WeedSmart is the industry voice that delivers science-backed weed control solutions.

As crops mature and harvesters begin reaping, consider the potential fate of seeds ripening on weeds that escaped in-crop control measures.
Peter Newman, WeedSmart’s western extension agronomist, says harvest time is an important opportunity to assess weed burden across the farm and be proactive about driving down the weed seed bank.
“Harvest can either be a super-spreader or a weed suppressing event,” he says. “Small patches of weeds can quickly expand when seed is blown out the back of the harvester. On the other hand, the harvester can be a powerful weed management tool if any one of the harvest weed seed control options are implemented.”
WeedSmart’s western extension agronomist, Peter Newman says efforts made to reduce the spread of weed seed at harvest will soon pay off for growers.
Australian growers have led the world in inventing and adopting harvest weed seed control tools such as impact mills, chaff carts, chaff decks and chaff lining, all of which can reliably destroy over 90 per cent of the weed seed that enters the front of the harvester.    
“In addition to harvest weed seed control there are several other actions in the WeedSmart Big 6 that growers can implement just prior to, during and immediately after harvest that will make a measurable difference to the weed burden in future growing seasons,” says Peter. “The WeedSmart Big 6 tactics are scientifically-proven to reduce the risk of herbicide resistance through diverse herbicide use and cultural control to prevent weed seed set.”
What can I do before harvest to manage late emerged weeds?
In brief: Scout for and map weedy patches. Consider sacrificing small areas of high density weeds. Swathing can be a very effective way to stop seed set of late emerged or resistant weeds. Collect weed seeds for herbicide susceptibility testing.
The details: Growers across Australia use a variety of methods to map weeds – from the simple to the sublime. ‘Dropping a pin’ using the tractor’s GPS mapping system as you travel through a weedy section when spraying or harvesting is easy and provides useful information about the distribution of weeds in the paddock. Many growers have their own drones and use them the collect images or video footage of the crop that can be viewed or analysed to identify high density weed patches.
Collect seed for herbicide susceptibility testing – knowing what still works is vital information for planning next season’s herbicide program. There are three herbicide testing facilities in Australia that are equipped to test weed seed samples – Plant Science Consulting, CSU Herbicide Resistance Testing and UWA Herbicide Resistance Testing.
Collecting weed seed before or at harvest is the most common method used. The collected seed must be mature, from green to when the seed changes colour. Before harvest, collect 30 to 40 ryegrass seedheads or several handfuls of wild oats seed. After harvest, it is common to find seedheads still in the paddock or samples of contaminated grain can be sent for analysis.
Keep samples from different locations separate and details noted on the bag. Only use paper bags (double layer) to collect and send seed samples. Ensure bags are sealed so that the samples don’t mix during transit.
Which harvest weed seed control tool is best for my situation?
In brief: There are six harvest weed seed control tools used in Australia – impact mills, chaff decks, chaff lining, chaff carts, bale direct and narrow windrow burning. Choose the one that best suits your system and budget.
The details: Impact mills are best suited to continuous cropping situations. Residues are retained and evenly spread. Chaff decks have lower capital cost and are well-suited to controlled traffic situations. Chaff carts are popular with grain producers who also run livestock. Bale direct is also expensive but has a good fit in locations where there is access to straw markets. Chaff lining is currently the best ‘entry level’ system and can be used in CTF or non-CTF systems, with best results where the harvester runs on the same track each year. Chaff lining has essentially superseded narrow windrow burning, overcoming the time required and risks involved in burning and reducing the loss of nutrients from the system.
If you haven’t used harvest weed seed control tools before, it doesn’t take long to build and fit a chaff lining chute ready for use this harvest season.
What should I be ready to do straight after harvest?
In brief: Spraying weeds immediately after harvest is fairly common practice. Weeds present may be close to maturity or fresh germinations of summer-active weed species.
The details: Some growers get in early with knockdown herbicide applied under the cutter bar when swathing barley or canola. Consider using the double knock strategy, heavy grazing pressure and possibly a soil residual herbicide that is compatible with your planned crop rotation. Pay particular attention to any weedy patches identified before or during harvest. Stopping seed set at every opportunity is the crux of an effective weed management program.
Give some thought to what might be the underlying cause of weedy patches – fixing problems such as pH and soil nutrition imbalances, waterlogging and spray practices that routinely deliver low doses of herbicide.

Belinda Chase, agronomist at Dalby Rural
It has been a few years since growers on the Darling Downs have been able to plan for a summer crop. With the uncertainty surrounding planting opportunities came some hesitancy to fully invest in weed control.
Belinda Chase, an agronomist at Dalby Rural Supplies, says it is a welcome change to be actively planning for summer cropping and being more proactive with weed control leading up to planting.
“Our main weeds here over summer are feathertop Rhodes grass, barnyard grass and broadleaf species like flaxleaf fleabane and peach vine,” she says. “In terms of known resistance we have some Group 1 [A] and glyphosate resistance in barnyard grass and paraquat plus glyphosate resistance in tall fleabane. Annual ryegrass seed in hay from other states has also established in small areas, potentially bringing resistance that has evolved elsewhere.”
With more soil moisture than growers have had for several years, many have been planning ahead to manage weeds in winter fallows destined from summer crops – mainly cotton and sorghum, along with some millet, corn, mungbean and sunflower.    
“Pre-emergent, or residual, herbicides have played an important role in keeping weed numbers low in winter fallows,” says Belinda. “The next rain event will most likely also be the trigger for planting, and weed control in those early weeks of crop growth will be crucial.”
The WeedSmart Big 6 tactics can be applied to all crops to reduce the risk of herbicide resistance through diverse herbicide use and cultural control to prevent weed seed set.
What are the resistance risks with pre-emergent herbicides in summer crops?
In brief: Avoid over-using the same mode of action group throughout the crop rotation.
The details: There has been widespread use of knockdown herbicides with residual activity, such as Group 14 [G] flumioxazin (e.g. Valor), in winter fallows to maintain low weed numbers. With multiple use patterns in fallow (on its own or as a tank mix ‘spike’) and post-sowing pre-emergent in a number of summer crops, there is a risk that resistance to this mode of action could develop. There is currently no known resistance to Group 14 [G] herbicides in Australia but resistance has been identified in both grass and broadleaf species in other countries.
Valor has a plant back period of one month pre-sowing for pigeon pea, maize, sorghum and navy bean, two months pre-sowing for cotton, sunflower and mungbean and up to nine months pre-sowing for canola. It can be applied after sowing and pre-emergent to protect the crop as it establishes.
Tank mixes and co-formulations are an effective way to reduce the risk of resistance to Group 14 [G] herbicides. Significant escapes must be prevented from setting seed.
Another option for pre-emergent weed control in sorghum is s-metolachlor, a Group 15 [K] herbicide such as Dual Gold. Applied as a split application before and immediately after planting sorghum gives optimal weed control, provided the seed safener (Concep II or Epivio C) has been applied at planting.
Atrazine, a Group 5 [C] herbicide, also provides a pre- and post-emergence option for broadleaf weeds and some pre-emergent control of annual grasses like barnyard, crowsfoot, summer and love grass in crops such as sorghum and maize.
How can I best incorporate imi-tolerant summer crops?
In brief: Imi-tolerant (IT) hybrids are now available for sorghum, maize, sunflower and soybean, providing for alternative in-crop control options for a range of grass and broadleaf weeds.
The details: IT sorghum seed is in short supply this year, but there are other options to consider. Growers who have Group 1 [A] resistance in barnyard grass or who are looking for alternative products to control or suppress other key summer weeds may consider growing these herbicide-tolerant crops.
With an imi-tolerant hybrid, growers will be able to safely plant into paddocks where imidazolinone (Group 2 [B]) chemistry has been recently applied, say after an IT maize, wheat, canola or pulse crop, negating the current 12+ month plant-back requirement, which otherwise restricts the choices that growers have in summer. In this situation it is advisable to avoid applying Group 2 [B] herbicides to the current crop to break the ‘imi-cycle’ that can develop where imi herbicides are used in successive crops – resulting in a high risk scenario for the evolution of herbicide resistance.
The second major benefit is that an imi-tolerant hybrid can be planted into a paddock that has a broadleaf weed burden—a situation that would not be advisable for a non-IT variety.
Imi-tolerant (IT) hybrids are now available for sorghum, maize, sunflower and soybean.
Is patch management a good tactic for feathertop Rhodes grass?
In brief: Feathertop Rhodes grass is a serious challenge, but summer cropping can help reduce the weed seed bank.
The details: Controlling FTR can be easier in a competitive mungbean crop on narrow row spacing, than in sorghum and sunflower, which are typically grown on wider row spacings and do not quickly cover the inter-row.
Controlling feathertop Rhodes grass can be easier in a competitive crop on narrow row spacing, where the crop quickly covers the inter-row.
After applying suitable pre-emergent herbicides (if FTR numbers are low) or cultivating if FTR pressure is high, do everything you can to maximise crop competition. Consider narrower row spacing, optimal plant density, avoid planting gaps and provide adequate crop nutrition.
After the crop has established, use inter-row cultivation where necessary and harvest early. If there are areas with low density escapes of this prolific seeding weed, chipping is worthwhile before the seed sets. If patches of escapes are evident it can be very beneficial to accept a ‘short-term pain for long-term gain’ by ploughing or spraying out the affected patch of crop before the FTR plants set seed.
At harvest, record any patches of FTR and return soon after harvest to remove established plants and apply a pre-emergent herbicide to manage new germinations.
Further information

Feathertop Rhodes grass and sorghum – does this work?
Avoid getting stuck in the imi-cycle
Competitive sorghum to reduce FTR and BYG pressure

Chaff carts were invented in Canada in the 1970s as a fodder collection system, so they have always had an association with livestock. In the late 1980s they were introduced to Australia and used as one of the first harvest weed seed control (HWSC) tools to manage herbicide resistant ryegrass and wild radish in Western Australia.
Depending on the grower’s situation, chaff heaps in Australia have been burned, grazed then burned, sold as loose chaff for horse feed, grazed or knocked down then seeded through (unburned) and most recently, baled and used on-farm as conserved fodder.
Different model chaff carts incorporate more or less straw with the chaff, which can affect the feed value of the bales.
This latest method fits very well with the move to confinement feeding of sheep and brings the use of the chaff cart almost full circle. Given the long-standing association with fodder conservation, the value of the chaff as a feed source is well known, particularly for sheep.
A HWSC Calculator developed by WeedSmart western extension agronomist Peter Newman, now enables growers to quickly estimate the potential value of baling chaff from chaff heaps and feeding it out in on-farm confinement areas. Upper Eyre Peninsula mixed farmer, Bruce Heddle, instigated the addition of the new feature to the calculator and provided the energy figures from his own testing of different chaff types. Planfarm livestock consultant, Paul Omodei, worked with Peter to make sure the calculator took into account the important considerations for livestock production.
Although there is variability in feed value results, ballpark energy figures are in the order of 6.7 MJ/kg DM for wheat chaff, 7.7 MJ/kg DM for barley chaff, and 8 MJ/kg DM for canola and pulse chaff. Peter says the cost to deliver baled barley chaff to an on-farm feedlot is around 1.5 c/MJ compared to barley grain for 2.3 c/MJ, based on these feed values that Bruce Heddle provided.
“Chaff carts incorporate more or less straw with the chaff depending on the cart model, and while some extra straw makes baling and handling easier, it also reduces the feed value slightly, so that needs to be factored in,” he says. “Some growers are already set-up for hay making while others might choose to employ a contractor.”
Bruce has a contractor to bale their wheat, barley, canola and lentil chaff dumps, and feeds the bales along with barley grain and a lick feeder to sheep held in confinement from January to the end of seeding. All the chaff is beneficial although the canola and lentil chaff has the highest feed value.
Bruce says having the sheep confined during this period reduces their workload as the sheep are easy to look after and feeding out is only needed once a week. They are able to easily monitor the flock and attend to any problems quickly, and don’t have to spend time moving sheep around during seeding.
The pastures also benefit from a spell from grazing for these months, so the sheep are turned out into refreshed pasture after their confinement.
Jarred Tilley is another mixed farmer who has been making chaff bales for use in their family’s sheep enterprise at Kapunda in South Australia for the last two years. Jarred has been fencing off less productive areas on their farms and using them as confinement yards for their sheep in May and June. Their high density baler creates large bales that weigh 500 to 550 kg each from chaff dumps that are knocked down using a telehandler and then raked to save time when baling. Having their own hay-making equipment helps justify the chaff baling operation.
“We have fed 500 ewes in an 8 hectare paddock for six weeks with a diet of chaff bales, regular hay and a lick feeder,” he says. “The benefits are probably marginal for us using wheat chaff, but the canola and pulse chaff is a better feed supplement. The feed quality of the chaff is not always as good as we would like.”
Jarred says that the sheep spread the chaff out in the confinement paddocks and often leave more ground cover in those paddocks than when the sheep went in.
Chaff is a low-cost feed that is only sufficient for maintenance energy levels for sheep. Other feeds need to be included in the feed ration to promote growth or support lactation. Studies that Ed Riggall at AgPro Management conducted at multiple sites in Western Australia over three years demonstrated that sheep with access to chaff heaps from various crops gained an average of 2 kg in the first three weeks. This was 500 g more than sheep grazing stubbles where the chaff was spread out by the harvester.
Livestock containment paddocks boost productivity while stopping the spread of herbicide resistant weed seeds.
At the end of six weeks grazing the sheep with access to chaff piles had gained about 100 g, while sheep without access to the chaff piles had lost almost 2 kg compared to their starting weight. Additional benefits might be expected in a confinement feeding situation where the chaff is easier to forage and is potentially available to the stock for longer as part of a mixed ration.
By making use of the chaff within their own operations these growers are gaining benefit from a resource that is otherwise wasted, and avoid the risks associated with burning chaff heaps.
All harvest weed seed control methods provide similar levels of weed control – collecting, concentrating or destroying over 90 per cent of the weed seed that enters the front of a well set-up harvester. Australian farmers are spoiled for choice when it comes to the options for collecting and destroying weed seed at harvest and keeping downward pressure on herbicide resistance. The WeedSmart HWSC Calculator tool allows growers and agronomists to test different scenarios using their own figures to estimate the costs and benefits of the different systems available.
More information

Download the HSWC costs calculator
More lambs, less weeds in confinement systems
Sheep can turn weed seeds and chaff into cash
Ben & Emily Webb case study

These weeds are now the subject of a pilot ‘area-wide management’ project to trial cooperative and cost-effective methods to reduce the movement of these weeds and the herbicide resistance traits they have evolved.
The cross-industry project has Australian government funding to target weeds that are a common problem to all industries in an area and have ‘mobile’ seed and pollen – that is, they spread easily. Weed species that fit the criterion include flaxleaf fleabane, feathertop Rhodes grass and annual ryegrass.
In three distinct regions – in the Darling Downs region of Queensland, the Riverina region of NSW and the Sunraysia region of Victoria – project teams are devising and implementing area-wide management programs to tackle target weeds of concern in their region.
The University of Adelaide is providing targeted herbicide resistance testing within the pilot areas and mapping the spread of weeds, based on genetic testing conducted at the University of Queensland.
Dr Rick Llewellyn, senior principal research scientist (agricultural systems) with CSIRO is leading the ‘Area wide management for cropping systems weeds’ project to better understand the importance of weed mobility, and test the opportunities for this collaborative approach. He says the idea is to draw together industries and land managers to ‘find a collaborative solution to a common problem’ where a strong value proposition can be established.
Dr Rick Llewellyn, CSIRO, says more coordinated awareness and information sharing can channel effort and innovation into weed management improvements that benefit both the individual land manager as well as the district.
“Area-wide management has been very effective in the management of invasive animal pests and for some mobile insect pests,” says Rick. “We know that some weeds are particularly good at moving across the landscape, either as contaminants or borne on the wind or in flood water – and most farmers have experienced a weed incursion from a source beyond their farm boundary. We are testing collaborative and cost-effective ways to reduce the spread of cropping weeds across diverse farming landscapes.”
Each of the three pilot areas have identified the highest priority mobile weeds to target in their initial on-ground project. In the Sunraysia region the Mallee Sustainable Farming and horticulture organisations have partnered to develop strategies that minimise spray drift while also providing effective control of important weeds like fleabane. Dr Chris Preston, The University of Adelaide, is assisting the Sunraysia project team as they investigate application techniques and product choice for summer weed control in this diverse cropping region.
“Where a range of different crops are grown in close proximity there is a risk of damage through off-site movement of herbicides. To reduce this risk, growers using some products, such as phenoxy herbicides, must work within narrow application windows; but to prevent large populations of weeds setting mobile seeds, growers need cost-effective herbicide options,” says Rick. “The area-wide management trials led by Mallee Sustainable Farming compared weed fallow control efficacy of six alternative products registered for use in optical sprayers, as well as options for better control of mobile and resistance-prone weeds like sow thistle in horticulture.”
The area-wide management trials led by Mallee Sustainable Farming compared weed control efficacy of six alternative products registered for use in optical sprayers, as well as options for better control of mobile and resistance-prone weeds like sow thistle in horticulture. Photo: Frontier Farming Systems.
In the Riverina’s Murrumbidgee Irrigation Area, the focus of the area-wide trials led by the local Irrigation Research and Extension Committee (IREC) is to reduce the movement of weed seed within the irrigation scheme through better management of channel bank vegetation.
Establishment of weed-suppressive grassed channel banks is being tested to assess its feasibility as a weed management strategy that could bring benefits to the many industries that utilise the irrigation scheme. The area-wide activities in the Riverina also attracted extra support from the wine industry with additional trials established to prevent seed set of mobile weeds in vineyards.
The Darling Downs pilot is addressing the management of drainage line and roadside vegetation, particularly where farmers are aiming to maintain weed-free fallows. The project also included the demonstration of new innovations, such as the mechanical weed chipper for control of glyphosate resistant weeds in fallows. The focus weeds in this project are fleabane and feathertop Rhodes grass, both of which can be difficult to control with knockdown herbicides such as glyphosate.
The Darling Downs area-wide weed management project is focused on the management of road-side vegetation, particularly alongside cropping paddocks.
“In each of the pilot areas we have conducted herbicide resistance testing to build a picture of the extent and pattern of resistance across the landscape, on public lands, orchards, vineyards, and on grain and cotton farms,” says Rick. “We have also used genetic testing to map related populations to determine where weeds have come from. For example, we are interested in whether the weeds in the cropping paddocks are related to those found on adjacent roadsides, or did they originate from a distant site.”
Initial findings suggest that there is not usually evidence of ‘resistance fronts’ moving across a district, but rather resistant weed outbreaks are usually scattered. This points to the importance of localised ‘neighbourly’ action to reduce the overall cost of weeds, in addition to broader regional approaches to prevent the introduction and spread of problem weeds.
The WeedSmart Big 6 tactics can be applied to area-wide management as well as within a cropping enterprise to tackle resistance through strategic patch management and diverse control methods that result in low weed densities and prevent seed set of mobile weeds.
Rick says more coordinated awareness and information sharing can channel effort and innovation into weed management improvements that benefit the individual land manager as well as the district.
“There has been an increase in the diversity of food production industries in many districts over recent decades, so there’s more and more opportunity for a collaborative approach to reduce weed costs and risks as ‘new neighbours’ become established in many dryland grain growing areas,” he says.
Research and development partners involved with the project include Grains Research and Development Corporation, Cotton Research and Development Corporation, AgriFutures Australia, CSIRO, University of Queensland, University of Adelaide, University of Wollongong, Mallee Sustainable Farming, Millmerran Landcare Group, Irrigation Research & Extension Committee Inc, together with Wine Australia, the Toowoomba Regional Council and a range of additional local industry organisations.
This project is supported through funding from the Australian Government Department of Agriculture as part of its Rural R&D for Profit program and the Grains Research and Development Corporation and the Cotton Research and Development Corporation.
More resources

Area wide management of weeds project updates
Social attitudes to area-wide management – Preliminary report, Darling Downs

Although an emerging weed in the northern grain growing region, wild turnip steals crop yield and requires additional control measures that together cost over $10 million every year across Australia.
Bhagirath Chauhan, Professor in Weed Science at The University of Queensland, Gatton, says recent ecological studies have shed light on tactics that growers can use to rein in this invasive weed that could become a threat, particularly in production areas with marginal soil moisture in the northern region.
Bhagirath Chauhan, Professor in Weed Science, The University of Queensland, Gatton says if wild turnip plants are prevented from setting seed, it is possible to rapidly deplete the seedbank in a no-till system using a 6 to 12-month fallow and or competitive cropping.
“Wild turnip is considered a winter weed, but with sufficient soil moisture and mild temperatures it can also establish and set seed over summer,” he says. “Some biotypes of Brassica tournefortii have evolved resistance to chlorsulfuron, and other Group 2 [B] herbicides, first identified in South Australia in 1996.”
With investment from the GRDC, Dr Gulshan Mahajan conducted the recent studies on four biotypes collected in the northern region investigated the differences in seed dormancy, drought tolerance, effect of competition by wheat and chickpea crops, and seed persistence on the surface and at various burial depths.    
“These experiments clearly demonstrated the invasive capability of wild turnip,” says Dr Chauhan. “This species can produce vast quantities of seed with variable dormancy, meaning there can be multiple germination cohorts, mainly associated with rainfall events. However, it is a poor competitor when faced with a crop such as wheat that achieves canopy closure quickly.”
‘Grow competitive crops’ is one of the WeedSmart Big 6 tactics, providing season-long weed suppression and maximising the value of early weed control efforts.
How long does wild turnip seed persist in the seedbank?
In brief: Seed persists for up to 18 months on the soil surface and 5 per cent of seed was still viable after being buried at a depth of 2 cm for 30 months.
The details: Fresh seeds initially have high dormancy when placed on the soil surface. The seed coat extends dormancy of fresh wild turnip seed and light inhibits germination.
Once the seed coat has degraded somewhat, seedlings readily emerge from the surface after rainfall events, generating multiple cohorts between February and October. Emergence peaks from March to May, potentially challenging crops sown from the end of April to June.
Keeping the weed seed on the surface in a no-till system and minimising soil disturbance at planting, coupled with pre-planting knockdown, pre-emergent herbicide and vigorous early crop growth can reduce germination and weed seed production in-crop.
Germination rates of up to 14 per cent occurred when seed was buried at a depth of 2 cm in soil with sufficient moisture (>25 per cent off water holding capacity) and alternating day/night temperature of 25/15 °C.
Wild turnip seedlings did not emerge from a depth of 5 cm. A one-off deep tillage event could be an effective control tactic to bury the existing seedbank, provided there was no seedbank replenishment or subsequent tillage.
A single, large wild turnip plants can produce 10,000 seeds (left). Wild turnip can become a problematic weed in no-till systems because emergence of seeds in the surface layer is greater than for buried seeds (right).
How much seed does a wild turnip plant produce?
In brief: A wild turnip plant growing in a fallow or fenceline situation can produce around 10,000 seeds.
The details: Early emerged cohorts achieve greater plant height and shoot biomass, resulting in greater seed production than later emerged cohorts. The early emerged plants also enjoyed a longer growing season, reaching flowering stage after 87 days while later emerged plants reached flowering after only 70 days.
Although water stress (25% WHC) reduced the seed production to 3000 seeds per plant, this is still ample seed to establish an infestation capable of reducing crop yield.
Wide-row and slow growing crops such as chickpea do not inhibit wild turnip growth or seed production. On the other hand, a fast growing, dense wheat crop suppressed weed growth and seed production by 78 per cent for the early sown crop (15 May), 96 per cent for the crop sown on 5 June, and 65 per cent for the late sown crop (25 June). This reduction in seed production was achieved without the application of herbicide. The vigorous growth of the wheat crops sown on the latter two planting dates prevented wild turnip plants from producing enough seeds for re-infestation.

What is the best strategy to drive down wild turnip numbers?
In brief: Pre-plant knockdown, delayed sowing, pre-emergent herbicide and a fast growing, competitive crop.
The details: Wild turnip can, and has, become a problematic weed in no-till systems because emergence of seeds in the surface layer is greater than for buried seeds. The retention of stubble supports higher soil moisture at the soil surface, creating a favourable environment for germination over a long period. However, if emerged plants are prevented from setting seed, it is possible to rapidly deplete the seedbank in a no-till system during a 6 to 12-month fallow.
Shallow tillage may result in the buried seeds remaining viable for more than 2.5 years, with the potential for seed to be brought to the surface during subsequent planting operations, triggering the re-infestation of the paddock.
Avoid slower growing and wide-spaced crops such as chickpea in paddocks with a large wild turnip seedbank.

Just as Australia led the way with the development and adoption of ‘green-on-brown’ weed detection and spot spraying in fallow situations, now Australian researchers are developing technologies that will deliver ‘green-on-green’ weed recognition and targeted control in-crop.
Imagine a machine that can identify one weed species from another and apply the best treatment to each weed, even in-crop. While expert human brains can make these differentiations and decisions relatively easily, training artificial intelligence technologies to do the same thing is challenging.
With investment from GRDC, a team of researchers led by Dr Michael Walsh, director weed research at the University of Sydney, have recently completed the pilot phase of crucial work that will underpin future developments for machine learning in weed recognition.
Dr Michael Walsh, director weed research at the University of Sydney, says the WeedAI image database will underpin future developments for machine learning in weed recognition.
Dr Walsh says there are several commercial interests developing machine learning technologies for site-specific weed control in Australia, but they all need access to a collection of relevant images to essentially ‘train’ computers in the development of weed recognition algorithms that can differentiate between crop and weed plants.
“We have set out to establish protocols for collecting and annotating images that will be stored in an open-source database that anyone with commercial or academic interests can contribute to and also use for future developments in this technology,” he says. “The pilot project has centred on collecting images and developing weed recognition algorithms to detect representative grass and weed species in wheat and chickpea crops.”
The WeedAI database currently contains thousands of images of annual ryegrass and turnip weed growing in chickpea and wheat crops. These images have been manually annotated and used to develop and test weed recognition algorithms for their accuracy in correctly identifying weeds growing in-crop.
“The images are all high quality, with annotation outlining the weed shown in the image and notes about the agricultural context, such as soil colour, location, crop type, and growth stages of the crop and or weed,” he says. “We are hoping to fast-track developments and take advantage of the machine learning technologies that have capability to accurately recognise and locate in-crop weeds to ultimately provide growers with the opportunity to specifically target these weeds with a range of weed control options.”
Machine learning offers the potential for high-level accuracy in weed recognition in-crop.
“We are hopeful that this will give growers access to a range of novel chemical and non-chemical weed control technologies that will add to the existing options available for in-crop weed control. This might include herbicides that are currently too expensive for blanket spray application.”
Dr Walsh says Australia is leading the way in developing weed recognition technologies for grain production systems and he believes the open-source database will reduce replication of effort and encourage technology companies to address more challenging scenarios, such as recognition of grass weeds in cereal crops.
Like the optical spray technology that brought tractor-mounted spot spraying to fallow management over 20 years ago, the green-on-green in-crop weed recognition systems in-crop will be used for site-specific weed control in situations where weed density is already quite low.
“At densities of less than one weed per 10 square metres, the area sprayed with herbicides would be 70 to 80 per cent less than when a blanket spray is applied,” says Dr Walsh. “The opportunities to introduce different herbicide modes of action or alternate methods of weed control such as targeted tillage or laser treatment can also be considered to reduce the risk of herbicide resistance.”
Ground speed is the enemy of real-time weed recognition systems, as accuracy increases considerably with speeds slower than those currently used for blanket spraying. With increasing computing processing speeds the expectation is that in-crop weed recognition systems will be accurate at 10 to 15 km/h. The introduction of autonomous platforms is reducing the need for higher speeds, and with a light source there will be the opportunity for round-the-clock operation of weed recognition equipped site-specific weed control systems.
A number of commercial companies are bringing in-crop spot spraying to market and will be on-hand at WeedSmart Week, Esperance to showcase their technology in mid-August. Ben White, Kondinin Group’s research manager will host the machinery session with spray and harvesting gear on display including Goldacres’ G6 Crop Cruiser series 2, weed detection technologies using drones, weed identifying cameras (green on green) and a range of harvest weed seed control options including impact mills from Seed Terminator, Redekop and iHSD (both hydraulic and belt-driven) and the Emar chaff deck. This flagship event always attracts growers keen to see how other farmers are keeping weed numbers low in different systems. Early bird registration is now open.
More resources

WeedSmart podcast – Farmers can now help improve green-on-green technology
Browse the Weed-AI image database
Video of Weed-AI workshop presentations (Day 1)
Video of Weed-AI workshop presentations (Day 2)

The single cause of herbicide resistance in weeds is selection pressure through herbicide use.
Annual ryegrass leads the charge, with resistance to multiple herbicide modes of action, and demands a readjustment in weed control strategies.
Dr Peter Boutsalis of Plant Science Consulting said that the introduction of several new herbicides over recent years has provided options for controlling some resistant populations, particularly for Group 1 [A] and Group 2 [B] resistant ryegrass, but this alone will not halt resistance evolution in ryegrass populations across Australia.
“Simply changing to another mode of action when older chemistry seems less effective is not a long-term solution. Any herbicide has the ability to select for resistance, especially in a genetically diverse species such as ryegrass,” he said. “The strategy needs to centre on increasing diversity in herbicides and non-herbicide tools, not just switching from an ‘old’ herbicide to a ‘new’ one.”
In 2020, Dr Peter Boutsalis, Plant Science Consulting, was sent 83 ryegrass samples from concerned growers in NSW and the Quick Test showed 79% of individual plants that survived paddock treatments were in fact resistant to glyphosate.
The Grains Research and Development Corporation has invested in random weed surveys in different regions within New South Wales each year from 2015 to 2019. These surveys have identified differences in the pattern of resistance between regions and other states but the trend toward multiple resistance mechanisms and resistance to increasing application rates is undeniable.
Dr John Broster, Charles Sturt University said the majority of annual ryegrass populations in NSW are resistant to Group 1 [A] ‘fop’ and Group 2 [B] herbicides with some variability between the surveyed sub-regions.
The random surveys conducted in NSW from 2015 to 2019 involved the collection and testing of 608 ryegrass populations by researchers from Charles Sturt University.
To date, no populations have been found that are resistant to the newer pre-emergent herbicides, however resistance has been reported in other states.
“Of particular concern is the percentage of ryegrass populations sampled in the random survey in some sub-regions that are resistant to glyphosate,” he said. “The extent of resistance in some areas was brought home strongly in the 2020 season when many growers were confronted with significant patches of ryegrass that clearly escaped pre-seeding glyphosate applications.”
The random surveys conducted in NSW from 2015 to 2019 involved the collection and testing of 608 ryegrass populations, with the results showing 5% of these populations were resistant to glyphosate. The highest level of resistance so far was found in the 2019 results from the eastern NSW region alone, where 14% of populations were resistant to glyphosate. A population is considered resistant to a herbicide when more than 20% of the plants grown from seed collected at a single site survive applications of registered rates of the herbicide in question.
In addition to the random sampling to provide the ‘big picture’ of resistance extent, Dr Boutsalis also conducts Quick Tests when growers and agronomists experience an apparent herbicide failure. In 2020, he was sent 83 ryegrass samples from concerned growers in NSW and the Quick Test showed 79% of individual plants that survived paddock treatments were in fact resistant to glyphosate.
“This suggested that although glyphosate resistance is generally a significant contributing factor to weeds ‘escaping’ herbicide treatment in the paddock, there are potentially other forces involved as well,” said Dr Boutsalis. “Poor application technique or application onto stressed plants, incorrect timing, sampling plants that were not exposed to glyphosate, antagonistic tank mixes, inferior glyphosate formulation, poor water quality, incorrect adjuvants, or a combination of these can also result in poor weed control in the field.”
“To keep any herbicide as a long-term option it is essential that high quality products are applied correctly and that survivor plants are prevented from setting seed,” he said. “Switching products is a very short term and inadequate solution. A better strategy is to implement a diverse program of both herbicide and non-herbicide tactics and be diligent about keeping weed numbers low.”
Other than confirming resistance, herbicide testing is a powerful way to identify modes of action that a resistant population is still susceptible to. Growers who are confronted with patches of ‘survivor’ weeds this season can send live plant samples in for the Quick Test to identify herbicide options that could be used to prevent seed set in the current season. If the escapes are not seen until seed has set, seed can be collected and sent to either CSU or Plant Science Consulting for testing against a wider range of herbicides, including pre-emergent herbicides.
Testing of ‘suspect’ seed samples sent to CSU last year resulted in 30% of populations testing positive to glyphosate resistance.
Back row = glyphosate resistant biotype, Front row = susceptibleLeft to right is 1.5 L/ha, 3 L/ha, 4 L/ha Glyphosate 540.
Patch management strategies such as cutting for hay, spraying out with paraquat, or chipping can be very effective in containing a potential blow-out. The WeedSmart Big 6 strategies for integrated weed management can then be implemented to apply long-term downward pressure on weed numbers.
The WeedSmart Big 6 tactics will be the centre of discussion at WeedSmart Week in Esperance, WA in August this year. This flagship event always attracts growers from interstate keen to see how other farmers are keeping weed numbers low in different systems. Early bird registration is now open.
Resources

Causes of poor ryegrass results and paraquat and glyphosate resistance 2020 season
Resistance and susceptibility testing

Ranked as the third most costly weed in Australian grain cropping, three weedy Avena spp. – wild oat, sterile oat and slender oat – are estimated to infest over two million hectares, causing crop yield losses of 114,596 t and a national revenue loss of $28.1 million.
In the southern and western regions, the main species found is wild oats (A. fatua), while in the northern region, sterile oat (A. sterilis ssp. ludoviciana) is the more problematic species. Both have evolved resistance to multiple herbicide groups in Australia.

QAAFI weed researchers Gulshan Mahajan and Bhagirath Chauhan have recently published a series of papers on their weed ecology studies of Avena spp., providing growers and agronomists with more information to use when formulating integrated management plans for these weeds in crops.
Practical tips

Both wild oat and sterile oat can survive in soil moisture conditions of 60 per cent water holding capacity (WHC). Sterile oat even produced seed at 40 per cent WHC.
Seedlings of these weeds can emerge from a depth of 10 cm, but greater emergence occurred from 2 and 5 cm depths. Emergence commenced at the start of winter (May) and continued until spring (October).
Early emergence plants produce the most seed, but later emergence plants can still produce enough seed to support reinfestation.
In a no-till system there is low persistence of seed on the soil surface. A 2-year assault on the weed seed bank can result in complete control of infestations.
Weed density of 15 wild oat and 16 sterile oat plants/m2 resulted in a 50 per cent reduction in wheat yield. Lower weed density (just 3 plants/m2) can still support reinfestation.
Sterile oat is a better candidate than wild oat for harvest weed seed control (HWSC).
Wild oat is best managed through early weed control (pre and post sowing) and strong crop competition.
An integrated approach to weed management can reduce Avena weed biomass by up to 90 per cent.

Experimental design features
We are summarising the finding from four related research papers:

Biological traits of six sterile oat biotypes in response to planting time. https://doi.org/10.1002/agj2.20507
Influence of soil moisture levels on the growth and reproductive behaviour of Avena fatua and Avena ludoviciana. https://doi.org/10.1371/journal.pone.0234648
Seed longevity and seedling emergence behaviour of wild oat (Avena fatua) and sterile oat (Avena sterilis ludoviciana) in response to burial depth in eastern Australia. https://doi.org/10.1017/wsc.2021.7
Interference of wild oats (Avena fatua) and sterile oats (Avena sterilis ludoviciana) in wheat. https://doi.org/10.1017/wsc.2021.25

Detailed findings
Sterile oats growth and seed production for early and late emergence cohorts
Six biotypes of sterile oats were collected from sites in southern Qld and northern NSW and planted in field conditions at the Gatton research farm in the winter cropping seasons of 2018 and 2019. The weed seed was sown early, mid and late season and the growth and reproductive potential of the six biotypes was monitored.
Averaged across the biotypes, the early planted weeds produced 2660 seeds/plant. Weeds sow mid-season produced 21 per cent less seed and the late-season weeds produced 84 per cent less seed than the early-season plants.
Although seed production was more prolific from the early and mid season plants, the late season plants produced sufficient seed to support reinfestation the following season.
A clean seed bed and competitive crop environment is the best strategy to suppress sterile oat seed production.

Effect of moisture stress on biomass and seed production of wild oats and sterile oats
Seeds of wild oat and sterile oat used in this study were collected from Warialda, NSW, in October 2017 and multiplied at the University of Queensland, Gatton Research Farm in the winter season of 2018. The pot trial to investigate the effect of 20, 40, 60, 80 and 100 per cent water holding capacity (WHC) on these two Avena weed species was conducted in 2019.
Results revealed that wild oat did not survive, and failed to produce seeds, at 20 and 40 per cent WHC. However, sterile oat survived at 40 per cent WHC and produced 54 seeds/plant, suggesting that this species is likely to compete strongly with crops in water stressed situations.
In favourable moisture conditions, both species will produce copious quantities of seed, suggesting that high infestation rates for both species may be a risk in irrigated crops.

Effect of seed burial on emergence, growth and persistence of wild oats and sterile oats
The seed longevity and emergence pattern of wild oat and sterile oat were monitored in field conditions at Gatton, Narrabri and St. George. Fresh weed seed was placed into nylon bags and buried at depths of 0, 2 and 10 cm in November 2017. Bags were exhumed at 6-month intervals over 30-months to evaluate seed germination, viability and decay.
For both species, 50 per cent of seeds at the surface and 10 cm depth had decayed within the first six months. Shallow burial (2 cm depth) of the seed increased persistence, with a significant percentage of seed being viable in the following winter cropping season.
The largest cohort of both species began to emerge at the start of the winter season (May). To ensure the seed bed is clean prior to planting, consider using tillage, herbicide application and cover crops to control this early cohort of Avena weeds. Tillage will bury seeds below their maximum depth of emergence and subsequent tillage should not be performed for 3–4 years to avoid bringing seeds back to the ‘emergence’ depth. Later emerging cohorts (through to October) will be suppressed using strong crop competition or a winter fallow if the infestation is severe.
The results of this research suggest that management strategies that can control all emerged seedlings over two years and restrict seed rain in the field could lead to complete control of weedy Avena spp. in the field.

Effect of wild oats and sterile oats infestation on wheat yield
The interference of wild oat and sterile oat in a wheat crop was examined through field studies in 2019 and 2020 at Gatton, Qld. Infestation levels of 0, 3, 6, 12, 24 and 48 plants m2 of both weed species were evaluated for their impact on wheat yield.
At an infestation level of 15 and 16 plants per m2 for wild oats and sterile oats respectively, wheat yield was halved as a result of reduced spike number per m2.
At the highest weed infestation level (48 plants per m2), wild oat and sterile oat produced a maximum of 4800 and 3970 seeds per m2, respectively. At wheat harvest, wild oat exhibited lower seed retention (17 to 39 per cent) than sterile oat (64 to 80 per cent), with most of the wild oat seeds having fallen from the seed heads before crop maturity.
The results of this study suggest that harvest weed seed control is likely to be a useful tactic in paddocks infested with sterile oat. An integrated weed management strategy that uses both chemical and nonchemical tactics is required to avoid severe crop yield loss, increased weed seed production and weed seedbank replenishment when these weed species are present.
This body of research highlights the benefits of an integrated weed management program that takes the ecology of the target weed into account.

This research was conducted by researchers from the University of Queensland, a WeedSmart scientific partner, with investment from the Grains Research and Development Corporation a WeedSmart sponsor.
Research papers

Mahajan, G., & Chauhan, B. (2021). Biological traits of six sterile oat biotypes in response to planting time. Agronomy Journal,113: 42-51 https://doi.org/10.1002/agj2.20507
Sahil , Mahajan G, Loura D, Raymont K, Chauhan BS (2020). Influence of soil moisture levels on the growth and reproductive behaviour of Avena fatua and Avena ludoviciana. PLoS ONE 15 (7): e0234648. https://doi.org/10.1371/journal.pone.0234648
Mahajan, G., & Chauhan, B. (2021). Seed longevity and seedling emergence behavior of wild oat (Avena fatua) and sterile oat (Avena sterilis ludoviciana) in response to burial depth in eastern Australia. Weed Science, 1-10. https://doi.org/10.1017/wsc.2021.7
Mahajan, G., & Chauhan, B. (2021). Interference of Wild Oats (Avena fatua) and Sterile Oats [Avena sterilis ssp. ludoviciana (Durieu)] in Wheat. Weed Science, 1-20.  https://doi.org/10.1017/wsc.2021.25