Key messages

• Our research has shown large differences in seed dormancy between brome and barley grass populations; high seed dormancy populations are more difficult to control with pre-sowing knockdown herbicides and delayed crop sowing.
• Brome grass seedbank tends to persist for three years and barley grass for two years. Therefore, single year management programs are unlikely to prevent rebound in populations of these weeds.
• Presence of resistance to group 1 (A) herbicides is still relatively low but, in some regions, resistant populations have been responsible for control failures.
• Integration of higher crop densities (seeding rate) with effective herbicide options has been consistently successful in minimising crop yield loss and reducing weed seed set of brome grass.

Aims

• Identify weed traits responsible for the increasing incidence of brome and barley grass in cereal crops in southern Australia.
• Quantify benefits of integrating non-chemical tactics such as sowing time and crop density with herbicides to improve weed control.

Introduction

Brome grass has climbed to be the fourth worst weed of grain crops in Australia in terms of the area infested, as well as yield and revenue loss (Llewellyn et al 2016). Barley grass has also increased in importance to become one of the top 10 weeds of Australian cropping. In this survey, barley grass was ranked as the seventh most costly weed to control by growers in South Australia and the Victorian Mallee and Mid-North, Lower Yorke and Eyre Peninsula. Given the increasing importance of these two grass weeds, it is important not only to understand why they have been increasing in grain crops but also to determine how they could be managed more effectively. Research to understand the biological factors responsible for the increasing incidence of these weeds has focused on changes in seed dormancy in response to cropping intensity and to persistence of their seedbank. Research is currently underway to determine how integration of non-chemical factors could be used to enhance weed control with herbicides.

Method

Seed dormancy

Barley grass populations from the low rainfall zones in New South Wales, Victoria, South Australia and Western Australia were collected during the summer of 2018. In this random survey, a total of 143 samples were collected from grower paddocks. Seeds of barley grass samples were removed from panicles and sown by weight (2g per tray) into seedling trays filled with potting mix in the first week of April. Seedling trays were placed outdoors at Roseworthy (SA) to experience natural rainfall and temperature conditions. There were two replicates of each barley grass population. Weed seedlings were counted and removed throughout the growing season to determine seedling emergence pattern (i.e., dormancy). Cumulative seedling emergence data were analysed in GraphPad Prism. An identical experimental approach was used to assess differences in seed dormancy between in-crop and non-crop populations of brome grass collected from growers’ fields in 2015.

Seedbank persistence

Seeds of barley grass and brome grass populations were placed in soil in micro-plots and seedling emergence was recorded regularly in subsequent seasons. Field sites for seedbank persistence were established at Karoonda (low rainfall), Roseworthy (medium rainfall) and Tarlee (high rainfall). Absence of new weed seedling emergence for the whole growing season was considered indication of complete exhaustion of the weed seedbank.

Herbicide resistance

Barley grass populations (n=143) collected from the low rainfall zones in New South Wales, Victoria, South Australia and Western Australia in the summer of 2018 were tested for resistance to all major groups used for selective and non-selective weed control. Herbicides were used at the recommended field rate. This included quizalofop and clethodim (group 1 (A)), imazamox + imazapyr (group 2 (B)), glyphosate (group 9 (M)) and paraquat (group 22 (L)). Plants that produced new growth after herbicide treatment were rated as resistant to that herbicide.

Brome grass management

Each year since 2018, three field trials have been undertaken to investigate the effect of crop sowing time, seed rate and herbicide treatments on brome grass control, weed seed set and crop yield. In this paper we have presented results of two field trials were undertaken in South Australia in 2019 to investigate brome grass management in Razor Cl Plus PBR symbol Clearfield® wheat.

Results

Seed dormancy

Based on extensive research over the last 10 years, it is clear that higher cropping intensities select for greater seed dormancy. Initial evidence for this trend emerged in brome grass collected from fence lines and adjacent cropping fields at Warner Town in South Australia. Since then, further research has confirmed similar trends in many other brome grass and barley grass populations. Essentially, all weed populations possess individuals with different levels of seed dormancy (i.e., genetic variation). Management systems (e.g., cropping) that effectively kill early germinating weeds (low dormancy) tend to increase seed dormancy in weed populations. Conversely, systems that allow all individuals to survive and set seed (e.g., pastures or fence lines) maintain lower levels of seed dormancy. Results from one of our recent studies can be used to highlight this principle. In 2019 we identified two barley grass populations with contrasting seed dormancy from upper Eyre Peninsula. Population SEP-AC3 came from a paddock with low cropping frequency whereas SEP-KV2 was from a paddock with high cropping frequency (Figure 1). These populations showed more than 3-fold difference in time required for 50% seed germination (13 days vs 46 days). These results clearly show how management practices within a region can have a large influence on seed dormancy of weed populations. From a practical viewpoint, populations with high seed dormancy will have a slow and staggered weed establishment, which will reduce effectiveness of pre-sowing knockdown herbicides and possibly of some pre-emergent herbicides as well.

As a group, populations from the southern plains of New South Wales (NSP) were the quickest to germinate and emerge due to low seed dormancy (t₅₀= 8.9 ± 1.08d). This contrasts with barley grass populations from the Upper Eyre Peninsula in South Australia (SEP), which had the highest t₅₀ (32.6 ± 3.17d). The average t₅₀ for the other regions ranged from 13.2 days for the central plains of New South Wales (NCP) to 18.5 days for the populations from the Victorian and South Australian Mallee. The average t₅₀ for Western Australian populations ranged from 13.8 to 17.7 days. Within most of the regions there were sizable differences between the least and the most dormant populations. Therefore, management practices used in the paddocks, where these samples were collected, also appears to have influenced seed dormancy status of barley grass.

Seedbank persistence

Brome grass: Field studies at three different locations in South Australia showed that the brome grass seedbank can persist for three years even though most of the seedling establishment occurs within the first year after seed dispersal. Karoonda South Australia, with sandier soils and a lower rainfall, showed greater persistence of the initial seedbank into the second and third year than sites with greater rainfall and heavier textured soils (Roseworthy and Tarlee SA). Similar pattern of seedbank persistence was observed at Wongan Hills in Western Australia. These results are consistent with previous studies which showed that a three-year effective management program can deplete field populations of this weed species.

Barley grass: Seedbank persistence of barley grass was investigated in South Australia at three field sites (Karoonda, Roseworthy and Tarlee). At the site with the highest rainfall (Tarlee), barley grass emergence only occurred in year 1, which indicates that remaining seed had decayed by the second

growing season. At Roseworthy (medium rainfall), there was only 0.2% emergence from the initial seedbank in year 2 and no emergence was observed in years 3 and 4. Karoonda, which has the lowest rainfall out of the three South Australian sites, showed much greater seedling emergence in year 2 (12%) than the other two sites. Even at Karoonda, there was no barley grass plants observed in years 3 and 4 of this study. Therefore, the barley grass seedbank in South Australia appears to be completely exhausted after two years. These results are consistent with the results from Western Australia, where most of the seedlings emerged within a year. However, some barley grass establishment was observed in the Western Australian trials even after 3-4 years. These results again highlight the difficulty in completely exhausting weed seedbanks in a single year and the need for a two-year management program incorporating rotations and herbicides.

Herbicide resistance

All populations of barley grass collected in NSW and Victoria were susceptible to the four herbicide groups used in the resistance screening (Figure 2). However, some samples from SA and WA showed resistance to group 1 (A) and 2 (B) herbicides. Resistance to the SU herbicide Atlantis® was identified in 16.1% of the populations tested. The presence of resistance to the imidazolinone herbicide Intervix® was relatively low (1.4%). Resistance to the FOP herbicide quizalofop (Leopard®) was detected in 4.2% of the barley populations tested. Four of these populations came from the upper Eyre Peninsula in SA and two were from WA. Survivors of this herbicide were vigorous and showed no inhibition in growth. There is no doubt that presence of resistance to group 1 (A) and 2 (B) herbicides in the southern and western region will complicate management of barley grass in break crops and pastures. There was no resistance detected to glyphosate or paraquat in barley grass samples in this survey. However, a subsequent survey in 2021 has identified populations with resistance to paraquat and glyphosate.

Based on resistance testing of barley grass over last three years, it can be stated that the overall level of herbicide resistance in still low but there are some regions such as Upper Eyre Peninsula where resistance levels are higher than other regions. However, herbicide resistant populations were also detected in WA and VIC. Similarly, herbicide resistance in brome grass remains at a much lower level than in annual ryegrass. However, some populations with resistance to group A (1), B (2) and M (9) have already been identified. Therefore, growers facing unexpected herbicide failures should send their seed samples to commercial laboratories for resistance testing.

Brome grass management

A two-week delay in sowing reduced brome grass density by 82% at Riverton compared to a 38% reduction at Mallala. As both sites received very similar rainfall during the month of May, the differences in effectiveness of delayed sowing in controlling brome grass are likely to be associated with seed dormancy in these two populations. At Riverton, delaying crop sowing by two weeks reduced brome grass seed set by 76% for TriflurX + Avadex Xtra and 93% for Sakura + Avadex Xtra treatments. TriflurX + Avadex Xtra fb Intervix completely prevented brome grass seed set at both times of sowing (Figure 3a). Similar effects of delayed sowing on herbicide efficacy on brome seed set were also observed at Mallala (P=0.026) (Figure 3b). Considering the current low levels of resistance in brome grass to imidazolinone herbicides (Boutsalis pers. comm.), Clearfield® systems are an attractive option for brome grass management especially in cereal crops in Australia and should be carefully integrated into management plans.

In TOS 1 at Riverton, when Intervix (POST) was applied after TriflurX +Avadex Xtra IBS (2.391t/ha), grain yield of wheat increased by 45% to 4.32t/ha (Figure 4a). The comparison of the same treatments in TOS 2 showed only a 15% increase in wheat grain yield from 3.42t/ha to 3.93t/ha. The large difference in brome grass plant density in the TriflurX +Avadex Xtra IBS treatment between TOS 1 and TOS 2 (Figure 3) is the most likely reason for these yield responses. As brome grass was almost completely controlled in TriflurX + Avadex Xtra fb Intervix (Figure 2), comparison of TOS 1 and TOS 2 for this treatment provides an indication of the yield penalty from delayed sowing. Wheat yield for this herbicide treatment was 4.32t/ha for TOS 1 compared with 3.93t/ha for TOS 2, which equates to a 9% yield penalty (Figure 4) or 130kg/ha/week.

Herbicide treatments also had a significant effect on wheat grain yield at Mallala (Figure 4b). The treatment of TriflurX + Avadex Xtra produced a wheat yield of only 1.11t/ha, which was significantly lower than the wheat yield in the herbicide mixture of Sakura + Avadex Xtra (1.81t/ha). However, when Intervix post-emergence herbicide was used, wheat yield increased further to 2.63t/ha. In this trial, integration of Clearfield® technology with pre-emergent herbicides not only prevented brome grass seed set (Figure 3), but it also produced the highest grain yields (Figure 4b).

Conclusion

Our research has shown large differences in seed dormancy between brome and barley grass populations, which can have a large effect on the performance of pre-sowing weed control and the success of delayed crop sowing for weed control. Growers are often unaware of the seed dormancy status of their weed populations but careful observation of paddocks for weed emergence after the opening rains can be helpful in ranking paddocks for seed dormancy. Other factors that can influence success of weed management include seedbank persistence and herbicide resistance status. A brome grass seedbank tends to persist for three years and barley grass for two years. Therefore, single year management programs are unlikely to prevent rebound in populations of these weeds. Integration of higher crop densities (seeding rate) with effective herbicide options has been consistently successful in minimising crop yield loss and reducing weed seed set of brome grass.

Acknowledgments

The research undertaken as part of this project is made possible by the significant contributions of growers through both trial cooperation and the support of the GRDC, the authors would like to thank them for their continued support.

References

Llewellyn RS, Ronning D, Ouzman J, Walker S, Mayfield A and Clarke M (2016) Impact of Weeds on Australian Grain Production: the cost of weeds to Australian grain growers and the adoption of weed management and tillage practices. Report for GRDC. CSIRO, Australia.

Contact details

Dr Gurjeet Gill
The University of Adelaide
Waite Campus, Glen Osmond SA 5064
Ph: 08 83137744
Email: gurjeet.gill@adelaide.edu.au