Improving weed control in sustainable agro-ecosystems: role of cultivar and termination timing of rye cover crop

Alternative strategies to control weeds are required at field level to reduce herbicides and derived pollution. Rye (Secale cereale L.) cultivation as cover crop is adopted mainly because of its allelopathic weed control, which takes place throughout a strong inhibition of germination and seedling growth in several grass and broad-leaved weeds. The present study consisted of: i) a field trial, focused on evaluation of biomass production and allelochemical concentration in the biomass, and in situ weed control at 30 days after termination (with two termination timings: T1 heading phase and T2 10 days later) of 8 rye varieties; ii) a pot experiment, focused on the inhibition effect of mulches derived by those 8 rye varieties on four summer weeds: velvetleaf (Abutilon theophrasti Med.), lambsquarters (Chenopodium album L.), redroot pigweed (Amaranthus retroflexus L.), and common purslane (Portulaca oleracea L). Results showed that biomass production was the highest with Protector, closely followed by Primizia, Sito 70, Hellvus, Forestal, and Hymonta. In any case, rye mulching always reduced the weed biomass, especially with Fasto and Forestal. The allelochemical concentration in the biomass was the highest with Fasto and Forestal, and decreased on average from T1 to T2 (-38% for total BX and -57% for isovitexin). Conversely, the rye biomass production increased (on average + 77%) passing from T1 to T2. We found also that the reduction of weed biomass, compared with the control, is highly Ac ce pt ed p ap er correlated with the allelochemical content in rye biomass in the case of T1 termination, while with the biomass production in the case of T2. In pots, a strong inhibitory effect on seedling growth due to rye mulching was observed for C. album (-76%), A. retroflexus (-56%), and P. olearcea (-84%), while not for A. theophrasti. We concluded that, whatever the variety, adopting rye as cover crop may be considered as a suitable practice to reduce weed pressure at the field level. Among all the varieties tested, Forestal and Protector showed the greatest weed suppression potential, as a consequence of high amount of allelochemicals production for Forestal, and high biomass production for Protector. Introduction Weed control strategies based on the use of herbicides are expensive and may affect negatively the quality of soil, water and air (Felsot et al., 2011). The excessive use of chemical herbicides in the last decades led to the development of herbicide resistance: 262 species of herbicide-resistant weeds have been detected on 93 crops in 70 countries (Beckie, 2020). Consequently, a growing interest in alternative strategies for weed management has been stimulated worldwide to address current economic and environmental challenges of crop production (Kumar et al., 2020). In addition, the European Commission recently stated ambitious goals for reducing the herbicide use (-50%) at the field level by 2030 (European Commission, 2021). The use of cover crops (CCs) has long been indicated as a good solution for limiting weed development in a broad series of agro-ecosystems, thus reducing herbicide use and cost (Barnes and Putnam, 1983). It was shown that selected CC species (e.g. gramineous plants) may suppress weeds due to their high competitiveness for space, light, water and nutrient use (Hiltbrunner et al., 2007). Also, CC residues that remains onto the soil surface can limit weed germination and development by reducing light transmittance and soil temperature (Wayman et al., 2015). In addition, some CCs produce allelochemical compounds as secondary metabolites [(e.g. benzoxazinoids, flavonoids, alkaloids, etc. (Scavo et al., 2019)], which are released both by living or decaying plant tissues, and Ac ce pt ed p ap er exert a strong control on weeds (Liebman and Davis, 2015). A plant species considered as a reference for the allelopathic weed control is rye (Secale cereale L.) (Tabaglio et al., 2008). Rye is one of the most commonly used winter cover crops, due to high biomass production, soil and climate adaptability, nitrous oxide emission mitigation, and weed suppression potential (Gavazzi et al., 2010; Fiorini et al., 2020). After termination, rye mulch can be left on the soil surface (i.e. in the case of no-till soil management) to ensure a persistent inhibitory effect on weed germination and development (Mirsky et al., 2013; Tabaglio et al., 2013). Beyond the suppressing effect of rye mulch on weeds, which highly depends on the thickness of the mulch layer (Teasdale and Mohler, 2000), rye residue maintained on the soil surface was reported to release the benzoxazinoids 2,4-dihydroxy-1,4 (2H)-benzoxazin-3-one (DIBOA) and benzoxazolin-2(3H)-one (BOA) (Schulz et al., 2013). Those allelochemical compounds strongly inhibit the germination and seedling growth of several grass and broad-leaved spontaneous plants (Macías et al., 2019). Under field conditions, the amount of benzoxazinoids (BX) released by rye is estimated to range between 0.5 and 5 kg ha-1, depending on variety and date of phenological phase of termination (Reberg-Horton et al., 2005). Rice et al. (2005) reported that BX concentration consistently decreases in rye tissues from tillering to flowering. However, previous studies did not consider the component of benzoxazinoids bound to the cell wall: this component is probably not available immediately after rye termination, but it can be released gradually afterwards, affecting the long-term allelochemical potential. Besides benzoxazinoids, also flavonoids have a high allelopathic potential: isovitexin was isolated in flavonoid fraction of oat (Avena sativa L.) (de Bertoldi et al., 2009), which demonstrated a good control of weeds. A clear assessment on relationships between rye varieties and allelochemical concentration under field conditions is missing in our soil-climate conditions and only few studies previously compared the effect of termination time on allelochemical concentration in different rye varieties, evaluating the effect on weed development under both field and greenhouse conditions. In addition, previous Ac ce pt ed p ap er studies did not assess the concentration of cell wall-bound BOA, which may contribute to a long-term allelopathic effect. The present study aims to: (i) identify the best rye varieties for benzoxazinoids and isovitexin concentrations and production in order to improving sustainable weed control, (ii) find the combination rye variety × termination timing that can exert the highest and lasting level of weed suppression. We hypothesized that terminating the cover crop early (at the heading phase) may enhance the alleochemical content and, consequently, the control of weeds. Materials and methods Field experiment The field trial was carried out at CREA Institute, located in Fiorenzuola d’Arda, Piacenza, Northern Italy (lat. 44.926383 N; long. 9.890661 E), from October 2014 to July 2015. Soil characteristics (0-30 cm soil depth) at the beginning of the experiment are reported in Table 1. The climate is temperate; mean annual rainfall and temperature are 774 mm and 12.8° C, respectively. Climatic data during the experiment were collected from an automated meteorological station situated close to the experimental field (Fig. S1). The experiment compared 8 treatments, corresponding to 8 rye varieties (6 cultivars: Dukato, Fasto, Forestal, Primizia, Protector, and Sito 70; 2 hybrids: Hellvus and Hymonta), each one replicated 4 times, for a total of 32 main plots (plot area of 10.2 m2). Then, each plot was divided into two subplots (sub-plot area of 5.1 m2) according to the two selected rye termination timings (i.e. at heading phase [T1] and 10 days after the full heading phase [T2]), thus obtaining a total number of 64 subplots. As a result, a split-plot (SP) experimental design was obtained, where the main factor was the rye variety, and the secondary factor was the rye termination timing. Rye was sown on 29th October 2014, using an 8-row plot-seeder (mod. Wintersteiger), the seeding rate was 110 kg ha-1. Four more plots (one for each block) without rye were used as a control. In the Ac ce pt ed p ap er control plots the vegetation was let to grow spontaneously. Fertilization was carried out twice: (1) before sowing using a ternary fertilizer (15-15-15) applying 45 kg ha-1 of N, 20 kg ha-1 of P and 37 kg ha-1 of K; (2) at the end of the tillering using ammonium nitrate (26% N) distributing 52 kg ha-1 of N. The phenological and yield data were determined in the sub-plot area according to the following termination timing: (T1) at the heading phase and (T2) 10 days after. The aboveground biomass production was obtained cutting manually the whole sub-plots biomass at the two termination timings. Prior to each cut, plant height was determined by measuring 20 randomly chosen plants from the ground level, for each subplot. After cutting, rye aboveground biomass was weighted in the field. A sub-sample of about 800 g was collected to be used (i) as rye mulch for the pot trial (only for subplots harvested at T1, as described below), and (ii) for the allelochemical analyses on a dry matter basis, after being oven-dried at 65 °C to constant weight. The remaining biomass was left onto the soil as mulch for evaluating the effect on weed development directly into the field. Weed density was measured in each sub-plot placing twice randomly a circle of 0.125 m2 30 days after both rye termination timing. Weeds in each sampling area were collected and oven-dried at 105 °C for 48 hours for calculating the total dry weight per plot. Greenhouse pot experiment A greenhouse experiment was set up using rye biomass harvested at T1. The experiment was conducted immediately after rye harvest and it was designed as a randomized complete block (RCB) with 8 treatments and 6 replicates. The 8 treatments were represented by the 8 rye varieties used in the field trial. The experiment was conducted using plast