Introduction
Weeds increase production costs and reduce marketability of container-grown nursery crops (Case et al. Reference Case, Mathers and Senesac2005). To manage weeds, growers apply preemergence herbicides every 6 to 8 wk, yet weeds continue to emerge throughout the growing season (Gilliam et al. Reference Gilliam, Foster, Adrain and Shumack1990; Neal et al. Reference Neal, Derr, Marble and Senesac2017). Lacking selective postemergence herbicides, growers rely on hand weeding to remove emerged weeds (Mathers Reference Mathers2003; Stewart et al. Reference Stewart, Marble and Pearson2017). Container nurseries use soilless substrates that are typically weed free; consequently, weed populations arise primarily from weed seed introductions (Cross and Skroch Reference Cross and Skroch1992). Documented sources of weed seed introductions to container nurseries include contaminated transplants, reused containers, wind, and dispersal from nearby mature plants (Bachman and Whitwell Reference Bachman and Whitwell1995; Cross and Skroch Reference Cross and Skroch1992). Despite available integrated crop management practices targeting these seed sources, weeds continue to emerge, leading growers to suspect overhead irrigation as a dispersal vector for weed seeds.
Hydrochory, the passive dissemination of seeds or plants by water, is an important mode of seed dispersal in natural riparian environments (Nilsson et al. Reference Nilsson, Brown, Jansson and Merritt2010; Rasran et al. Reference Rasran, Vogt, Trattnig and Bernhardt2023), flooded rice production (Li and Qiang Reference Li and Qiang2009), and other agroecosystems (Wilson Reference Wilson1980). Egginton and Robbins (Reference Egginton and Robbins1920) estimated 1.5 to 10.7 million seeds floated past fixed observation positions in flood-irrigation ditches during a 24-h period. The authors concluded that irrigation from those sources could significantly contribute to weed populations in adjacent fields. Overhead irrigation water drawn from such sources could also be an important vector for weed seed dispersal in agronomic cropping systems. Irrigation water from canals in Nebraska contained seeds of the 10 most detrimental weed species in adjacent farm fields and distributed an estimated 12,500 viable seeds per hectare in one irrigation event (Wilson Reference Wilson1980). Other researchers have suggested that hydrochory may be a significant vector for introducing new species (Kelley and Bruns Reference Kelley and Bruns1975).
Weed seeds have also been collected from source water used for irrigation in container nurseries. Williams and Sanders (Reference Williams and Sanders1984) reported collecting weed seeds from creeks and holding ponds in five nurseries. Ray et al (Reference Ray, LeBude, Altland, Harlow and Neal2026) reported germinable seeds of more than 75 taxa in irrigation water filtrates from container nursery irrigation ponds, including 28 weedy species commonly found in container nurseries. In those studies, the average number of seeds in a single irrigation cycle was between 30 and 61 ha−1. While this number may seem inconsequential, many common nursery weeds reach reproductive maturity quickly and forcefully disperse seeds significant distances (Neal and Derr Reference Neal and Derr2005). Weed seeds spread via irrigation could be an important vector for early season introductions.
Nurseries in the United States annually use about 775 billion liters of irrigation water on 212,147 hectares of crops (Paudel et al. Reference Paudel, Pandit and Hinson2016). Container nursery crops are irrigated daily except on rainy days or when evapotranspiration demands are low. While the volume of water applied each day is variable, a frequently cited standard for container nursery crops in the southeastern United States is 1.9 cm d−1, equivalent to about 187,000 L ha−1 d−1 (20,000 gal acre−1 d−1) of overhead irrigation (Fare et al. Reference Fare, Gilliam and Keever1992; Fernandez et al. Reference Fernandez, Pershey, Andresen and Cregg2019; Warsaw et al. Reference Warsaw, Fernandez, Cregg and Andresen2009). Most container nursery crop producers draw irrigation water from surface ponds (Fain et al. Reference Fain, Gilliam, Tilt, Olive and Wallace2000), often without filtration. In the southeastern United States, 75% of growers recapture excess irrigation and rainwater via drainage ditches and culverts, returning this water to surface irrigation ponds (Fain et al. Reference Fain, Gilliam, Tilt, Olive and Wallace2000; Yadzi et al. Reference Yazdi, Owen, Lyon and White2021). Through this practice, seeds produced from mature weeds in containers, nursery roadways, drainage areas, and pond edges are deposited into irrigation water.
Seed buoyancy is considered an important trait for hydrochorous transport (Boedeltje et al. Reference Boedeltje, Bakker, Brinke, Van Groenendael and Soesbergen2004; Vogt et al. Reference Vogt, Rasran and Jensen2006). Seeds with high buoyancy were dispersed over greater distances and in higher numbers than seeds with low buoyancy (Boedeltje et al. Reference Boedeltje, Bakker, Brinke, Van Groenendael and Soesbergen2004; Fryirs and Carthley Reference Fryirs and Carthley2022; Nilsson et al. Reference Nilsson, Andersson, Merritt and Johansson2002). de Rouw et al. (Reference de Rouw, Ribolzi, Douillet, Tjantahosong and Soulileuth2018) reported that heavy rainfall increased the number of weed seeds captured in runoff water of drainage channels of mountainous farmland, and floating weed seeds were transported in runoff greater distances than submerged seeds. High seed buoyancy is particularly important in slow-flowing systems, where seeds with lower buoyancy are more likely to sink and not be deposited in habitats that are suitable for germination (Nilsson et al. Reference Nilsson, Andersson, Merritt and Johansson2002). There are few reports on the buoyancy of seeds of weeds commonly found in container nurseries. Eclipta, a common weed of container nurseries in the southeastern United States has been reported to float for more than 400 h in flooded rice (Shi et al. Reference Shi, Li, Zhang and Qiang2020). However, another study by the same research group reported that eclipta and hairy bittercress (Cardamine hirsuta L.) seeds sank after 120 h (Zuo and Qiang Reference Zuo and Qiang2008). Seed buoyancy of other common container nursery weed species is largely unknown.
For weeds to be successfully distributed via irrigation, the seeds must survive in water until they are deposited in a conducive environment for germination. Hydraulic resistance time is the average time it takes water to flow from a discharge point to the intake pump. For chemical and pathogenic contaminants, Yazdi et al. (Reference Yazdi, Owen, Lyon and White2021) recommended a minimum of 72-h hydraulic resistance time for container nursery irrigation water systems. While seeds of many weed species maintain high germination percentages after months submerged in water, others do not. Comes et al. (Reference Comes, Bruns and Kelley1978) reported that 58 of 82 weed species they tested maintained high germination percentages after 12 mo submergence, but 13 species exhibited little or no germination following 3 mo of submergence. Bruns (Reference Bruns1965) reported that bigseed alfalfa dodder (Cuscuta indecora Choisy) and showy milkweed (Asclepias speciosa Torr) germinated at higher rates following 6 mo of submergence compared to dry-stored seeds. Similar research on weeds that are common to nursery crops is limited to two reports on spotted spurge. In one report, spotted spurge germination was 26% following 36 mo of submergence (Comes et al. Reference Comes, Bruns and Kelley1978); other researchers observed greater than 60% germination after 3 wk but no germination after 9 wk (Asgarpour et al. Reference Asgarpour, Ghorbani, Khajeh-Hosseini, Mohammadvand and Chauhan2015).
Data regarding seed buoyancy and germinability following submergence in water for the seeds of weed species commonly found in container nurseries are lacking. Therefore, the objectives of this research were to evaluate the buoyancy and survival after submergence of common container nursery weed seeds.
Materials and Methods
Seeds of common container nursery weeds used in these experiments were field collected from 2019 to 2021 at the Horticultural Field Laboratory in Raleigh, North Carolina (35.79161°N, 78.69783°W) and a local container nursery in Johnston County, North Carolina (35.59353°N, 78.39443°W). Species used in the study were eclipta, spotted spurge, rice flatsedge (Cyperus iria L.), redroot pigweed (Amaranthus retroflexus L.), fringed willowherb (Epilobium ciliatum Raf.), dogfennel, marsh yellowcress, large crabgrass [Digitaria sanguinalis (L.) Scop.], American burnweed [Erechtites hieraciifolius (L.) Raf. Ex DC.], livid amaranth (Amaranthus blitum L.), and common groundsel. Seeds of yellow woodsorrel and flexuous bittercress were originally collected from local nurseries, then seed numbers were increased by growing plants to maturity and harvesting seeds in greenhouses at North Carolina State University (35.78742°N, 78.67314°W). All collected seeds were air-dried in paper bags at room temperature then screened and stored at 3 C. These species were selected because they were present in nursery irrigation water filtration tests (Ray et al. Reference Ray, LeBude, Altland, Harlow and Neal2026), represent a diversity of seed-propagated cool-season and warm-season species commonly found in container nurseries (Neal and Derr Reference Neal and Derr2005), and for which we had sufficient inventories of locally collected and germinable seeds. Redroot pigweed is not common in container nurseries but was included for comparison to previous research.
Weed Seed Buoyancy
Methods for testing seed buoyancy were modeled after those described by Egginton and Robbins (Reference Egginton and Robbins1920) and by Shi et al. (Reference Shi, Li, Zhang and Qiang2020). Shi et al. (Reference Shi, Li, Zhang and Qiang2020) conducted seed flotation tests in 10-cm rings placed in a flooded rice paddy, whereas Egginton and Robbins (Reference Egginton and Robbins1920) conducted buoyancy tests in glass beakers indoors. Our tests were conducted indoors at the Horticulture Weed Science Laboratory at North Carolina State University, Raleigh, North Carolina (35.78839°N, 78.67286°W), maintained at approximately 23 C. In July 2021, buoyancy was tested on seeds of American burnweed, common groundsel, dogfennel, eclipta, fringed willowherb, large crabgrass, livid amaranth, marsh yellowcress, redroot pigweed, rice flatsedge, spotted spurge, and yellow woodsorrel. To measure the effect an attached pappus may have on seed settling, dogfennel and common groundsel seeds were included with the pappus intact and with the pappus removed. Descriptions of the seeds used in this test are listed in Table 1.
Description of weed propagules used in buoyancy testing. a
a Sources for propagule descriptions and sizes were the Flora of North America https://www.efloras.org/flora_page.aspx?flora_id=1, Neal et al. (Reference Neal, Uva, DiTomaso and DiTommaso2023), and measurements of seeds in the authors’ seed collection.
Ten seeds of a species were placed in a 180-mL cylindrical glass jar (Danyang EnCheng Glass Co.) filled with 120 mL of tap water. Jars were sealed and agitated for 30 s to break surface tension and ensure seed immersion (Egginton and Robbins Reference Egginton and Robbins1920), then placed on a level laboratory bench. The number of seeds that settled to the bottom of the jar was recorded 1 h after agitation and at approximately 24-h intervals for 168 h (7 d). During the study all seeds either floated or settled to the bottom of the jars. Seed buoyancy was calculated as a percentage of unsettled seeds. The experiment was conducted using a randomized complete block design with five replications of each species. The experiment was repeated in November 2021. Changes in percent buoyancy over time were tested using a Tukey HSD test at α = 0.05. There were no changes in percent seed buoyancy after 120 h for any tested species; therefore, we deemed our 168-h observation period to be sufficient. Shi et al (Reference Shi, Li, Zhang and Qiang2020) recorded seed floatation for more than 400 h, but settling for most species stabilized after 120 h. Differences in percent buoyancy were analyzed using a general linear models procedure in SAS software (v. 9.4; SAS Institute, Inc., Cary, NC) using percent buoyancy as the dependent variable and species as the independent variable. Means were separated using the Tukey HSD test at α = 0.05. Species were grouped into buoyancy classes based on these mean separations of percent buoyancy at 168 h. The effect of the presence of a pappus on seed buoyancy was tested using single-degree-of-freedom comparisons.
Seed Submergence and Viability
This experiment was conducted at four container nurseries in eastern North Carolina over two consecutive years. The study locations were near Raleigh (35.79161°N, 78.69783°W), Holly Springs (35.61105°N, 78.85265°W), Garner (35.70547°N, 78.55251°W), and Willow Springs, NC (35.60629°N, 78.68402°W). Seeds of spotted spurge, eclipta, flexuous bittercress, and yellow woodsorrel were submerged in irrigation ponds in November 2019, July 2020, November 2020, and July 2021. Hereafter, November submergence timings are referred to as Winter 1 and Winter 2, and July timings as Summer 1 and Summer 2. With the exception of eclipta, the same seed lot was used for all submersion deployments. In the first deployment (Winter 1) germination of eclipta was lower than expected based on the authors’ experience with this species. Therefore, a different eclipta seed lot with higher germination was used for the three subsequent deployments.
For each species at each submersion deployment (Winter 2 and 2; Summer 1 and 2), 200 seeds were placed in separate 150-µm mesh nylon bags. One bag of each species was tied together to create a bundle. Nine seed bundles were placed inside a perforated polyvinyl chloride (PVC) cylinder with a capped bottom and a threaded PVC lid to protect seeds from predation. The cylinder was large enough to allow water movement through the perforations. One cylinder was deployed at each location and kept at a submergence depth of 60 cm, the recommended depth of irrigation intakes, using a buoy and weighted anchor.
One seed bundle containing 200 seeds of each species was collected from each location 7, 15, 21, 30, 60, 90, 120, 240, and 360 d after submergence (DAS). Due to COVID-19 travel restrictions, the 120 DAS Winter 1 collection occurred 135 DAS. This 15-d deviation was considered negligible and data were analyzed and presented with other 120-d submergence data. Bundles were transported to the laboratory, where seed bags for each species were separated. Bags were rinsed for 60 s in tap water, surface-sterilized with 0.6% NaOCl plus one drop of dishwashing liquid (Dawn Original, The Procter & Gamble Company, Cincinnati, OH) for 30 s, then rinsed for 30 s with tap water to remove NaOCl. From each bag, four subsamples of 50 seeds each were placed in petri dishes double-lined with filter paper saturated with distilled water. Dishes were placed in germination chambers with 14-h photoperiods. Additional distilled water was added as needed to keep the filter paper moist for the duration of the germination period. Temperature regimes were optimized for species: 35/25 C day/night for spotted spurge and eclipta, and 23/13 C day/night for flexuous bittercress and yellow woodsorrel (Altom et al. Reference Altom and Murray1996; Andersen Reference Andersen1968; Asgarpour et al. Reference Asgarpour, Ghorbani, Khajeh-Hosseini, Mohammadvand and Chauhan2015; Kimata Reference Kimata1983). Higher temperatures in the chambers were synchronous with the 14-h photoperiod. To isolate submergence effects from natural declines in viability over time, and to ensure adequate conditions for germination, seeds stored dry at 3 C were added to each germination test to serve as a control. Germination was recorded every 5 d for 30 d after placement in germination chambers. Seeds were recorded as germinated when the radical, hypocotyl, and cotyledons reached 1 mm in length, and seedlings were removed after counting. Percent germination was calculated for each petri dish. The viability of seeds that did not germinate was not tested. The authors acknowledge that some of those seeds may have been viable but in a deeper dormancy that prevented germination. However, for the objectives of the current study, the number of germinable seeds was of greater relevance. Seeds with long-term dormancy that may be spread by irrigation would have little potential effect on overall weed populations in container nurseries where crops are grown in soilless substrates and are typically sold within 12 to 18 mo.
For data analyses, control seed germination was used as a covariate. The resulting adjusted LSMEANS for percent germination of submerged seeds were used for all subsequent analyses and are referred to as percent germination hereafter. Seed germination data were subjected to analysis of variance using a mixed model (GLMMIX procedure) and linear regression models (REML – TYPE 3) in SAS software. Year, season, species, and DAS were treated as fixed effects. Because pond replication within locations was not possible, the four 50-seed petri dish subsamples treated as replicates for testing for differences among locations. The resulting error term used to test location effects was ‘year*season*DAS*replicate’. For all other tests the four petri dish subsamples were averaged. Due to the high number of parameters within the error terms, a p-value of < 0.01 was used to indicate significance for all main effects and interactions. Given the differences in germination between the two eclipta seed lots, these data were analyzed separately to assess the effects of DAS and location. Because year and season were confounded with seed lot, only DAS and location were tested as main effects, and with a smaller number of parameters in the error term a value of P < 0.05 was used to indicate significance. The effect of DAS on percent germination of each species and deployment were plotted and best-fit regression lines were generated using Microsoft Excel.
Results and Discussion
Weed Seed Buoyancy
Seed buoyancy among species differed significantly. After 24 h in water, eclipta, marsh yellowcress, and both common groundsel and dogfennel with the pappi attached exhibited ≥89% buoyancy (Table 2). Seeds of dogfennel without a pappus, rice flatsedge and American burnweed were 64% to 78% buoyant. Common groundsel without pappus, large crabgrass, and spotted spurge seeds were between 36% and 41% buoyant. The least buoyant species were flexuous bittercress, fringed willowherb without the coma, livid amaranth, redroot pigweed, and yellow woodsorrel, ranging from 1% to 16% seed buoyancy (Table 2). Data for redroot pigweed are consistent with prior reports indicating seeds sink within 15 min (Egginton and Robbins Reference Egginton and Robbins1920) or within 12 h to 48 h (Shi et al Reference Shi, Li, Zhang and Qiang2020). There was little change in percent buoyancy between 24 h and 72 h, for all species except American burnweed, for which buoyancy declined from 74% to 10% during the same time interval (Table 2). Only minor changes in seed buoyancy of individual species were observed between 72 h and 120 h, and no significant changes for any species occurred between 120 h and 168 h.
Percent buoyancy of weed seeds following placement in water for up to 168 h and assigned buoyancy categorical ranking. a
a Species are listed in descending order by % buoyancy at 168 h. Analysis of variance indicated no difference between repetitions of the experiment; data presented are averages of the two repetitions. Means within a column followed by the same letter were not significantly different based on a Tukey HSD means separation procedure with α = 0.05.
b Buoyancy categories based on a Tukey HSD means separations of % buoyancy at 168 h and modeled after categories described by Shi et al (Reference Shi, Li, Zhang and Qiang2020).
Species were categorized as nonbuoyant, semibuoyant, buoyant, or super-buoyant based on means separations of percent buoyancy at 168 h. Nine species had less than 10% seed buoyancy and were categorized as nonbuoyant: American burnweed, common groundsel without a pappus, flexuous bittercress, fringed willowherb, large crabgrass, livid amaranth, redroot pigweed, spotted spurge, and yellow woodsorrel (Table 2). Yellow woodsorrel was the only species to exhibit 0% buoyancy at 168 h. Dogfennel without a pappus and rice flatsedge were semibuoyant with 47% and 54% buoyancy, respectively. Dogfennel and common groundsel seeds with pappi attached were 72% and 82% buoyant, respectively, and categorized as buoyant. Eclipta and marsh yellowcress seeds were 100% and 99% buoyant, respectively, and were categorized as super-buoyant (Table 2). These results align with those reported by Shi et al. (Reference Shi, Li, Zhang and Qiang2020) that eclipta seeds maintained >50% buoyancy for more than 400 h.
Shi et al. (Reference Shi, Li, Zhang and Qiang2020) reported that species found in wetland habitats often have highly buoyant seeds. The most buoyant species in our test, eclipta and marsh yellowcress, are common in wetlands (Neal et al. Reference Neal, Uva, DiTomaso and DiTommaso2023). These species were recovered from irrigation water filtrates at container nurseries and seed numbers were higher after it rained (Ray et al. Reference Ray, LeBude, Altland, Harlow and Neal2026). Although nonbuoyant seeds are generally considered poorly adapted for hydrochory (Andersson et al. Reference Andersson, Nilsson and Johansson2000), they may still be transported during in run-off water during heavy rains (Xiong and Nilsson Reference Xiong and Nilsson1997). Under such conditions, species with low buoyancy, such as yellow woodsorrel and spotted spurge, may be deposited into irrigation ponds.
The presence of a pappus increased the seed buoyancy of both dogfennel and common groundsel (Table 3). A significant difference between experiment repetitions was observed for dogfennel; therefore, data are presented separately by repetition. Data for common groundsel were pooled across repetitions (Table 3). Buoyancy of common groundsel seeds with pappi attached was 89% after 24 h in water and 82% at 168 h. In contrast, the buoyancy of seeds without pappi was 36% and 8% after 24 h and 168 h, respectively. Similarly, the pappus increased dogfennel seed buoyancy. The buoyancy of dogfennel seeds with and without the pappi averaged 72% and 47%, respectively, after 168 h in water. This anatomical difference was sufficient to shift common groundsel seeds from the nonbuoyant to the buoyant category, and dogfennel from being categorized as semibuoyant to buoyant (Table 2). These data are consistent with prior reports by Egginton and Robbins (Reference Egginton and Robbins1920) and Fryirs and Carthley (Reference Fryirs and Carthley2022), who noted that specialized seed appendages like a pappus facilitated hydrochory.
Effects of pappus removal on percent seed buoyancy of dogfennel and common groundsel 168 h after dispersion in water. a
a Repetition was significant for dogfennel, thus means for each repetition are presented. There was no significant difference between repetitions for common groundsel; therefore, the average of the two repetitions is presented for that species.
Nevertheless, the effect of a pappus may be transient. For example, dandelion (Taraxacum officinale Weber in Wiggers) seeds were highly buoyant for about 1 h but buoyancy declined to only 10% by 10 h (Egginton and Robbins Reference Egginton and Robbins1920). Despite this short duration of buoyancy, in that study high numbers of dandelion seeds were present in irrigation canals during seasonal seed shedding. Similarly, black willow (Salix nigra Marshall) seeds were present in irrigation water filtrates at container nurseries only during spring samples when willow trees were shedding seeds (Ray et al. Reference Ray, LeBude, Altland, Harlow and Neal2026). Most wind-dispersed seeds are deposited within a few meters of the parent plants. For example, the majority of horseweed (Erigeron canadensis L.) seeds were deposited within 6 m of the seed source (Regehr and Bazzaz Reference Regehr and Bazzaz1979). Managing weed populations in the immediate vicinity of irrigation ponds and drainage ditches is recommended to reduce the risk of weed seed spread via irrigation or wind.
Seed Submergence and Viability
The germination of control (nonsubmerged) seeds decreased over time. Control seeds of flexuous bittercress and spotted spurge displayed a slight, linear loss of germination over time (Figure 1A). The rate of decline in seed germination for yellow woodsorrel best fit a log function, with a more rapid decline up to about 90 d and a more gradual rate of decline thereafter (Figure 1A). Eclipta seed lot 1, used for the Winter 1 deployment, had overall low germination rates with a rapid decline in percent germination to about 90 DAS and a gradual decline thereafter (Figure 1B). Eclipta seed lot 2, used for all other deployments, averaged 93.7% germination throughout the study, with no decline in germination over time (Figure 1B).
Percent germination of control seeds for each species following 7, 15, 21, 30, 60, 90, 120, 240 and 360 d in cold storage. A) flexuous bittercress, spotted spurge, and yellow woodsorrel, and B) eclipta, seed lots 1 and 2. The regression lines are for the averages of all tested control seeds, except eclipta, for which separate lines are presented for the two eclipta seed lots. All functional forms tested for seed lot 2 were not significant; therefore, we present the statistical average (Ȳ) for that seed lot. DAS indicates days after submergence.
The main effect of species, and interactions between species and all other variables were highly significant (P < 0.0001); therefore, data were analyzed and are presented by species. Although differences among locations were statistically significant, the trends in the data were consistent and the differences were inconsequential in relation to the primary objective of determining whether weed seeds remain viable long enough in water to be spread via irrigation. Therefore, to test for differences between seasons, years, and days after submergence, the error terms for locations were pooled.
Germination of submerged seeds for all tested species declined over time but the rates of decline differed. Seeds of all species germinated 240 DAS in both years, both seasons, and at all locations (Figure 2A–D). Eclipta, flexuous bittercress, and spotted spurge maintained some germination for 360 d, supporting prior reports that seeds of many weeds germinated following submergence in water (Bruns Reference Bruns1965; Bruns and Rasmussen Reference Bruns and Rasmussen1958).
Percent germination for each species following 7, 15, 21, 30, 60, 90, 120, 240, and 360 d after submergence (DAS). A) flexuous bittercress, B) yellow woodsorrel, C) eclipta, and D) spotted spurge. SUM 1, SUM 2, WIN 1, and WIN 2 are abbreviations for the four seed deployments in summer year 1, summer year 2, winter year 1, and winter year 2, respectively. The regression lines are for the averages of least squares means for all deployments except for eclipta for which there was no significant change in percent germination over time for seed lot 1 used in WIN 1. The regression line for seed lot 2, used in three deployments, is presented.
Flexuous Bittercress. Flexuous bittercress maintained greater than 80% germination following 360 d of submergence in three of four locations for both seasons and both years. Decline in germination was gradual and linear in both years and both seasons (Figure 2A). There was no significant seasonal effect (Table 4), indicating that seeds entering irrigation ponds at any time of year will maintain high germination percentages for a year or more.
Probability values for F-tests of main effects and interactions for percent germination of weeds submerged in irrigation ponds.a–d
a Abbreviations: DAS, days after submergence; NS, not significant.
b Seeds were deployed in winter and summer over 2 yr at four locations. Seeds were removed 7, 15, 21, 30, 60, 90, 120, 240, and 360 DAS and their ability to germinate was tested.
c Where NS is indicated, comparisons with P-values greater than 0.01 were deemed NS because the number of variables in the model contributed to very high degrees of freedom for tests.
d Combined analysis for all deployments with both seed lots. See Table 5 for separate analysis of variance for the two eclipta seed lots.
Yellow Woodsorrel. Germination of yellow woodsorrel declined rapidly between 7 and 120 DAS, after which time the rate of decline was more gradual (Figure 2B). There were significant differences among seasons (Table 4), with seeds submerged in the winter maintaining greater than 50% germination at 90 DAS, whereas summer-deployed seeds exhibiting <20% germination. Yellow woodsorrel seeds may have greater survival in water during cool seasons or colder climates. Regardless of season, seed germination declined to less than 35% by 240 DAS and less than 15% by 360 DAS.
Eclipta. Season and location effects were highly significant (Table 4), but the overall model could not account for different responses between seed lots. When analyzed separately, percent germination of seeds for seed lot 1, used in Winter 1, did not decline over time (Table 5; Figure 2C). For seed lot 2, there were no significant differences between deployments (Table 5) and data were combined for presentation in Figure 2C. While there were differences among the four locations (Table 5), the overall trends in responses over days after submergence were similar. For that seed lot there was a small but significant linear decline in percent germination from 99% at 7 DAS to ≥89% at 360 DAS (Table 5; Figure 2C). The data indicate that eclipta seeds can persist in irrigation ponds for more than a year and maintain high germination rates.
Probability values from F-tests of main effects and interactions for percent germination of two seed lots of eclipta submerged in irrigation ponds at four locations.a–c
a Abbreviation: DAS, days after submergence.
b Seeds were removed from ponds 7, 15, 21, 30, 60, 90, 120, 240, and 360 DAS and their germination was then tested.
c Probability values were obtained from analysis of variance F-tests. Seed lot 1 was used in only in the first winter deployment; therefore, only main effects of location and DAS could be tested. Seed lot 2 was used in the second winter and both summer deployments, which provided sufficient degrees of freedom to test for two-way interactions.
Spotted Spurge. There were significant differences in the rate of decline in spotted spurge seed germination among seed deployments, as well as significant year by season interaction (Table 4). In Winter 1, percent seed germination increased from 50% to 80% between 7 and 120 DAS and then declined (Figure 2D). In contrast, the percent germination of seeds in Summer 1 declined rapidly, and percent germination for Winter 2 and Summer 2 declined more gradually over the same time frame (Figure 2D). Spotted spurge seed germination 90 DAS was greater than 50% in three of four deployments. Despite variable initial rates of decline, by 240 DAS the percent germination was similar for all deployments. After 360 DAS, seed germination was between 15% and 34% for all deployments. Similarly, Comes et al. (Reference Comes, Bruns and Kelley1978) reported that spotted spurge seeds exhibited 26% germination following 9 mo of submergence, but germination declined thereafter. Conversely, Asgarpour et al. (Reference Asgarpour, Ghorbani, Khajeh-Hosseini, Mohammadvand and Chauhan2015) reported that germination declined rapidly after submergence, with no germination after 63 d. The reasons for differences between studies has not been examined.
Implications for Dispersal
While the length of time weed seeds remained germinable following submergence in irrigation water was variable, all four weed species studied maintained 50% or greater germination for 3 mo or more. Thus, seeds of weeds that reach nursery irrigation ponds should have sufficient viability to serve as inoculum for weed infestations in container nursery crops. Previous research suggests that seed buoyancy increases the likelihood that weed seeds will be distributed by hydrochory (Boedeltje et al. Reference Boedeltje, Bakker, Brinke, Van Groenendael and Soesbergen2004; Vogt et al. Reference Vogt, Rasran and Jensen2006). The current study demonstrates that eight of 13 species tested had nonbuoyant seeds. With the pappi attached seeds of common groundsel and dogfennel were buoyant, rice flatsedge and dogfennel without the pappi were semibuoyant, and seeds of eclipta and marsh yellowcress were super-buoyant. Ray et al. (Reference Ray, LeBude, Altland, Harlow and Neal2026) reported a high correlation between eclipta seed numbers in irrigation filtrates and eclipta populations in container nursery crops. This correlation was driven by very high numbers of eclipta seeds in irrigation filtrates at a single nursery where high populations of eclipta were present in crops and drainage paths. Additionally, the number of seeds in irrigation samples increased significantly when it rained, from about 187 eclipta seeds ha−1 before the rain to 3,399 seeds ha−1 after the rain stopped. In that irrigation filtration, eclipta seeds accounted for 97% of all seeds collected. The effect of rain was short-lived, with seed numbers returning to pre-rain levels almost immediately after the rain stopped. It is possible that water turbulence caused by the force of rain drops on the pond surface or by increased runoff could have mixed floating seeds into the water column for a short time. Avoiding irrigation during rain would decrease the spread of highly buoyant seeds via irrigation.
Flexuous bittercress, spotted spurge, and yellow woodsorrel seeds were nonbuoyant. Based on this we would expect seeds of those species to settle rapidly in irrigation ponds resulting in very few seeds available for uptake by irrigation. However, all three species were recovered from irrigation filtrates at numbers similar to buoyant species, dogfennel, rice flatsedge, and marsh yellowcress (Ray et al. Reference Ray, LeBude, Altland, Harlow and Neal2026). In that study, the average number of seeds in irrigation water filtrates for nonbuoyant species, American burnweed, flexuous bittercress, large crabgrass, livid amaranth, and yellow woodsorrel, was between 2.5 and 3.5 seeds ha−1 d−1. Seeds of semibuoyant weeds, dogfennel and rice flatsedge, were present at 3.0 and 2.9 ha−1 d−1, respectively. Highly buoyant species, eclipta and marsh yellowcress, were present at 9.1 and 1.9 seeds ha−1 d−1, respectively. These results suggest that buoyancy is not required for hydrochory in source water used for irrigation in container nursery operations. Ray et al. (Reference Ray, LeBude, Altland, Harlow and Neal2026) concluded that the low number of seeds in daily irrigation was likely to be inconsequential compared with the potential number of seeds that are spread by existing weeds in the nursery, but irrigation could be a vector for introduction of new weed species to newly potted crops.
Controlling weeds in crops and drainage areas before they go to seed can reduce weed seed movement to irrigation ponds. In particular, species with highly buoyant seeds, such as eclipta and marsh yellowcress, have higher potential for hydrochorous spread and should be controlled before mature plants produce seeds. Diverting drainage ditches to settling ponds before water transfer to irrigation ponds could reduce the incidence of nonbuoyant seeds in irrigation water (Yazdi et al Reference Yazdi, Owen, Lyon and White2021). However, settling ponds are unlikely to reduce the presence of buoyant seeds. In well-designed irrigation systems, water intakes are suspended below the surface to minimize the intake of floating debris, including weed seeds. Water intakes are also recommended to be placed as far from the point where runoff enters the pond to increase hydraulic resistance time and minimize contaminants in irrigation water (Yadzi et al. Reference Yazdi, Owen, Lyon and White2021). To reduce the intake and spread of super-buoyant species like eclipta and marsh yellowcress, growers should avoid irrigation during rain or other water disturbances to reduce the spread of buoyant species.
Practical Implications
Weed seeds remain viable in irrigation ponds long enough to be disseminated by irrigation, but seed buoyancy affects the numbers of seeds likely to be disseminated via irrigation. Species with highly buoyant seeds, such as eclipta and marsh yellowcress, appear to have higher potential for movement to irrigation ponds and dissemination via irrigation. Other species with low buoyancy seeds, such as flexuous bittercress, spotted spurge, and oxalis, survived for months in irrigation ponds but likely settle to the bottom soon after entering the pond. Despite rapid settling, these species have been collected in irrigation water filtrates. Reducing the transport of weed seeds into irrigation ponds would decrease the number of weed seeds suspended in the water and reduce the potential for weed seed spread via irrigation. Adopting weed management practices that control weeds along irrigation pond edges and run-off pathways, especially those species with high buoyancy and survivability in water, would reduce the number of seeds transported and deposited into surface irrigation ponds. Additionally, drainage areas designed with long flow paths, settling ponds, or other interventions before reaching the irrigation pond could reduce the presence of nonbuoyant weed seeds.
Acknowledgments
We thank Consuelo Arellano, Professor Emeritus, Department of Statistics, North Carolina State University, for providing advice and assistance with statistical analysis.
Funding
Funding for this research was provided via cooperative agreement 58-5082-8-026 with the U.S. Department of Agriculture–Agricultural Research Service, and in connection with Hatch projects NC-02707 and NC-02807, which are funded by the U.S. Department of Agriculture–National Institute for Food and Agriculture.
Competing Interests
The authors declare they have no competing interests.
Open access
