur. J. Agron. 13, 225–244. Blackshaw, R., Johnson, E., Gan, Y., May, W., McAndrew, D., Barthet, V., McDonald, T., Wispinski, D., 2011. Alternative oilseed crops for biodiesel feedstock on the Canadian prairies. Can. J. Plant Sci. 91, 889–896. Bueckert, R.A., Clarke, J.M., 2013. Review: annual crop adaptation to abiotic stress on the Canadian prairies: six case studies. Can. J. Plant Sci. 93, 375–385. Campbell, C.A., Zentner, R.P., Gameda, S., Blomert, B., Wall, D.D., 2002. Production of annual crops on the Canadian prairies: trends during 1976–1998. Can. J. Soil Sci. 82, 45–57. Cutforth, H.W., McGinn, S.M., McPhee, K.E., Miller, P.R., 2007. Adaptation of pulse crops to the changing climate of the Northern Great Plains. Agron. J. 99, 1684–1699. Dessureault-Rompré, J., Zebarth, B.J., Burton, D.L., Georgallas, A., Sharifi, M., Porter, G.A., Moreau, G., Leclerc, Y., Arsenault, W.J., Chow, T.L., Grant, C.A., 2012. Prediction of soil nitrogen supply in potato fields using soil temperature and water content information. Soil Sci. Soc. Am. J. 76, 936–949. Dia, M., Wehner, T.C., Arellano, C., 2016. Analysis of genotype × environment interaction (G×E) using SAS programming. Agron. J. 108, 1838–1852. Döring, T.F., Knapp, S., Cohen, J.E., 2015. Taylor’s power law and the stability of crop yields. Field Crops Res. 183, 294–302. Ellouze, W., Esmaeili Taheri, A., Bainard, L.D., Yang, C., Bazghaleh, N., Navarro-Borrell, A., Hanson, K., Hamel, C., 2014. Soil fungal resources in annual cropping systems and their potential for management. Biomed Res. Int. 2014, 531824. Finlay, K.W., Wilkinson, G.N., 1963. The analysis of adaptation in a plant-breeding programme. Aust. J. Agric. Res. 14, 742–754. Gan, Y., Angadi, S.V., Cutforth, H., Potts, D., Angadi, V.V., McDonald, C.L., 2004. Canola and mustard response to short periods of temperature and water stress at different developmental stages. Can. J. Plant Sci. 84, 697–704. Gan, Y., Hamel, C., O’Donovan, J.T., Cutforth, H., Zentner, R.P., Campbell, C.A., Niu, Y., Poppy, L., 2015. Diversifying crop rotations with pulses enhances system productivity. Sci. Rep. 5, 14625. Gan, Y., Hamel, C., Kutcher, H.R., Poppy, L., 2017. Lentil enhances agroecosystem productivity with increased residual soil water and nitrogen. Renew. Agr. Food Syst. 32, 319–330. Gliessman, S.R., 2004. Integrating agroecological processes into cropping systems research. J. Crop Improve. 11, 61–80. Government of Saskatchewan, 2017. AGR Market Trends. http://applications. saskatchewan.ca/Default.aspx?DN=6dd2ea91-22e6-4095-9d2a-67ef8c822897&l= English. Harker, K.N., O’Donovan, J.T., Turkington, T.K., Blackshaw, R.E., Lupwayi, N.Z., Smith, E.G., Johnson, E.N., Gan, Y., Kutcher, H.R., Dosdall, L.M., Peng, G., 2014. Canola rotation frequency impacts canola yield and associated pest species. Can. J. Plant Sci.

31