Alternative Options to Glyphosate for Control of Large Echinochloa colona and Chloris virgata Plants in Cropping Fallows

The over-reliance on the herbicide glyphosate for knockdown weed control in fallows under minimum and zero-till cropping systems has led to an increase in populations of glyphosate-resistant weeds. Echinochloa colona and Chloris virgata are two major grass weeds in the cropping regions of northern New South Wales and southern Queensland, Australia, that have become harder to kill due to a steady rise in the occurrence of glyphosate-resistant weed populations. Therefore, to help growers contain these hard to kill fallow weeds, an alternate approach to glyphosate application is needed. With this purpose in mind, a pot study was carried out during the summer seasons of 2015 and 2016 at the Tamworth Agricultural Institute, Tamworth, NSW, Australia, to evaluate the efficacy of tank mixtures and sequential applications of Group H (4-hydroxyphenylpyruvate dioxygenase (HPPD) inhibitor), Group C (inhibitors of photosynthesis at photosystem II), Group A (ACCase inhibitors) and Group L (photosystem I inhibitor) herbicides on late tillering E. colona and C. virgata plants. These herbicide groups are a global classification by the Herbicide Resistance Action Committee. Highly effective results were achieved in this study using combinations of Groups H, C, A and L herbicides applied as tank mixtures for controlling large E. colona plants. Additionally, sequential applications of Group H, C and A herbicides followed by (fb) paraquat were shown to be very effective on large E. colona plants. Late tillering C. virgata plants were generally well controlled by tank mixtures, and sequential applications proved to be highly effective on this grass weed as well. Haloxyfop in combination with paraquat as a tank mixture, via sequential application or as a stand-alone treatment, was highly effective for C. virgata control; however, using combinations of herbicide groups is the preferred choice when combating herbicide resistant weed populations. There was a clear synergy shown using Group H, Group C and Group A herbicides in combination with the Group L herbicide paraquat in this study for controlling advanced E. colona and C. virgata plants. These combinations were shown to be successful on plants grown under glasshouse conditions; however; these treatments would need to be tested on plants grown in a field situation to show whether they will be a useful solution for farmers who are trying to control these weeds in fallow.


Introduction
Echinochloa colona (L.) Link and Chloris virgata SW. are two major fallow weeds that have proliferated under no-till regimes in the crop growing regions of New South Wales (NSW) and Queensland (QLD), of preserving moisture for subsequent crops. Therefore, the sequential application method has evolved into two applications of herbicide using different MOA groups for the first and second application. The length of time separating the two applications will vary according to what weeds are being sprayed, the growth stage of weeds and what herbicide MOA groups are being used [2]. Generally, the second application will be applied between 1 and 14 days after the first application. The sequential application method has the potential to slow down the spread of glyphosate resistance. Modelling work on Lolium rigidum Gaud. predicted glyphosate resistance in only 17 of the 1000 simulation runs using the scenario of glyphosate followed by paraquat [15]. E. colona plants at varying growth stages can be well controlled using the sequential application technique. Over five experiments, a sequential application of glyphosate followed by paraquat achieved 100% control of E. colona plants ranging from early to late tillering in the growth stage [16]. In one of the experiments, control dropped to between 96 and 99.4%, which was likely due to the high density of plants being sprayed which may have impeded proper coverage of plants with herbicide [16]. The sequential application technique also worked well on three growth stages of L. rigidium, using glyphosate as the first application and paraquat as the second. In this study, sequential applications were more effective than using glyphosate or paraquat as a single treatment [17]. The sequential application method works well on both grasses and broadleaf weeds. In a previous study, a Group I (global classification by Herbicide Resistance Action Committee) herbicide (disrupters of plant cell growth) followed by an application of paraquat 7 days later on Conyza bonariensis (L.) Cronq. plants proved to be an effective combination when compared to single applications of herbicides [18]. Sequential application is an effective tool to control populations of herbicide-resistant weeds and hard to kill and/or glyphosate-tolerant weeds. Additionally, this method can help alleviate selection pressure for resistance in selective in-crop herbicides. Reduced numbers of herbicide-resistant weeds resulting from the effectiveness of using sequential applications contribute to less selection pressure for the selective herbicides that are used in-crop [2].
The use of herbicide mixtures can be an effective part of an integrated approach to dealing with resistant weed populations, provided herbicide combinations come from different MOA groups and are synergistic [19]. Herbicide mixtures also have the potential to reduce the chance of selecting for resistance in herbicide-resistant weeds, if the herbicides used are from different MOA groups [20]. Under a modelling scenario, tank mixtures of herbicides from two or more MOA groups appeared to be more effective at slowing down the onset of resistance than rotating herbicides from different groups, especially when areas of weed populations were 100 hectares or less [21]. A tank mixture of propanil with either anilofos or piperophos applied to propanil-resistant Echinochloa crus-galli (P.) Beav. proved to be very effective in controlling this weed in southern United States rice crops, with limited damage to the crop [22]. The continued use of single applications of herbicides to control fallow weeds has the potential to expedite the onset of herbicide resistance. The longevity of herbicides such as glyphosate may be prolonged if mixed with herbicides from other MOA groups when applying to fallow weeds such as Sonchus oleraceus L. [23]. A tank mixture of paraquat plus imazapic was shown to be very effective for the control of E. colona plants in two field experiments conducted on at Pittsworth and Dalby, the Darling Downs in north-eastern Australia [16]. Control levels ranged from 96-99% when assessed using a biomass reduction score (% of untreated plots).
A tank mixture of the Group H (4-hydroxyphenylpyruvate dioxygenase (HPPD) inhibitor) herbicide tembotrione and Group C (photosystem II inhibitor) herbicide atrazine was synergistic for controlling large Digitaria sanguinalis (L.) Scop, Panicum miliaceum L., Chenopodium album L., and Abutilon theophrasti plants in maize crop experiments in Illinois, Oregon, Washington, U.S.A. and Ontario, Canada [24]. This tank mix treatment was superior by 3% to 45% to single applications for weed control in this experiment. Control of propanil-resistant E. colona was improved by tank mixing propanil with the Group D (microtubule inhibitors) herbicide pendimethalin in glasshouse and field experiments in Costa Rica [25]. Tank mixtures of pendimethalin and propanil lowered rates of herbicide actives needed to control three-to-four leaf resistant E. colona plants. Pendimethalin and propanil tank mixed at rates of 0.23 kg ha −1 and 1.08 kg ha −1 , respectively, exhibited similar levels of control to 2.16 kg ha −1 of propanil applied alone [25].
Paraquat is routinely used as the final herbicide in the sequential application technique, predominantly because of its dissimilar MOA to the systemic herbicides used in the first application. Paraquat is widely reported as being effective for this purpose and plays a significant part in this successful tactic for controlling hard to kill and or glyphosate-resistant weeds. A tank mix of systemic herbicides followed by a paraquat-based herbicide as the second application 7 days later, was found to be very effective for controlling varying growth stages of C. bonariensis plants [26]. The use of sequential application treatments using paraquat to control glyphosate-resistant E. colona proved to be potent as well, showing the usefulness of this tactic on fallow grass weeds [8]. A study on C. bonariensis using the sequential application technique using paraquat as the secondary herbicide was very effective for controlling this problematic weed [18]. Glyphosate-resistant S. oleraceus can also be successfully controlled in the field using paraquat as the second application [27].
The following pot study was carried out to gauge whether tank mixtures and/or sequential applications specifically using Group H (4-hydroxyphenylpyruvate dioxygenase (HPPD) inhibitor), Group C (inhibitors of photosynthesis at photosystem II), Group A (ACCase inhibitors) and Group L (photosystem I inhibitor) herbicides are effective for controlling advanced E. colona and C. virgata plants. Advanced plants, nearing or in a reproductive stage, provide a more robust test for treatments than smaller plants and more closely resemble the portion of larger weeds farmers would have in their paddocks after harvesting crops.

Experimental Location and Potting Mix Type
The glasshouse experiments were conducted during summer of 2015 and 2016 at the Tamworth Agricultural Institute, Tamworth (31.0927 • S, 150.9320 • E), NSW, Australia. The glasshouse was not temperature controlled; however, evaporative cooling did avoid extremes of hot and cold temperatures. The potting mix used for germination of weed seedlings and the duration of the experiment was a premium mix made by Scotts Australia Pty Ltd, Bella Vista, Australia. The potting mix comprised of composted hardwood sawdust and pine bark, cocopeat fibre and nutrients consisting of calcium, ammonium nitrate, gypsum, iron sulfate, potassium sulfate, phosphorus and sulphur. In addition, this potting mix contained water storage crystals and a wetting agent.

Weed Plant Management
Seeds of E. colona were collected from plants grown in the glasshouse in the autumn of 2014 from a population originally sourced from a property near Bellata (29.9019 • S, 149.7884 • E), NSW, Australia. The C. virgata seeds were collected off plants located in paddocks on the Tamworth Agricultural Institute research farm in 2015. E. colona was susceptible to glyphosate but C. virgata is not mentioned on the label of glyphosate. After collection, seeds from both species were placed in paper envelopes and stored at room temperature. To commence the experiments, seeds were germinated in shallow plastic trays filled with premium potting mix. When seedlings reached the three-leaf growth stage, they were transplanted as single plants into 8 cm square and 15 cm deep pots. These potted plants were then placed in trays with water so as to eliminate moisture stress, and fertilised with a soluble plant fertiliser every 3 weeks up until plants reached late tillering to inflorescence emergence (Z50-59). The fertiliser comprised of 25% nitrogen, 5% phosphorous, 8.8% potassium, 4.6% sulphur, 0.5% magnesium, 0.18% iron, 0.01% manganese, 0.005% boron, 0.005% copper, 0.004% zinc and 0.001% molybdenum. Once plants reached late tillering to inflorescence emergence (Z50-59), they were moved into trays containing water outside the glasshouse for 14 days (d) before herbicide application. This process was carried out so that plants were exposed to weather conditions similar to paddock plants before herbicide application.

Treatments and Experimental Design
There were four tank mixture application and four sequential application experiments carried out in this study. The treatments were the same for both E. colona and C. virgata; thus, there was a tank mixture application and a sequential application experiment for each species in each year. This resulted in four experiments per year and eight experiments in total over the two years. The herbicides, mode of action (MOA) herbicide group and herbicide doses are shown in Tables 1 and 2. The herbicide treatments were applied using a powered backpack sprayer with a 2 m wide four-nozzle boom, delivering 100 L ha −1 of spray solution. Nozzles used were flat fan with an orifice size of 02 (ASAE S-572 spray tip classification) and a spray coverage angle of 110 degrees. After herbicide application, pots were taken back into the glasshouse for the duration of the experiment. The tank mixture experiments had 12 treatments, comprising of 11 herbicide treatments and one nontreated, with every treatment replicated six times. The sequential application experiment had 13 treatments, comprising of 12 herbicide treatments and one nontreated, with every treatment replicated six times. The timing of the sequential application treatments was 7 d between the first and second applications. Tank mixture treatments involved mixing of herbicides in a single tank mixture and spraying as one application. Abbreviations: fb = followed by 2nd application of herbicide 7 days later.

Sampling Techniques and Observations
Herbicide efficacy was evaluated at 42 d after herbicide treatment for each experiment. Plants alive at 42 d after herbicide application had any remaining green plant material extracted and placed into a separate paper bag for drying. The paper bags containing the green plant material were then oven-dried at 80 • C for a minimum of 72 h. At the completion of the drying period, plant material was removed from the paper bag and weighed. Plants that were completely dead at 42 d after treatment were given a value of zero for green biomass.

Statistical Analyses
Experiments were conducted using a completely randomized design with six replications in each year. Experimental data was analysed using a generalised linear modelling program called ASReml (VSN International Ltd, Hemel Hempstead, UK). Comparisons of means were implemented based on the least significant difference test at 0.05 probability. Original values were included in the analysis as transformation did not significantly improve homogeneity of variance. Therefore, original data values are presented. Data are presented separately for each year to better understand variation between the years.

Effect of Tank Mixture Herbicide Treatments on Echinochloa colona
The tank mixtures of isoxaflutole plus paraquat, atrazine plus paraquat, simazine plus paraquat and terbuthylazine plus paraquat were very effective in controlling large E. colona plants, with complete control at 42 d after treatment in the 2015 experiment. These treatments were also effective in the 2016 experiment, exhibiting high control levels on large E. colona plants (Table 3). Additionally, the haloxyfop plus paraquat tank mixture was very effective and had similar results compared to the other tank mixture treatments over both experiments. The tank mixture treatments gave better (p < 0.05) control than the single herbicide treatments with the exception of paraquat. Paraquat as a standalone treatment was very effective over both experiments with similar efficacy levels to the tank mixture treatments. The single herbicide treatments of isoxaflutole, atrazine, simazine, terbuthylazine and haloxyfop had lower weed biomass compared to nontreated E. colona; however, they were inferior to the tank mix treatments. Furthermore, haloxyfop was similar to the tank mixture treatments in the 2016 experiment with 91% control. Predominantly, the tank mixture treatments were clearly superior to the single herbicide treatments. The addition of paraquat to the single herbicides and combining as a tank mix appears to have had a synergistic effect for controlling large E. colona plants. Paraquat as a stand-alone treatment was similar in efficacy to the tank mixture treatments, which may question the practice of using herbicides in combination when paraquat is just as potent. One feasible justification for using a tank mixture over paraquat is that the use of herbicides from different MOA groups combined as a tank mixture will slow down selection to resistance [19,20]. Furthermore, tank mixtures of paraquat with systemic herbicides may have improved performance when compared to paraquat alone. Previous studies on control of glyphosate-resistant C. canadensis in a fallow paddock at Leland, Mississippi, USA pointed to tank mixtures being more effective than paraquat as a stand-alone treatment [28]. Tank mixtures of paraquat plus 2,4-D and paraquat plus dicamba resulted in 88 and 89% control levels, respectively, compared to 63% control for the paraquat treatment [28].

Effect of Sequential Application Herbicide Treatments on Echinochloa colona
All of the sequential treatments were highly effective for controlling large E. colona plants over the 2015 and 2016 experiments. Isoxaflutole fb an application of paraquat 7 days later achieved 99% control in 2015 and 100% control in 2016 at 42 d (Table 4). Atrazine fb paraquat, simazine fb paraquat, terbuthylazine fb paraquat, haloxyfop fb paraquat and paraquat fb paraquat all achieved very high levels of control on large E. colona at 42 d. Paraquat was the only single product treatment that was effective in controlling E. colona with control levels similar to the sequential treatments. The remaining single treatments of isoxaflutole, atrazine, simazine, terbuthylazine and haloxyfop were inferior to the sequential application treatments and the single treatment of paraquat. Values in parentheses are percentage control relative to nontreated. Abbreviations: fb = followed by 2nd application of herbicide 7 days later; LSD, least significant difference at the 5% level of significance (0.05).
Generally, results were mainly consistent across both experiments with the exception of few treatments which showed a sizeable disparity between the 2015 and 2016 experiments. A possible reason for this disparity in results may be due to differences in humidity when treatments were applied to plants from the separate experiments. If temperatures are similar but humidity is significantly different, especially when spraying in warmer times of the year, then herbicide uptake can be affected [29]. There have been documented cases of this occurring when using the herbicide glufosinate on grass species such as Hordeum vulgare L. cv. 'Samson' and Setaria viridis (L.) Beauv [29]. In a glasshouse study, it was found that the efficacy of glufosinate ammonium applied at 100 g ha −1 on S. viridis was significantly reduced in conditions with lower humidity, even though temperatures were similar. When this herbicide was applied at 40% relative humidity, the S. viridis plants accrued 70% of the dry weight of nontreated plants; however, when the same herbicide at the same rate was applied under 95% relative humidity, plants were totally killed [30]. Differential plant size between the 2015 and 2016 experiments may also influence herbicide efficacy.

Effect of Tank Mixture Herbicide Treatments on Chloris virgata
The tank mix treatments of isoxaflutole plus paraquat, atrazine plus paraquat, simazine plus paraquat, terbuthylazine plus paraquat and haloxyfop plus paraquat were effective in controlling large C. virgata plants across the two experiments (Table 5). Values in parentheses are percentage control relative to nontreated. Abbreviations: LSD, least significant difference at the 5% level of significance (0.05).
The tank mix combination of haloxyfop and paraquat was the only treatment to achieve 100% control in both the 2015 and 2016 experiments. The tank mixing of paraquat with single herbicides was shown to be synergistic for controlling large C. virgata, plants. Tank mixes were more effective (p < 0.05) than the single herbicide treatments, with the exception of the single haloxyfop treatment, which achieved 98% and 99% control, respectively, across the two years. Generally, the tank mixes were superior to paraquat, although in the 2016 experiment it had similar control levels to the tank mix treatment treatments, excluding isoxaflutole plus paraquat and haloxyfop plus paraquat which were more effective. Previous research indicates that single knockdown herbicides such as glyphosate will not effectively control large C. virgata plants [12]. In previous work involving on-farm winter fallow experiments in central Queensland, Australia, control of C. virgata was improved using a combination of herbicides from different MOA groups rather than single herbicides alone [12]. Although this work was carried out on earlier weed growth stages, it appears as though there is a synergistic effect of mixing certain herbicides for control of C. virgata regardless of its size.

Effect of Sequential Application Herbicide Treatments on Chloris virgata
The sequential application treatments of isoxaflutole fb paraquat, atrazine fb paraquat, simazine fb paraquat, terbuthylazine fb paraquat and haloxyfop fb paraquat were very effective in controlling large C. virgata plants across the 2015 and 2016 experiments (Table 6). Values in parentheses are percentage control relative to nontreated. Abbreviations: fb = followed by 2nd application of herbicide 7 days later; LSD, least significant difference at the 5% level of significance (0.05).
Isoxaflutole fb paraquat and haloxyfop fb paraquat achieved 100% control over both years. In the 2016 experiment, atrazine fb paraquat and paraquat fb paraquat also achieved 100% control, with no weed biomass remaining at 42 d. All of the sequential application treatments had similar levels of control. The sequential application treatments were clearly superior to the single herbicide treatments, although paraquat as a stand-alone treatment in the 2016 experiment had similar efficacy to the sequential application treatments with 91% control at 42 d. Paraquat fb paraquat achieved 96% and 100% control, respectively, over the 2015 and 2016 experiments which shows the benefit of using two applications of paraquat 7 days apart over a single application. As suggested previously, although paraquat as a single stand-alone treatment has been moderately effective in controlling C. virgata plants in this study, the use of mixtures and or sequential applications is the preferred option. Previous studies under glasshouse and field conditions conducted in Toowoomba, southern Queensland, showed that sequential applications of systemic herbicides followed by paraquat were more effective than single applications [18]. Additional work from central Queensland conducted in cropping fallows showed that single applications of herbicides were inferior to sequential applications for the control of C. virgata at a range of growth stages [12].

Conclusions
Tank mixtures of Group A, C, and H herbicides mixed with the Group L herbicide paraquat were very effective in controlling large E. colona plants, showing a synergism between the systemic MOA herbicide groups and Group L. Additionally, paraquat as a stand-alone treatment had high efficacy levels, and haloxyfop was effective in one experiment and not the other, which may have been due to varying humidity levels at time of application, as mentioned previously. Overall, sequential treatments of Group A, C, and H herbicides followed by paraquat 7 d later were very effective in controlling large E. colona plants. Results from this study suggest that large E. colona plants can be well controlled with tank mixtures of Group A, C, H and L herbicides and sequential applications of Group A, C, and H followed by a Group L herbicide 7 days later, and are generally superior to single herbicide applications. This synergism between the MOA groups provides an important alternative control option to manage large E. colona plants in fallow situations. Tank mixture treatments effectively controlled large C. virgata plants and were markedly better than the single herbicide treatments with the exception of haloxyfop. Interestingly, haloxyfop as a single treatment achieved excellent control in both experiments. Haloxyfop plus paraquat was the most effective tank mix treatment with 100% control for both experiments at 42 d. The sequential treatments were superior to the single treatments for controlling large C. virgata plants; however, haloxyfop on its own was very effective with 96% and 98% control, respectively, over the 2015 and 2016 experiments. There seems to be a clear synergistic effect of combining Group A, C, and H herbicides with the Group L herbicide paraquat, as either tank mixes or sequential applications to control large E. colona and C. virgata plants. It is quite feasible that these treatments will not be as effective under field conditions, due to plants being more stressed under paddock conditions and consequently some adjustments of treatments may be required, possibly through the addition of other herbicide groups or an increase in rate of active. If, however, these treatments are effective under field conditions, then there is promise to use these herbicide combinations as an alternative to glyphosate to managing these weeds in a fallow scenario. The use of Group H and C herbicides when mixed with paraquat do provide the additional benefit of residual activity in conjunction with post-emergent control. The use of these herbicides comes with inherent risk to subsequent crop safety, and more thorough planning would be required to integrate them into crop management systems.