Resistance to herbicides in the broadleaf weeds Raphanus raphanistrum, Sisymbrium orientale, and Sonchus oleraceus as determined by random weed surveys in south-eastern Australia

Introduction

Weeds are a major constraint to grain production around the world. In Australia, weeds and their control cost the grains industry AU$3.3 billion per annum (Llewellyn et al. 2016). Herbicides are the main weed management tool used in grain production systems in Australia. A consequence of the intensive use of herbicides is the evolution of resistance to herbicides. Management of herbicide resistant weeds in Australia is estimated to cost grain growers AU$187 million per annum (Llewellyn et al. 2016).

While grass weeds remain the most important weeds of grain production systems in Australia, there are a number of important broadleaf weeds of concern. Many of these have also evolved resistance to herbicides (Walsh et al. 2001; Lu et al. 2007; Owen et al. 2015a; Long et al. 2019). Raphanus raphanistrum (wild radish), Sisymbrium orientale (Indian hedge mustard), and Sonchus oleraceus (sowthistle) are the 2nd, 9th, and 11th most costly weeds in Australian grain cropping in terms of revenue loss (Llewellyn et al. 2016).

R. raphanistrum, S. orientale, and S. oleraceus are all weeds of winter crops in south-eastern Australia, emerging in autumn or early winter with the crop. S. oleraceus can also emerge in spring or early summer and can also be a problem weed of summer fallows (Widderick et al. 2010). R. raphanistrum is a highly competitive weed that can reduce wheat (Triticum aestivum) yield by up to 40% (Eslami et al. 2006). It produces seeds enclosed in pod segments that can break off and contaminate grain (Cheam 1986). R. raphanistrum has a long-lived seed bank and can emerge from deep in the soil profile (Reeves et al. 1981; Chauhan et al. 2006a). S. orientale is a small-seeded weed species that germinates close to the soil surface (Boutsalis and Powles 1998; Chauhan et al. 2006b). It produces a large number of seeds that are held in siliques, which break open after maturity (Boutsalis and Powles 1998). S. oleraceus is also a small-seeded weed species, germinating close to the soil surface (Chauhan et al. 2006c; Widderick et al. 2010). S. oleraceus produces a large number of seeds each bearing a pappus that allows them to disperse long distances (Sheldon and Burrows 1973). In addition to being an important weed of summer fallows, S. oleraceus can also cause problems by its presence at harvest resulting in discolouration of harvested grain (Widderick et al. 2010).

Resistance to the ALS-inhibiting (aetolactate synthase inhibiting) herbicides was first reported in Australia in 1990 for both S. orientale and S. oleraceus (Boutsalis and Powles 1995), and in 1997 for R. raphanistrum (Hashem et al. 2001). Since then, S. orientale has evolved resistance to 2,4-D, atrazine and diflufenican (Preston et al. 2013; Dang et al. 2017, 2018), S. oleraceus has evolved resistance to chlorsulfuron, glyphosate, and 2,4-D (Cook et al. 2014; Merriam et al. 2018; Krishnan et al. 2025) and R. raphanistrum has evolved resistance to atrazine, 2,4-D, diflufenican, and glyphosate (Walsh et al. 2004; Ashworth et al. 2014).

Surveys of weed resistance can be used to better understand the extent and distribution of herbicide resistance in farming systems. Surveys of herbicide resistance in grass weeds in south-eastern Australia have identified high frequencies of resistance to herbicides in Lolium rigidum (Broster et al. 2022a) and lower levels of resistance in Avena spp. and other grass weeds (Broster et al. 2023a). A survey of R. raphanistrum from Western Australia (WA) found resistance to chlorsulfuron, 2,4-D and diflufenican was widespread in the state (Owen et al. 2015a). Despite these findings there have been few surveys of broadleaf weed resistance in south-eastern Australia (Merriam et al. 2018). This study was undertaken to determine the extent and distribution of herbicide resistance in R. raphanistrum, S. orientale, and S. oleraceus in south-eastern Australia.

Materials and methods

Crop survey

From 2013 to 2017, surveys were conducted in the cropping regions of South Australia (SA), Victoria (VIC), Tasmania (TAS), and southern New South Wales (NSW) (Fig. 1) to collect seeds of residual weeds from crop fields. Fields were visited at random within designated regions (Table 1) by travelling for a predetermined distance (5 or 10 km depending upon the size of surveyed area) with two or three regions visited in any year. At each stop, a single field was surveyed with the aim to visit a minimum of 120 crop fields in each region and collect Raphanus raphanistrum L., Sisymbrium orientale L., and Sonchus oleraceus L. Seeds were collected by removing mature seed pods or heads from the plants. After collection, the samples were stored for after-ripening with all other seed samples collected in the survey. A more detailed description of the sampling methodology is described by Boutsalis et al. (2012) and Broster et al. (2022a). However, the samples collected represent the potential seed bank input and therefore the resistance status of the weed populations present next year.

Fig. 1.

Survey area across south-eastern Australia. Each point represents a field where one or more of the three broadleaf weeds were collected. The different symbols and colours represent different weed species present: Raphanus raphanistrum (yellow circles); Sisymbrium orientale (red triangles); Sonchus oleraceus (blue squares); R. raphanistrum and S. orientale (blue circles); R. rapahnistrum and S. oleraceus (orange circles); S. orientale and S. oleraceus (pink circles); and all three species (green circles).

Table 1.Number of fields visited and samples collected for three broadleaf weeds (Raphanus raphanistrum, Sisymbrium orientale, and Sonchus oleraceus) by region and year.

Region	Year surveyed	Fields visited	Number of samples collected
			R. raphanistrum	S. orientale	S. oleraceus
SA
Eyre Peninsula	2014	167	0	14	24
Mid-north	2013	150	0	16	47
Mallee	2017	164	4	0	29
South-east	2017	74	9	0	90
VIC
Wimmera Mallee	2015	140	0	19	29
Southern	2014	116	14	0	18
North-east	2016	122	0	0	72
NSW
Western	2015	164	2	23	59
Plains	2016	152	6	12	35
Southern	2017	162	4	8	32
Slopes	2013	199	3	1	33
High Rainfall	2014	75	7	2	8
TAS	2014	75	25	0	12
Total		1760	74	95	488

Results

A total of 1760 crop fields were visited; however, broadleaf weeds were identified in only a minority of the fields. R. raphanistrum was collected from only 74 fields and S. orientale from 95 fields (Table 1). S. oleraceus was more common, being collected from 488 crop fields. The distribution of R. raphanistrum and S. orientale were localised, being absent from 4 and 5 of the 13 regions surveyed, respectively. In contrast, S. oleraceus was collected in every region (Table 1).

Due to the low number of weed samples collected from some cropping regions, resistance frequencies were calculated on a state basis (Table 3). Of the herbicides tested, the highest frequency of resistance was to the ALS-inhibiting herbicide chlorsulfuron (Table 3). Resistance was present to this herbicide in all three species in all four states. Overall, the frequency of resistance to chlorsulfuron was higher in S. oleraceus (both P = <0.0001; Table 4) than the other two weed species, and higher in S. orientale than in R. raphanistrum (P = 0.0001; Table 4).

Imazamox + imazapyr is an ALS-inhibiting herbicide combination from a different chemical class to chlorsulfuron. Only R. raphanistrum and S. orientale were tested for resistance to this herbicide. Resistance to imazamox + imazapyr was identified in R. raphanistrum in all states, except NSW. The frequency of resistance to imazamox + imazapyr in R. raphanistrum at 15% was similar to the frequency of resistance to chlorsulfuron (17%) (P = 0.74; Table 4). Resistance to imazamox + imazapyr in S. orientale was identified in SA and VIC only. In contrast to R. raphanistrum, resistance to imazamox + imazapyr in S. orientale occurred at a lower frequency compared to resistance to chlorsulfuron (P = <0.0001; Table 4). Overall, only 6% of S. orientale samples tested had resistance to imazamox + imazapyr, compared to 34% with resistance to chlorsulfuron (Table 3).

Atrazine is a PSII-inhibiting herbicide widely used to control broadleaf weeds in Australia. Resistance to atrazine was identified in S. orientale samples from SA and VIC, but not NSW (Table 3). In contrast, resistance to atrazine was found in R. raphanistrum samples from NSW, but not SA, VIC or TAS. S. oleraceus was not tested to this herbicide. Overall, the frequency of resistance to atrazine was low; 2% of R. raphanistrum samples and 6% of S. orientale samples with S. orientale having more resistance to this herbicide than R. raphanistrum (P = 0.019; Table 4).

Diflufenican is a phytoene-desaturase-inhibiting herbicide used in Australia for the control of Brassicaceous and other broadleaf weeds (Dang et al. 2018). As with resistance to atrazine, resistance to diflufenican was identified in S. orientale samples from SA and VIC, but not NSW (Table 3), and resistance in R. raphanistrum was identified only in NSW. Overall, the frequency of resistance to diflufenican was greater in S. orientale (16%) than R. raphanistrum (5%) (P = <0.0001; Table 4).

2,4-D is an auxin mimic herbicide widely used for broadleaf weed control in cereal crops and chemical fallows. Resistance to 2,4-D was found in all three weed species in SA and Vic (Table 3). Resistance to 2,4-D was present in 39% of R. raphanistrum samples tested from SA and 32% of S. orientale samples tested from VIC. Resistance to 2,4-D was also identified in S. orientale in NSW and R. raphanistrum in TAS (Table 3). Overall, the frequency of resistance to 2,4-D was 12% in R. raphanistrum, 9% in S. orientale, and 3% in S. oleraceus. The frequency of resistance to 2,4-D was greater in both R. raphanistrum and S. orientale than in S. oleraceus (both P = <0.0001; Table 4).

Glyphosate is an EPSPS (enzyme 5-enolpyruvylshikimate-3-phosphate synthase) inhibiting herbicide that is widely used prior to crop seeding and in glyphosate tolerant (RoundupReady) canola (Brassica napus) (Hudson and Richards 2014). No resistance to glyphosate was found in any of these weed species in this survey (Table 3).

Of the samples collected, 66% of R. raphanistrum samples were susceptible to all herbicides tested, compared with 55% of S. orientale samples and 30% of S. oleraceus samples (Table 5). Most resistant populations of all three species were resistant to only one herbicide mode of action. A total of 7% of R. raphanistrum samples, 8% of S. orientale samples, and 2% of S. oleraceus samples were resistant to two herbicide modes of action. Only S. orientale had samples resistant to three herbicide modes of action (5%) and no species had samples with resistance to more than three herbicide modes of action (Table 5).

Discussion

This survey of more than 1700 crop fields in south-eastern Australia identified the broadleaf weeds R. raphanistrum, and S. orientale were not as common as grass weeds in this region and S. oleraceus was less common than ryegrass and wild oats, but more common than brome grass (Bromus spp.) and barley grass (Hordeum spp.) (Broster et al. 2022a, 2023a). Previous herbicide resistance surveys in south-eastern Australia identified L. rigidum in 91% of sample fields in VIC, 86% of fields in SA (Boutsalis et al. 2012), 76% and 65% of fields in southern (Broster et al. 2011a) and south-western (Broster et al. 2013) NSW, respectively, and 90% of fields in TAS (Broster et al. 2012). Avena spp. were found in 43% and 54% of fields in southern (Broster et al. 2011b) and south-western (Broster et al. 2013) NSW, respectively, and 20% of fields in TAS (Broster et al. 2012). Bromus spp. were less common than the other two grass weed species, and were found in 41% of fields in SA, 25% of fields in VIC (Boutsalis et al. 2014), 6% of fields in Vic (Broster et al. 2010) and 8% of fields in TAS. Of the broadleaf weeds, S. oleraceus was most common being present in 28% of fields across the region, followed by S. orientale in 5% of fields, and R. raphanistrum in 4% of fields (Table 1). R. raphanistrum was more common in a survey of WA fields where it was collected from 21% of fields (Owen et al. 2015a) compared with south-eastern Australia.

As in south-eastern Australia, grass weeds were more common in WA crop fields with L. rigidum collected from 78% of fields (Owen et al. 2014), Avena spp. collected from 23% of fields (Owen and Powles 2016) and Bromus spp. collected from 20% of fields (Owen et al. 2015b). Likewise, Beckie et al. (2013) also found grass weeds were more common in crops in the Canadian prairies. In their surveys, the most common weeds were the grass weeds Avena fatua (present in 68% of fields), Lolium persicum (present on 21% of fields), followed by the broadleaf weeds Thapsia arvense and Galium spp. at 11% of fields each, and Polygonum convolvulus and Chenopodium album in 9% of fields each. There are several factors that influence the diversity of weed species in cropping systems including: type of crop, soil type, and environment (Fried et al. 2008, Hawes et al. 2010). Within crop types, management factors can become important (Fried et al. 2008). Cropping systems with high frequency of cereal crops in rotations tend to contain higher frequencies of grass weeds (Blackshaw et al. 1994; Derksen et al. 2002; Murphy and Lemerle 2006). Broster et al. (2022b) reported that broadleaf weeds were less common than expected in cereal crops compared with broadleaf crops in a study that comprised 752 of the NSW fields in this study. In south-eastern Australia, 78% of fields visited were sown to the cereal crops, wheat and barley (Hordeum vulgare) (Broster et al. 2022a). This is similar to the frequency of cereal crops in WA at 81% (Owen et al. 2014), but higher than the 60% of fields sown to cereal crops in the Canadian prairies (Beckie et al. 2013). The high frequency of cereal crops likely explains the dominance of grass weed species in all three systems.

Resistance was present in all three broadleaf weeds collected in south-eastern Australia. The levels of resistance in broadleaf weeds are generally low, except for resistance to chlorsulfuron (Table 3). Resistance to ALS-inhibiting herbicides, like chlorsulfuron, is common in broadleaf weeds. Of the 176 weed species with resistance to ALS-inhibiting herbicides, more than 100 are broadleaf weed species (Heap 2024). Resistance to ALS-inhibiting herbicides occurs rapidly (Boutsalis and Powles 1995; Preston and Powles 2002). The frequency of resistance to sulfonylurea herbicides in R. raphanistrum in south-eastern Australia at 17% of samples was much lower than resistance in WA, where 73% of R. raphanistrum samples tested had >20% survival to chlorsulfuron (Owen et al. 2015a). In the Canadian prairies, all Sonchus asper samples, 40% of Stelleria media samples and 16% of Galium spp. samples tested had resistance to ALS-inhibiting herbicides (Beckie et al. 2013). Resistance to ALS-inhibiting herbicides is often the result of target site mutations in ALS gene; however, non-target site mechanisms can also occur (Yu and Powles 2014). Differences in specific mutations provide differences in resistance to sulfonylurea and imidazolinone herbicides. For example, in R. raphanistrum in WA, there was a much higher frequency of resistance to sulfonylurea herbicides than imidazolinone herbicides (Owen et al. 2015a). In south-eastern Australia, there was no difference in the frequency of resistance between these two groups in R. raphanistrum, but there was in S. orientale (Table 4). These differences suggest differences in the selection intensity between regions of Australia in imidazolinone tolerant crops.

Resistance was also present to other herbicides in all three broadleaf weed species (Table 3). Resistance to atrazine was present in both R. raphanistrum and S. orientale at low frequencies. Atrazine resistance is more common in broadleaf weeds overseas than in Australia. Similar to the results here, atrazine resistance was identified in only one of 96 R. raphanistrum samples collected in a survey in WA (Owen et al. 2015a). Triazine herbicides are mainly used in canola and pulse crops in Australian grain production. As these represent less than 20% of the crops grown in south-eastern Australia (Broster et al. 2022a), the selection pressure in Australian grain production for resistance remains low.

Of greater concern, is resistance to 2,4-D, which was present in all three weed species (Table 3). 2,4-D is a relatively inexpensive herbicide, which is widely used for broadleaf weed control in cereal crops (Preston et al. 2013). The loss of 2,4-D and the ALS-inhibiting herbicides due to resistance reduces broad-spectrum broadleaf herbicide options, making broadleaf weed control in cereals more difficult (Walsh et al. 2004; Preston et al. 2013). Resistance to 2,4-D was not found in a weed survey of the Canadian prairies (Beckie et al. 2013) or a survey of Amaranthus rudis in Missouri, USA (Schultz et al. 2015); however, about a third of R. rapahnistrum samples collected from WA had greater than 20% survival when treated with this herbicide (Owen et al. 2015a). Multiple resistance, defined here as resistance to two or more herbicide modes of action, was relatively rare in this survey, affecting 7% of R. raphanistrum samples, 13% of S. orientale samples, and 2% of S. oleraceus samples (Table 5). This is in contrast to a survey of R. raphanistrum in WA where Owen et al. (2015a) concluded that multiple resistance was widespread and prevalent in this weed species. Likewise, more than half of the A. rudis populations in Missouri had multiple resistance (Schultz et al. 2015). The lower level of multiple resistance in broadleaf weeds in south-eastern Australia provides an opportunity for farmers to use a diverse range of herbicides to control these weeds. Rotation of herbicide modes of action or herbicide mixtures involving different modes of action will be necessary to reduce the selection pressure for resistance.

Conclusion

This research has shown that while broadleaf weeds are less common in crop fields of south-eastern Australia than grass weeds, herbicide resistance has evolved in these weeds. Resistance was present in at least two of the weed species to herbicides in four different modes of action. This highlights the risks of relying on herbicides alone for broadleaf weed control in grain crops. Additional weed control practices will need to be introduced. These could include tillage, crop competition or harvest weed seed control (Cirujeda et al. 2003; Walsh et al. 2013; Van der Meulen and Chauhan 2017). However, tillage is unlikely to be effective on R. raphanistrum, as its seeds can easily emerge from depth (Reeves et al. 1981). Likewise, harvest weed seed control is unlikely to be effective on S. oleraceus, as its seeds are easily dislodged from the plant and dispersed by wind (Widderick et al. 2010), therefore, different management practices will be needed for each of these weed species because of inherent differences in their biology which will affect their suitability for various management practices.