| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Agric. and Agri-Food Canada (AAFC), Lacombe Res. Center, 6000 C&E Trail, Lacombe, AB, Canada T4L 1W1
b AAFC, Lethbridge Res. Center, Box 3000, Lethbridge, AB, Canada T1J 4B1
c AAFC, Beaverlodge Exp. Farm, Box 29, Beaverlodge, AB, Canada T0H 0C0
d AAFC, Scott Res. Farm, Box 10, Scott, SK, Canada S0K 4A0
e AAFC, Semiarid Prairie Agric. Res. Center, Box 1030, Swift Current, SK, Canada S9H 3X2
f Dep. of Plant Sci., Univ. of Saskatchewan, Saskatoon, SK, Canada S7N 5A8
g AAFC, Brandon Res. Center, Box 1000A, R.R. 3, Brandon, MB, Canada R7A 5Y3
h Dep. of Plant Sci., Univ. of Manitoba, Winnipeg, MB, Canada R3T 2N2. Lacombe Res. Centre Paper no. 1080
* Corresponding author (harkerk{at}agr.gc.ca)
Received for publication June 6, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: CT, conventional tillage • FALL, post-harvest interval • GR, glyphosate-resistant • HDS, high-disturbance, direct seeding • LDS, low-disturbance, direct seeding • POST, post in-crop spray interval • PREP, preplanting interval • PRES, pre in-crop spray interval
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In the course of growing and harvesting canola, seeds are inadvertently scattered on the soil surface. Seed losses vary greatly from year to year (Bowerman, 1984), and also by crop life cycle (winter or spring) and harvest method (Price et al., 1996). Lutman (1993) suggested that oilseed rape losses in the United Kingdom can reach 10 000 seeds m–2. Gulden et al. (2003a) reported average canola losses from 35 Saskatchewan fields to be 3000 viable seeds m–2. Both estimates are many times greater than normal canola seeding rates. Clearly, there is remarkable potential for a volunteer canola problem.
Studies on crop yield losses due to volunteer canola interference are very limited. In Britain, Brain et al. (1999) demonstrated that an average of 52 rapeseed plants (Brassica napus L.) reduced winter wheat yields >50%. There are no similar reports published in North American scientific journals. However, wild mustard [Brassica kaber (DC.) L.C. Wheeler] has a similar growth habit to canola, and perhaps also has a similar competitive index. Dahl et al. (1982) determined that 6 to 107 uncontrolled wild mustard plants m–2 reduced wheat yields by 17 to 55%. Twenty wild mustard plants m–2 reduced navy bean (Phaseolus vulgaris L.) yield by 52% (Wall, 1993).
Notwithstanding the large number of seeds remaining on the soil surface after canola harvest, there is great variation in the proportion of volunteers that are recruited the following year. In Quebec, the average density of volunteer canola in the year following canola crops was 5 plants m–2 (Simard et al., 2002). Conversely, Lawson (2005) found much higher volunteer densities in Manitoba farmer fields; depending on year, site, and tillage differences, a range of 6 to 2015 plants m–2 emerged in the year following a canola crop. Volunteer canola densities after weed management practices have been employed are generally low. In the most recent western Canada postmanagement weed surveys, in fields where volunteer canola was detected, average canola densities in Alberta, Saskatchewan, and Manitoba field crops were 4.7, 5.4 and 2.7 plants m–2, respectively (Leeson et al., 2001, 2002, 2003). In the same surveys, average volunteer canola density across all surveyed fields ranged from 0.4 to 0.6 plants m–2.
There are no reports documenting the amount of unharvested canola seed that succumb to fungi or soil microbes, or to insect, avian, or small mammal predation. Some canola seeds also undoubtedly lose viability due to abiotic stressors such as mechanical damage, frost, desiccation, or flooding. In western Canada it is probable that, given favorable soil moisture conditions, much of the unharvested canola seed germinates in the fall and dies over the winter. Given adequate soil moisture, it is not uncommon to observe canola seedlings at densities above 500 plants m–2 in the fall succeeding canola harvest (K.N. Harker, personal observation, 2004).
Canola seeds may extend their persistence and viability in the soil seedbank by developing secondary dormancy. Canola secondary dormancy and persistence are likely to be reduced when unharvested seeds are not subjected to darkness (usually via tillage) immediately after harvest (Gruber et al., 2004; Pekrun and Lutman, 1996; Pekrun et al., 1998; Sparrow et al., 1990). Specific canola genotypes have differing potentials for secondary dormancy development (Gulden et al., 2003b; Pekrun et al., 1997). Gulden et al. (2004) found that canola genotype contributed to between 44 and 82% of the total variation in secondary seed dormancy found among 16 cultivars. Osmotic stress and temperature also influence secondary dormancy in canola seeds (Momoh et al., 2002). Weedy relatives of canola generally have much greater soil longevity than canola (Chadoeuf et al., 1998; Chepil, 1946). Chadoeuf et al. (1998) reported that interspecific hybrids involving canola had lower seed viability than canola itself.
Simard et al. (2002) reported that volunteer canola can persist at least 4 to 5 yr in Quebec cropping systems. Similar observations have been made in western Canada (A.G. Thomas, personal observation, 2005). The proportion of persisting plants from seeds shed in original planted fields vs. plants from seeds of volunteers recruited in subsequent years is unknown. Given the diverse markets that canola can or could be sold into, it is important to be able to segregate and preserve the identity of different canola types (Friesen et al., 2003). Pollen flow among different genotypes will complicate identity preservation considerably (Hall et al., 2000).
Our objective was to determine GR canola persistence in several western Canadian seeding systems in continuous cropping and fallow rotations.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
The experimental design was a randomized complete block with a factorial treatment arrangement consisting of a continuous cropping or fallow-crop rotation with five seeding system combinations in four blocks. The five seeding system combinations were early, normal, and late seeding in low-disturbance direct seeding (LDS), and high-disturbance direct seeding (HDS) and conventional tillage (CT) at normal seeding dates. Early and late seeding were defined as the time period when <20% and >80% of local seeding for a particular crop was completed, respectively.
Research-scale hoe-, knife-, or disc-drills were used to plant locally adapted cultivars at the different sites. Conventional tillage plots were tilled once in the fall ("spike-tooth") (including 2000 after seed spread) and once in the spring ("V-blade sweep") before seeding. High-disturbance direct seeding was accomplished with a V-blade sweep opener attachment. Crop row spacing varied from 17 to 30 cm. Target seeding rates for wheat, field pea, and barley were 250, 120, and 250 seeds m–2. All plots were fertilized according to soil test recommendations. Field peas were inoculated with an appropriate Rhizobium strain to facilitate N2 fixation.
The continuous cropping rotation was wheat–field pea–barley from 2001 to 2003. The fallow–crop rotation was fallow–field pea–fallow from 2001 to 2003. Plot size varied from 8 by 10 m to 8 by 15 m at all sites except Swift Current, where plots were 5 by 16 m.
Herbicides or tillage, or both practices were utilized to ensure that no canola volunteers set seed; specifically, seed production from newly recruited plants was prevented. Preseed glyphosate at 450 to 1320 g a.e. ha–1 was applied in LDS systems and tank-mixed with 2,4-D [(2,4-dichlorophenoxy)acetic acid] at 420 to 560 g a.i. ha–1 when necessary for volunteer canola management. In 2002, when pre-seed volunteer canola management was required, a commercial formulation of bromoxynil [3,5-dibromo-4-hydroxybenzonitrile]–MCPA [4-chloro-2-methylphenoxy)acetic acid] at 560 g a.i. ha–1 replaced 2,4-D in the glyphosate preseed treatment to reduce the risk of field pea injury. In the early LDS seeding system, preseed herbicide treatments were often not required as few weeds had emerged. Any volunteer canola plants that escaped herbicide and tillage treatments were hand-weeded before maturity.
Registered herbicides were used for general in-crop weed control including volunteer GR canola control. In wheat (2001), clodinafop (2-propynyl(R)-2-(4-(5-chloro-3-fluoro-2-pyridinyloxy)-phenoxy)-propionate)] at 56 g a.i. ha–1 was applied in a tank-mix with a commercial formulation of MCPA at 407 g a.e. ha–1, mecoprop [(±)-2-(4-chloro-2-methylphenoxy)propanoic acid] at 93 g a.e. ha–1, and dicamba [3,6-dichloro-2-methoxybenzoic acid] at 93 g a.i. ha–1. In field peas (2002), a commercial formulation of imazamox {2-[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2-yl]-5-(methoxymethyl)-3-pyridinecarboxylic acid} at 15 g a.e. ha–1 and imazethapyr {2-[4,5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2-yl]-5-ethyl-3-pyridinecarboxylic acid} at 15 g a.e. ha–1 was applied. If local monocot weed infestations warranted, supplemental herbicide, clethodim [(±) 2-{(E)-1-{3-chloroallyloxyimino]propyl]-5-{2-(ethylthio)propyl}-hydroxycyclohexen-2-one] at 45 g a.i. ha–1, was tank-mixed with the imazamox at 15 g a.e. ha–1 and imazethapyr at 15 g a.e. ha–1. In barley (2003), tralkoxydim {2-[1-(ethoxyimino)propyl]-3-hydroxy-5-(2,4,6-trimethylphenyl)-cyclohex-2-enone} at 200 g a.i. ha–1 and a commercial formulation of bromoxynil at 280 g a.i. ha–1 and MCPA at 280 g a.e. ha–1 were applied. Where necessary, 2,4-D at 420 to 560 g a.e. ha–1 was applied to all plots in the late fall for winter annual weed management. When applicable, herbicides were applied with recommended and registered adjuvants (Ali, 2005).
Fallow treatments in LDS fallow rotations were all chemical–fallow. Weeds (other than volunteer canola) were controlled with glyphosate at 450 to 900 g a.e. ha–1. All volunteer canola was treated with 2,4-D at 560 g a.e. ha–1 (broadcast over entire plot). In HDS fallow rotations, the first fallow treatment was chemical fallow and the remainder of fallow treatments were tillage. Cultivation in HDS and CT fallow rotations was before canola volunteers reached the four-leaf stage. In CT fallow rotations, an average of two or three fallow cultivations were completed during each growing season.
Volunteer canola counts were determined preplanting (PREP), in-crop prespray (PRES), in-crop post-spray (POST) (3–4 wk after herbicide application), and in the fall (FALL) from 2001 to 2003 in 20 randomly placed 0.25-m2 quadrats (10 quadrats at Winnipeg) in cropped areas in each plot. In fallow treatments, volunteers were counted and grouped into intervals that corresponded with the counts in the continuous cropping rotation. At maturity, crops were harvested and threshed with a combine.
At the end of the 2003 growing season in September, soil samples were taken from all plots to assess the canola seedbank. Twenty, 10-cm diameter, 10-cm deep soil cores were taken from a W pattern in each plot. The samples for each plot were bulked, mixed thoroughly, and a 10-L subsample was retained. The subsamples were immediately frozen until November. In the winter of 2003 to 2004, the soil samples were thawed for at least 3 d and placed in the greenhouse in 43 by 55 cm trays (5 cm deep) (Cardina and Sparrow, 1996) and watered and fertilized to encourage plant growth. Three weeks later, all emerged plants were treated with glyphosate at 900 g a.e. ha–1 and GR canola plants were counted and removed 2 wk after glyphosate application. Soil samples were then subjected to additional freezing (3 wk), thawing (3 d), soil mixing, fertilizer, and watering cycles until no further canola plants emerged. Two cycles were completed at all sites and no site required more than three cycles to exhaust the canola seedbank.
Statistical Analyses
An analysis was conducted to examine the probability of detecting volunteer canola plants in a particular rotation (Fig. 2
). Canola density data were converted to presence (1) or absence (0) based on a binomial proportion; the number of quadrats per plot containing one or more volunteer plants divided by total number of quadrats per plot. The analysis was conducted with the GLIMMIX macro in conjunction with the PROC MIXED procedure of SAS (Littel et al., 1996; SAS Institute, 1999). Data were averaged across seeding systems to reduce the number of zero values in the analysis to facilitate model convergence. Only 2001 and 2002 data were analyzed because of the infrequent occurrence of volunteer canola in 2003. Blocks and locations were assigned as random effects, and time period and rotation system were assigned as fixed effects. The analysis used the binomial proportion with a corresponding binomial probability distribution and logit link function model specifications.
|
–10
).
|
|
|
|
|
|
|
|
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Dry conditions prevailed at Beaverlodge in May and June for the entire study (Fig. 1). In 2001, with the exception of July, less than normal precipitation was received every month from April to October. Spring precipitation was higher than normal in June and July in 2000 and 2001. In 2002, Lacombe received much less than half of normal precipitation from April to July; 2003 was also dry in Lacombe. In 2001, Lethbridge received much less than normal precipitation from May through October. In contrast, record precipitation led to widespread flooding at Lethbridge in 2002. Indeed, more precipitation was recorded in Lethbridge in June of 2002 than is received in normal years from May through October. At Saskatoon, May was drier than normal from 2000 to 2004. Overall, moisture deficits were more common at Saskatoon than moisture abundance. At Scott, precipitation very rarely registered above normal and was often limiting. Early season precipitation at Swift Current was greater than normal in 2000. At Swift Current, as at Lethbridge, 2002 was also a year of excess precipitation, while in 2001 mean growing season precipitation was only 60% of normal. Precipitation was rarely limiting at Winnipeg for the entire study. In 2000, Winnipeg received 200% of normal precipitation in June, July, and August.
During most of the growing season, mean monthly temperatures were normal at all locations (data not shown). However, in 2002, April or October was unseasonably cold at several locations. For example, in April 2002, mean monthly temperatures were below 0°C at Beaverlodge, Lacombe, Saskatoon, Scott, and Swift Current; normal April temperatures at those locations average 4.2°C. In October 2002, mean monthly temperatures were also below 0°C at Saskatoon and Winnipeg; normal October temperatures at those locations average 4.9°C.
Volunteer Canola Occurrence Probabilities
It was necessary to average all data across seeding systems to facilitate ANOVA; an abundance of zero occurrence values precluded ANOVA model convergence and statistical comparisons among different seeding systems (Fig. 2). No statistical analyses on volunteer occurrence probabilities were possible in 2003.
The majority of canola volunteers emerged in the 2001 growing season following their initial dispersal on the plots in October 2000 (Fig. 2). Without additional seed production, canola volunteers in 2002 and 2003 were usually scarce. In 2002, the probability of finding one or more canola seedlings at the PRES interval in a continuous cropping or fallow rotation plot was 5 or 4%, respectively (lower 95% confidence intervals were 1 and 0%, respectively). Detection probabilities for all other intervals in 2002 averaged 0.5%. In 2003, detection probabilities for volunteer canola were all 0.
In the year following canola seed dispersal, volunteer canola occurrence probabilities varied between the continuous cropping and fallow rotation systems (Fig. 2). At PREP, PRES, and FALL intervals, canola volunteers were most likely to be growing in the continuous cropping rotation. However, at the POST interval, canola volunteers were most likely in the fallow rotation. Midseason tillage operations in HDS and CT seeding systems may have stimulated greater canola emergence (Staniforth and Wiese, 1985) in the fallow rotation at the POST interval. Overall, data from 2001 and 2002 suggest that volunteer canola management opportunities are greatest at the PRES interval.
Canola Seedbank: September 2003
By the fall of 2003 we detected almost no GR canola in the soil seedbank. Two canola seeds germinated from a single treatment at Beaverlodge and one canola seed germinated from a different treatment at Lacombe; at all other sites no viable canola seeds were detected. These results confirm that seed dormancy and viability in canola is much shorter than in weedy Brassicaceae species (Chadoeuf et al., 1998; Chepil, 1946). Accordingly, when seed production is prevented in newly recruited plants, canola seedbanks in the soil are rapidly depleted.
Location Variance Components: Canola Density
As expected, given the site diversity across western Canada, location was responsible for much of the variation in canola density (Table 2). Nevertheless, the three-way interaction among location, rotation, and seeding system was significant most of the time. In all but one case (2001, PRES), the latter interaction explained at least 25% of all location and location interaction variance. Location and the interaction of location with rotation were never significant. Therefore, much of the discussion below will focus on the interaction among location, rotation, and seeding system (Fig. 3
–9).
Individual Locations
Beaverlodge
Most canola volunteers emerged in 2001, but canola was also detected at low densities in 2002 and 2003 at Beaverlodge (Fig. 3). Previous research also indicated that the majority of canola volunteers occur in the year after the canola is grown (Simard et al., 2002). An overall rotation effect (P = 0.041) indicates that canola densities at the PRES interval were slightly greater in the continuous cropping rotation compared with the fallow rotation. Seeding at the "normal" date led to slightly greater canola numbers than at the other dates (P = 0.043). The HDS led to slightly greater canola numbers than LDS (P = 0.025) or CT (0.007).
At the 2001 POST interval, canola densities in the fallow rotation ranged from 14 to 21 plants m–2, whereas those in the continuous cropping rotation ranged from 0 to 2 plants m–2 (P < 0.001). Tillage operations in the fallow rotation may have stimulated canola emergence (Staniforth and Wiese, 1985). Higher than normal precipitation in July may have also promoted canola emergence at the POST interval (Fig. 1). In both rotations, canola density was lower in LDS than in CT systems (P = 0.048). The latter is consistent with research showing that tillage shortly after seed dispersal increases secondary dormancy in canola and canola persistence (Gruber et al., 2004; Pekrun and Lutman, 1996; Pekrun et al., 1998). In addition, limited seed mortality in CT vs. LDS (Mohler, 1993) may have led to higher canola densities in the CT than LDS at the 2001 POST interval.
In 2002, there were sufficient nonzero values to complete ANOVA at the PRES and FALL intervals, but no sources of variation or contrasts were statistically significant. The overall mean canola density for 2002 was 0.5 plants m–2.
Lacombe
At Lacombe, most canola volunteered at the PRES interval in 2001 (Fig. 4). Canola density was much greater in the continuous cropping vs. the fallow rotation (P < 0.001). In both rotations, LDS and HDS led to greater canola densities than CT. Similar to the findings at Beaverlodge, at the POST interval, canola density was greatest in the fallow rotation (P = 0.017). However, because canola density was lower in HDS and CT than LDS, it was not tillage that promoted canola emergence in the Lacombe fallow rotation.
The proportion of recruited canola plants was higher at Lacombe than at any other location. Averaged across seeding systems in the continuous cropping rotation, total recruitment in the PREP and PRES intervals was 123 plants m–2. Therefore, recruitment proportion from the original 770 seeds m–2 was 16%. In Manitoba, Lawson (2005) reported volunteer canola recruitment levels ranging from 1 to 9% in controlled field experiments, but noted that recruitment could be much greater in some farm fields.
At the POST interval in 2001, LDS led to greater canola densities than HDS or CT in both rotations. Seeding date effects in LDS varied between rotations (P = 0.041). The interaction was due to much higher canola densities at the normal LDS seeding date in the fallow vs. the continuous cropping rotation.
At 17 and 32 plants m–2 in the continuous cropping and fallow rotations, respectively, canola densities at the POST interval in 2001 were considerably greater than the 3 to 5 plants m–2 reported in western Canadian postmanagement field surveys in fields where canola was detected (Leeson et al., 2001, 2002, 2003). In 2001, substantially higher than normal June and July precipitation levels at Lacombe (Fig. 1) may help explain the relatively high levels of volunteer canola at the POST interval. Also, in this study spreading canola seed immediately before winter "freeze-up" in October may have greatly reduced seed losses to arthropod and avian predation as well as seedling mortality during the winter. Assuming the latter is true, more viable canola seed would be available for germination given the constraints of this study as opposed to natural seed shatter before or during canola harvest in early September.
Canola densities at Lacombe were very low in 2002 and 2003 (Fig. 5). Overall densities averaged 1.1 plants m–2 in 2002 and 0 plants m–2 in 2003.
Lethbridge
Canola emergence was significant at the PREP and PRES intervals in 2001, but not at any other time period (Fig. 5). In 2002 and 2003, canola densities averaged 0 and 0.1 plants m–2, respectively. At the PREP interval in 2001, CT led to greater canola densities than LDS or HDS in both rotations. Fall tillage in the CT plots may have improved seed germination conditions (Staniforth and Wiese, 1985) for canola early in the spring at the PREP interval. Over the long term, 2003 continuous cropping data tend to support the conclusion that surface seed pools in LDS were probably more rapidly depleted than the continually buried seed pools in CT (Mohler, 1993).
At the PRES interval in 2001, treatment effects interacted with rotation (P < 0.001). Higher canola densities were found in early LDS plots compared with the other seeding dates in the continuous cropping rotation, but the seeding date effect was not apparent in the fallow rotation. In addition, LDS and HDS led to greater canola densities in the continuous cropping rotation, but not in the fallow rotation.
In contrast with Lacombe, canola densities at the POST interval in both rotations in 2001 averaged 0.5 plants m–2, well below those reported by Leeson et al. (2001, 2002, 2003). However, also in contrast with Lacombe, growing season precipitation at Lethbridge in 2001 was dramatically lower than normal (Fig. 1).
Saskatoon
Volunteer canola densities at Saskatoon were relatively high at the PRES interval in 2001; especially in the continuous cropping rotation (Fig. 6). Average canola densities ranged from 12 plants m–2 in the fallow rotation to 36 plants m–2 in the continuous cropping rotation. Canola densities at the POST interval in 2001 were below those determined by Leeson et al. (2001, 2002, 2003). For the remainder of the study at Saskatoon, canola densities were zero. May precipitation was well below normal in 2002 and 2003 (Fig. 1).
At the PRES interval in 2001, both LDS and HDS led to higher canola densities than CT in both rotations. No LDS seeding date effect was apparent in the fallow rotation, whereas in the continuous cropping rotation, early and normal seeding led to greater canola densities than late seeding (P = 0.002). The HDS led to higher canola densities than LDS in the continuous cropping rotation, but the opposite occurred in the fallow rotation (P = 0.001).
Scott
Similar to Lacombe and Saskatoon, at Scott most canola volunteers occurred at the PRES interval in 2001 (Fig. 7). However, too many zero values in LDS fallow rotations precluded a valid statistical analysis at that interval. Nevertheless, it was apparent that continuous cropping led to higher volunteer densities than fallow, and that higher disturbance seeding systems also favored higher canola densities. At the POST interval in 2001, there were interactions between rotations and seeding systems (P < 0.001). In the continuous cropping rotation, LDS led to slightly higher canola densities than HDS (P = 0.038). In the fallow rotation, CT led to much higher canola densities than the other seeding systems (Mohler, 1993).
Canola densities in 2002 and 2003 were low, averaging 0.3 and 0.1 plants m–2, respectively. In 2002, early spring precipitation in April and May was very limited (Fig. 1).
Swift Current
Below normal moisture conditions in April of 2001 may be responsible for the absolute lack of canola germination at the PREP interval in 2001 (Fig. 8). All other sites had some canola germinate at the PREP interval in 2001. In the continuous cropping rotation, an average of 13 plants m–2 emerged at the PRES interval in 2001, but almost no plants emerged in the fallow rotation. Perhaps seeding operations in the continuous cropping rotation provided sufficient soil disturbance to stimulate canola germination and emergence. At the POST interval, canola density was greater in the continuous cropping than in the fallow rotation. In both rotations at the POST interval, LDS and HDS led to higher canola densities than CT.
Even though very few canola volunteers emerged in 2002, emergence was consistent enough in all plots at the PRES interval for a statistical analysis. In both rotations, slightly more canola emerged in the LDS system after early seeding than at the other two seeding dates (P = 0.44).
Winnipeg
At Winnipeg, canola density at the PREP interval in 2001 was often greatest with higher disturbance seeding systems (Fig. 9). Canola density in CT was usually at least double that of other seeding systems in both rotations. At the PRES interval in both rotations, LDS (P = 0.042) and CT (0.006) led to higher canola densities than HDS. In the fallow rotation, more than 100 canola plants m–2 emerged in the CT plots at the POST interval. The latter may have been stimulated by cultivation (Staniforth and Wiese, 1985) combined with 200% of normal precipitation levels in July (Fig. 1). Volunteer canola density averaged 0.3 plants m–2 in 2002.
Total recruitment averaged across 2001 seeding systems at the PREP and PRES intervals in the continuous cropping rotation was 25 plants m–2, representing a 3.2% recruitment proportion. These recruitment levels are confirmed by other field research in Manitoba (Lawson, 2005) where volunteer canola recruitment levels ranged from 1 to 9% in various tillage systems.
All Locations: Random Effects Model
Across all locations and treatments, at intervals where herbicides can be effectively used for volunteer management (PREP and PRES), canola emergence averaged 6.2, 0.7, and 0.0 plants m–2 in 2001, 2002, and 2003, respectively (Fig. 10). Simard et al. (2002) similarly observed an average of five volunteer canola plants in fields seeded to canola in the previous year. At the PRES interval in 2001, there was a trend (P = 0.133) for higher canola volunteers in continuous cropping vs. fallow, similar to findings reported by Derksen et al. (1994). Total recruitment across PREP and PRES intervals averaged across all seeding systems in the continuous cropping rotation was 3% (25 plants m–2). This is in accordance with results from a separate study in Manitoba where recruitment ranged from 1 to 9% (Lawson, 2005). Volunteer spring canola persistence densities were greater in the current study than were volunteer spring wheat persistence densities in a parallel study at the same locations (Harker et al., 2005a).
A few weeks after herbicide application (POST) in 2001, in the continuous cropping rotation, we observed an average of 1 plant m–2. Western Canadian post management surveys reported an average three to five canola volunteers in fields where canola was detected (Leeson et al., 2001, 2002, 2003). However, in western Canadian canola fields, it is likely that higher levels of canola seed enter the seedbank (3000 seeds m–2) (Gulden et al., 2003a) than were scattered in this study (770 seeds m–2).
The highest canola densities occurred at the PREP and PRES intervals in the year following seed spread (Fig. 10). Clearly, these two intervals are the most important times to implement volunteer canola management. High volunteer canola recruitment at PREP and PRES intervals confirms other western Canadian research (Lawson, 2005). In the Northern Hemisphere, volunteer canola seedling recruitment occurs mostly in May and June; typical of a summer annual weed (Gulden et al., 2003b). Generally, few canola volunteers emerged in 2002 and 2003 (Fig. 10). In this study we prevented seed production from newly recruited plants. Therefore, subsequent to October 2000, the volunteer canola seedbank was not replenished. A low but noteworthy number of canola volunteers at the PREP and PREP intervals in the spring of 2002 suggest that some of the initial seeds scattered in 2000 were dormant or experienced induced dormancy (Gulden et al., 2003b; Momoh et al., 2002; Pekrun et al., 1997). It is unlikely that growers would find it expedient to manage an average of 0.7 canola plants m–2 at PREP or PRES intervals (Fig. 10, 2002), unless factors beyond weed–crop competition (identity preservation, restricted markets, gene flow considerations, seed production, and so forth) are involved.
Canola densities in this study were probably markedly influenced by variable soil and environmental conditions at each site. Therefore, few consistent treatment effects prevailed across all sites. Blackshaw et al. (2001, 2005) and Derksen et al. (1993) suggest that environmental conditions may influence weed population dynamics more than specific agronomic practices. Indeed, of all of the rotation and seeding system comparisons across all locations, only one effect in 1 yr and one time interval had a significant P value: 2001, PRES, R x LDS:HDS – N (Fig. 10, P = 0.013). The latter contrast confirms that in the continuous cropping rotation, volunteer canola densities were greater in HDS than in LDS seeding systems; farmer field survey data from Manitoba are confirmatory (Lawson, 2005).
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
High levels of weed control are common in GR crops (Askew and Wilcut, 1999; Blackshaw and Harker, 2003; Harker et al., 2005b; Johnson et al., 2000; VanGessel et al., 2001). In LDS cropping systems, growing GR canola may necessitate an incremental cost for controlling canola volunteers. There are herbicide alternatives for preseeding GR canola control (Rainbolt et al., 2004) that are desirable from a glyphosate resistance management perspective. However, the vast majority of LDS growers prefer the broad spectrum control, efficacy, cost, and low soil residual properties of glyphosate as a "burn-off" herbicide (Friesen et al., 2003). The cost of adding a herbicide such as 2,4-D to glyphosate burn-off treatments is not usually a deterrent for growers cultivating GR canola (Johnson et al., 2004; Serecon Management Consulting, 2005).
Our data suggest that careful volunteer canola management in the year after a canola crop may preclude additional management concerns in subsequent years. However, canola seed harvest losses can be much greater than those we simulated in this study (Gulden et al., 2003a; Lutman, 1993). In addition, volunteer canola management across millions of prairie hectares is <100% (unlike the current study), and seed from newly recruited volunteers may considerably augment canola seedbanks. Therefore, it would be prudent to use LDS systems that increase volunteer canola seed mortality and decrease secondary seed dormancy (Gulden et al., 2003b; Mohler, 1993; Pekrun et al., 1998) and to grow canola genotypes that are less prone to secondary dormancy and persistence in the soil seed bank (Gulden et al., 2003b, 2004; Pekrun et al., 1997).
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
