Sustainable Weed Control: In-Row Weed Cultivation Strategies for Midwest Vegetable Growers

Part 1: Soil Surface Effects on Flextine Cultivation Efficacy in Vegetables

Soil surface and weed density responses to compost, rolling and molasses

Surface roughness.  The surface roughness of soil just before flextine cultivation was influenced by rolling or compost additions but not by molasses (Table 1.3). As expected, rolling reduced surface roughness in both runs of the experiment.  In run 1, there was an interaction between rolling and compost addition:  Rolled soils which had received historic compost additions had 26.3% greater roughness than rolled soils which had not received compost (Figure 1.4). Surprisingly, the effect of rolling was greatest where no compost had been applied (Figure 1.4). In run 2, the same trend held, but surface roughness in rolled compost plots was not different than rolled plots without compost (Figure 1.4).

Soil micro-penetrometer resistance. Soil hardness was influenced by rolling, but not by compost application (Table 1.3). The effect of molasses application on hardness was only detected in run 2, and only in treatments that had not been rolled (Table 1.3; Figure 1.5).  Rolled plots were 73% and 85% harder compared to unrolled plots, in run 1 and 2, respectively (Table 1.4). In run 2, molasses application increased hardness by 24% in unrolled treatments, but had no effect in treatments that had been rolled (Figure 1.5).

Soil moisture.  Rolling, compost and molasses factors had no detectable effect on soil moisture prior to flextine cultivation with one notable exception (Table 1.4):  In run 1, rolled soils were 26.3% wetter than unrolled soils prior to sweet corn cultivation (Table 1.4).

Flextine effects on weed mortality and crop mortality following compost, rolling and molasses

Flextine efficacy. In both runs of the experiment, ambient weed mortality following flextine cultivation was influenced by the main effect of rolling, but not by historic compost or molasses application (Table 1.5). Rolling reduced flextine efficacy on ambient weeds by 1.7% in run 1 and 2.6% in run 2 (Table 1.6). In run 1, red amaranth surrogate mortality was also higher in unrolled treatments, but no effect was detected for mustard surrogate weeds. In run 2, germination of both surrogate weeds was poor, so no surrogate mortality could be evaluated.  As anticipated given our calibration process, crop mortality in both runs of the experiment was low, averaging approximately 8% for sweet corn and 5% for beans, with no obvious effects of compost, rolling or molasses (Table 1.6).

Flextine effects on surface roughness.  Surface roughness following flextine cultivation was influenced by the main effect of rolling in run 2, but not in run 1 (Table 1.3). Rolled soils had 38.5% lower surface roughness than unrolled soils following flextine cultivation in beans (Table 1.4). However, following cultivation in corn, surface roughness was not different between rolled and unrolled soils (Table 1.4).

Relationship between mortality and soil surface characteristics

Results from our regression analysis show that among the variables we tested, weed mortality was negatively correlated with soil hardness for all species in both runs (Figure 1.6). We were not able to detect correlations between weed mortality and soil moisture or surface roughness. Regression coefficient suggest that each one newton increase in soil penetrometer resistance was associated with a reduction in flextine efficacy of approximately 1 to 2 percentage points.

Contrary to expectations, we found that rolling a seedbed to create a level soil reduced efficacy of flextine cultivation in most cases. In both runs of the experiment, rolling resulted in reduced efficacy of flextine cultivation for control of ambient weeds (Table 1.6).  Although efficacy was greater than 95% regardless of rolling, the greater density of escaped weeds in rolled plots is likely to represent a real potential cost to growers in the form of higher handweeding costs or reduced quality and yield.  For example, in run 1 of this experiment, ambient weed density following flextine cultivation was 6-fold higher in rolled compared to unrolled treatments.

            Our results also suggest several mechanisms that help explain this counterintuitive effect of rolling on flextine efficacy.  As expected, rolling beds prior to planting produced a more level soil surface in most cases (Table 1.3; Figure 1.4), which —other things equal—would likely result in the more uniform working depth to improve efficacy.  However, although rolling reduced surface roughness, we were not able to detect any correlation between surface roughness and flextine efficacy (Figure 1.6). Rolling also affected other soil conditions which may have had a negative effect on flextine efficacy.  In particular, rolling soils increased soil micropenetrometer resistance (hardness) in both runs of the experiment, and increased soil surface moisture content in run 1 (Table 1.4).  Our regression analysis shows that among these factors, soil hardness  was most clearly associated with weed mortality (Figure 1.6).  Regression coefficients estimates suggest that a 1 N increase in soil penetrometer resistance was associated with a reduction in efficacy of 1 to 2 percentage points.  Lack of observed surface roughness effects on efficacy may have been due in part to the relatively small differences in the range of roughness evaluated in this trial;  although differences existed across rolled and unrolled bedsi (Table 1.4),  the soil surface was probably sufficiently level to facilitate tine working depths that effectively killed weeds.

            Rolling soils prior to planting represents a cost to growers in time, labor, and increased soil disturbance, and we found no benefit to this practice for flextine cultivation of large seeded vegetable crops on our loamy-sand soils. In fact, our results suggest that rolling can reduce flextine efficacy through its adverse effects on soil penetrometer resistance.  However, these results may not be generalizable to other soil textures, crops, or cultivation tools.  For example, soils with finer textures  may be rougher following primary tillage and may benefit from leveling with a roller to improve cultivation equipment function and reduce potential safe-sites for weeds. Levelling may also provide benefits for smaller seeded crops such as carrots or beets by improving planting depth, stand establishment and tolerance to cultivation.   

            Our results also did not support our initial hypotheses that historic compost additions would improve cultivation efficacy (Table1.5). We anticipated that compost additions would improve efficacy by improving tool movement through the soil as previously asserted in several studies (Bowman 1997; Mohler 2001).  However, these potential benefits on tool action may have been offset by several negative indirect effects.  First, compost applications may have influenced flextine efficacy through effects on the density or growth of weeds and hence their ability to tolerate cultivation.  For example, in run 1, historic compost addition was associated with higher densities of ambient weeds (Table 1.6). Although no clear differences in weed growth prior to cultivation were observed, it is also possible that weeds growing in composted soil had more vigorous root systems that improved their tolerance to cultivation. Second, it is interesting to note that compost itself in some cases contributed to greater surface roughness (Figure 1.4) which may have also offset any benefits of compost for flextine efficacy.  We speculate that this effect may have been due to aggregates present in the compost applied in the spring just before initiation of the experiment. Such aggregates may have produced safe-sties for weeds that offset any longer term benefits for cultivation of compost addition on soil tilth.

            The lack of compost effects on flextine efficacy may have also been due to the limited effects of that compost on soil tilth. Note that our study was conducted on a loamy-sand soil common in Midwest vegetable production. Potential changes in soil organic matter and structure in such soils is limited compared to more fine textured soils.  In our case, soil organic matter (SOM) in treatments receiving annual compost for 12 years was 1.6%, compared to 1.3%, in soils receiving no compost.   Moreover, both compost and non-compost treatments had identical cover crop and tillage management over this 12 year period, which may have overshadowed compost effects on SOM . It is possible that this difference in SOM between treatments was simply not sufficient to impact soil tilth and tool function. Indeed, we observed no impacts of compost on soil moisture or hardness (Table 1.4) that might have been expected in soils with significant changes in SOM. Cultivation tool function may be improved in soils with increased SOM and tilth, but we were not able to create sufficiently distinct soil conditions in this study to adequately test this hypothesis.

            Artificial induction of soil crusting using molasses also did not have detectible effects on cultivation efficacy. Although we observed crusting due to molasses in these soils in greenhouse testing, molasses effects in the field appear to have been relatively small, with little or no detectable effect on soil penetrometer resistance (Table 1.4). However, we did observe that the soil surface with molasses appeared to create visible differences in soil crusting that were not adequately captured by resistance measured with our micropenetrometer.   Again, the loamy sand soil where this trial was conducted has a lower propensity for crusting compared to finer textured soils, and our ability to produce a detectable soil crust with molasses was limited. Heavier soil types with poor soil structure are more likely to produce a crust that inhibits function of cultivation tools (Bresson and Boiffin 1990).

            Our study demonstrates that soil conditions can play an important role in cultivation tool function. Characterization of soil conditions before and after cultivation events can provide insights into the mechanisms responsible for these effects. Observed discrepancies between tool function across years and sites may be attributable to differences in soil characteristics including roughness, moisture, and crusting.  Unfortunately, soil conditions are often overlooked in cultivation studies, making inferences about the mechanisms of tool efficacy difficult or impossible. Ultimately, we’d like to be able to tell growers which tool will work best under which conditions, and give insights into management decisions. For instance, we saw in this study how soil compaction effects from rolling on efficacy might not be obvious when preparing level beds for cultivation. Including mechanistic information in cultivation studies is critical for providing growers with the information they need. Further research opportunities exist to more deeply explore how these characteristics impact cultivation efficacy and make management recommendations to growers.

Part 2: Integrated Physical and Cultural Weed Management in Table Beets

Cultivar x depth studies

Greenhouse

Total number of emerged seedlings.  The final number of seedlings seed^-1 to emerge was influenced by the main effect of cultivar but not by depth of planting (Table 2.4). B and TG varieties had the greatest total seedling emergence, and M had the least (Table 2.4). Final emergence of CG was higher than M in run 1, but lower than B and TG in run 2 (Table 2.4).

Percent emergence days 4-8.  The percentage of final seedlings that had emerged at 4-8 days appears to have been influenced by both cultivar and depth (Figures 2.2 and 2.3).  In run 1, B and CG planted at 4 cm appear to have emergence delayed by 1 day, relative to other depths, without a noticeable reduction in yield (Figure 2.2a and 2.2b). TG and M in run 1 seem to have reduced germination at 4 cm depth (Figure 2.2c and 2.2d). In run 2, delay in B and CG emergence based on depth was not noticeable, but TG planted at 4 cm is much lower than other depths at day 7, before catching up with other depths at day 8 (Figure 2.3).

Field trial       

            Final emergence was influenced by the main effects of cultivar and planting depth, however no interactions between cultivar and depth were detected (Figure 2.4). Beets planted at 4cm had lower final emergence compared to beets planted at 1 and 2cm. B germinated with more seedlings seed^-1 than M and TG varieties (Figure 2.4). 

            Shoot biomass was influenced by the main effect of cultivar, while root biomass and the root-shoot ratio (RSR) were influenced by cultivar and depth main effects (Table 2.5). Both root and shoot biomass were greater for B than M and TG. TG had particularly small roots, with biomass representing only 24% that of B and 42% that of M.  Lower root biomass of TG was due to both smaller overall plant growth, and lower partitioning to root tissue relative to shoot tissue (lower RSR).

            The effects of planting depth on root biomass and RSR—but not shoot biomass—differed for B and M (Table 2.5). For B, the root biomass and RSR were lower when planted at 1 cm compared to 2 cm. RSR in Moneta was also lower when planted at 1cm compared to 2cm.

Cultivar x tool field cultivation study

            At the time of cultivation, beet varieties differed in their height, anchorage force, root biomass and shoot biomass, but not in their RSR (Table 2.6). The root and shoot dry weights and anchorage forces of B and M were greater than those of TG.  B also had greater shoot length than CG and TG (Table 2.6).   No differences in dry weights, anchorage forces or heights were detected between the CG and TG cultivars, nor between TG and M cultivars. 

            Beet survival following cultivation was influenced by beet cultivar but not by tool type or tool by cultivar interactions (Table 2.7). CG beet survival following cultivation was significantly greater than TG variety, where 17.1% of beets did not survive cultivation (Table 2.7). However, no differences in survival were detected between B, M, and the other cultivars. 

Cultivation efficacy on the surrogate mustard weed was greater for finger weeders than hilling disks (Table 2.7). Compared to density prior to cultivation, mustard density three days after cultivation was 65% lower following finger weeding, but 17% greater following hilling compared to pre-cultivation density, suggesting some mustard seeds did not germinate until after cultivation (Table 2.7). However, carpetweed control was not influenced by tool, and neither weed species’ survival was influenced by beet cultivar.

            The biomass of surrogate mustard weeds was affected by both cultivar and tool, but not their interaction (Table 2.8). Mustard biomass was the lowest in B plots, while weed dry weights from CG, M, and TG plots were not different (Table 2.7). Final mustard biomass was lower in plots that had been cultivated with the finger weeder compared to the hilling disks (Table 2.7). However, yield loss due to surrogate mustard was affected neither by beet cultivar nor cultivation tool (Table 2.7).

            Yield was affected by cultivar but not by tool or tool x cultivar interaction (Table 2.8). B was the highest yielding variety among the four cultivars tested, both in marketable and total beet yield (Table 2.8). CG and M had significantly higher yield than TG, which was the lowest yielding cultivar (Table 2.8).

            As hypothesized, we found that planting beets deeper than the typical recommended 2cm depth may delay germination by a day or two, allowing for more flexibility in implementing a stale seedbed. However, this result was not consistent across cultivars or trials (Figure 2.2, 2.3, and 2.4).  Under field conditions, deep planting of beets did not obviously delay emergence, and final emergence was reduced at depths greater than 3 cm (Table 2.4). For polyembryonic beet cultivars, a slight reduction in stand may reduce the need for thinning of beets but greater stand losses could be costly for growers.  Our results are similar to those obtained by Romaneckas et al. (2009), who also found that deep planting delayed sugar beet emergence in the greenhouse, but reduced emergence under field conditions.  Such discrepancies between greenhouse and field studies likely reflect more uniform and optimal growing conditions in the greenhouse.  For example, differences by depth in soil moisture, temperature or soil resistance are much more common in the field, and undoubtedly influence relationships between seed depth and seedling emergence in some cases.  Tradeoffs associated with different planting depths must be considered based on specific field conditions.  Our results suggest that although planting beets deeper may not be practical in all soil conditions, it has the potential to give growers more flexibility in implementing a pre-emergent stale seedbed.

            A second central hypothesis of our study was that cultivars differ in their tolerance to in-row cultivation, and that this difference might be exploited to improve tool selectivity and lower weed management costs.  Our results were consistent with this hypothesis (Table 2.7), although the magnitude of differences in cultivar tolerance were relatively small.  Our primary finding was that TG tolerance to both finger weeding and hilling was lower than that of the CG variety. Poor relative tolerance of TG to cultivation is not surprising given that it was consistently among the cultivars with the shortest height and lowest anchorage force (Table 2.6).  The CG cultivar appears to both establish more slowly, and partition fewer resources to early root development than other beet cultivars (Table 2.5), and hence—other things equal—is not a good choice for improving weed management in beets.    

            We anticipated that such differences in beet cultivar partitioning to root vs shoot tissue might confer preferential tolerance to tools that killed weeds either by uprooting (fingers) versus burial (hilling disk).  However, we were unable to identify specific plant traits related to cultivation tolerance because beet mortality was too low and beet characteristics too similar. B was our most vigorous cultivar, with the greatest root and shoot biomass, and anchorage force. However, B survival was not different than CG and M cultivars. Nor did we detect any differences in beet tolerance based on tools differing in their mode of action. 

            We were surprised to find that our estimate of mustard “survival” following cultivation with the hilling disk was greater than 100% (Table 2.7). The tool hilled soil nicely in the crop row, covering mustard and ambient weed seedlings but not beets. However, buried mustards were able to recover, and some mustard seeds that had been on the soil surface prior to cultivation may have been induced to germinate by hilling, resulting in greater post-cultivation counts.

            Our study also provided insight into the relative competitive ability of beet cultivars with weeds. In particular, surrogate mustard biomass was lowest in plots with the B cultivar (Table 2.7).  This apparent weed-suppressive ability of the B cultivar was presumably due to the fact that it was among the cultivars with the highest shoot biomass and height (Table 2.6), traits shown in studies in other crops to improve weed competitive ability through shading (Colquhoun et al. 2009; Garrity et al. 1992).  Though B survival following cultivation was not different than other cultivars, its weed competitive ability makes it a good cultivar choice for suppressing weeds.          

            Cultivar choice may be an important consideration for growers managing weeds through physical and cultural approaches. Differences in both cultivation tolerance and weed competitiveness among cultivars of the same species can be  important factors in determining weed management success in certain crops (Colquhoun et al. 2009; Garrity et al. 1992; Wick et al. 1986).  However, within the conditions of our study, cultivation tolerance  appears to have been less important than weed competitive ability in determining agronomic success. It can be difficult to draw broad conclusions from weed competition in one study because weed impact is dependent on weed populations, crop conditions, soil type, and a host of other factors. Nevertheless, future research should consider both cultivation tolerance and weed competitive ability as  potentially important characteristics for reducing weed management costs in vegetable crops including table beets.  Identification of specific traits associated with these characteristics may be also be helpful for plant breeders wishing to select useful cultivars for improving weed management  However, our results suggest that straightforward selection criteria to improve weed management success are unclear.  For example, given the greater efficacy of uprooting (finger weeding) relative to burial (hilling disk) in controlling weeds in our study, selection of cultivars which partition more early season resources to roots relative to shoots might be desirable;  however, partitioning resources to roots at the expense of shoots might reduce the capacity of that cultivar to suppress weeds through shading. Early season partitioning to shoots may be a desirable trait for weed competitiveness, along with improved cultivation tolerance to burial with hilling disk. Beet breeders could also select for cultivars with delayed emergence from shallow planting to facilitate a longer stale seedbed window, further improving weed competitiveness of a shoot-partitioning cultivar.