Serotiny facilitates kochia (Bassia scoparia) persistence via aerial seedbanks

Abstract Serotiny results in aerial seedbanks that enable persistence of some plant species by evading decay, predation, or lethal germination in the soil seedbank. Although more common in forestry, this concept has received little focus in weed science. On average, kochia [Bassia scoparia (L.) A.J. Scott] retained 2091 seeds plant−1 (∼18.5% of seeds) in aerial seedbanks among 109 sample sites in spring, 7–8 months after senescence and following subsequent cohort emergence from the short-lived soil seedbank. Therefore, asynchronous seedling recruitment of kochia persisting in simultaneous aerial and soil seedbanks could represent an adaptive bet-hedging strategy for successful establishment in high-stress drought-prone environments.


Introduction
Kochia [Bassia scoparia (L.) A.J. Scott] is a halophytic summerannual tumbleweed that inhabits agroecosystems and ruderal areas in western North America, and can cause substantial crop yield losses (Geddes and Sharpe 2022). Kochia exhibits tolerance to several abiotic stressors, including heat, drought, and salinity (Kumar et al. 2019). High genetic variability, rapid seedbank turnover, prolific seed production (up to 100 000 seeds plant −1 ), tumbleweed seed dispersal, and protogynous flowering--causing obligate proceeded by facultative cross-pollination--result in rapid increase of herbicide resistance traits under recurrent herbicide treatment (Beckie et al. 2016;Kumar et al. 2019;Geddes et al. 2022b). Kochia exhibits resistance to up to four herbicide sites of action, including acetolactate synthase (ALS) inhibitors [Herbicide Resistance Action Committee (HRAC) Group 2], synthetic auxins (HRAC Group 4), photosystem II inhibitors (HRAC Group 5), and the 5-enolpyruvylshikimate-3-phosphate synthase inhibitor glyphosate (HRAC Group 9). Recent surveys in the Canadian prairies found that all kochia populations sampled post-harvest were ALS inhibitor-resistant, while the majority were also glyphosate-resistant (Beckie et al. 2019;Geddes et al. 2022b). Synthetic auxin-resistant kochia, discovered in Canada in 2015, continues to impact prairie farms and is of increasing concern for farmers across western North America (Kumar et al. 2019;Geddes et al. 2022aGeddes et al. , 2022b. The short (1-2 years) longevity of kochia seed in the soil seedbank has been well documented (Schwinghamer and Van Acker 2008;Dille et al. 2017;Beckie et al. 2018), and represents a weak point in the kochia life cycle that can be targeted for effective management (Geddes and Davis 2021). However, despite management strategies aimed at depleting the soil seedbank, kochia populations continue to increase in relative abundance among annual field crops .
The soil seedbank represents an important life stage for the persistence of many annual weed species; however, little focus has been given to the persistence of weed seed in plant canopies (i.e., aerial seedbanks). Serotinous species retain at least part of their seed in an aerial seedbank on the plant, thereby evading decay, predation, or lethal germination often subject to seed in the soil seedbank (Lamont 1991). Serotiny is a well-documented mechanism of seed persistence that aids in successful establishment of several woody perennial tree species, with fire being a common trigger for seed release (i.e., pyriscence) (Lamont and Enright 2000). This adaptive strategy serves as a mechanism for seed storage and release when conditions are suitable for successful germination and establishment. Simultaneous aerial and soil seedbanks also aid in persistence and asynchronous establishment of some halophytic Amaranthaceae species in high-stress drought-prone environments like sand dune ecosystems (Gao et al. 2014;Bhatt et al. 2019).
While rarely discussed with respect to the persistence of weed populations, serotiny too could aid in persistence of weed species that tend to thrive in high-stress environments, such as kochia. The objectives of this study were to determine (i) whether kochia plants retain seed in aerial seedbanks, and if so (ii) quantify the number of viable seeds retained on kochia plants, 7-8 months after senescence and during planting of crops the subsequent spring.

Materials and methods
A randomized-stratified roadside survey was conducted to determine the number of seeds retained on kochia plants and tumbleweeds (i.e., a subset of kochia plants that are sphericalshaped, detached from the soil, and blown by wind) following the winter period in southern Alberta, Canada. The survey was conducted during the first week of May, an estimated 12-13 months after emergence of the sampled plants and 7-8 months after their fall senescence according to previous research (Schwinghamer and Van Acker 2008;Dille et al. 2017;Geddes and Davis 2021). This time frame coincides with the typical time of crop planting, and took place after the majority of subsequent kochia cohorts had emerged from the soil seedbank in the spring (C.M. Geddes, personal observation). The survey was conducted in four counties spanning the mixed grassland and moist mixed grassland ecoregions, where kochia was among the most abundant weed species found midseason in annual field crops . Sites were randomly selected by visually identifying senesced kochia plants or tumbleweeds. The number of sites sampled within each county was stratified based on cultivated area, following the methodology described by Geddes et al. (2022aGeddes et al. ( , 2022b. A single kochia plant was collected at random from each of 109 sites, spanning the counties of Lethbridge (27 sites), Taber (18 sites), Warner (30 sites), and Forty Mile (34 sites) (Fig. 1).
The sampled sites were characterized based on site description: field, field access, roadside ditch, fenceline, shelterbelt, or other (bale yard, irrigation ditch, or oil well site).
The nearest preceding crop was identified as cereal, oilseed, pulse, other, pasture, or not specified. The stem was documented as either attached to or detached from the soil (i.e., a tumbleweed), and the plant status was recorded as either full or partial (missing branches or decapitated from harvest operations). The GPS coordinates of each site were recorded and mapped using QGIS 3.12 (QGIS Geographic Information System, Open Source Geospatial Foundation) (Fig. 1).
All aboveground biomass of each kochia plant was sampled, and plants were bagged individually in cotton sacs, dried at 40 • C, and dry weight (DW) determined. Each plant was threshed by hand, and the cleaned seeds were weighed and subsampled. A 2.0 g seed subsample was used to determine the number of viable seeds retained on each kochia plant following the methods of Geddes and Davis (2021). Each seed subsample was planted (1 mm depth) in a 52 cm × 26 cm × 5 cm greenhouse flat filled with Cornell soil-less potting mixture and fertilized to 756, 958, and 505 mg N P K L −1 mixture. The flats were placed in the greenhouse under 20/18 • C day/night temperature, 16 h photoperiod, and 230 μmol m −2 s −1 light. Each flat was watered daily by soaking from the bottom.
The number of kochia seedlings emerged in each flat was counted and removed weekly for 4 weeks. The total number of seedlings was considered viable seeds in the subsam- Fig. 2. The number and percentage of seeds retained on kochia plants after the winter period in southern Alberta as a function of the site, nearest crop, stem attachment, and plant status. P values indicate statistical significance of each main factor based on square root transformed data. Within subfigures, different letters indicate significant differences based on the Dunn-Sidak means comparison (α = 0.05). The boxplot shows the median seed retention with first and third quartiles, while the tails indicate the smallest and largest values within 1.5 times the interquartile range. Mean seed retention is indicated by ×.
ple, which were used to back-calculate the number of viable seeds retained on each plant. Kochia seed exhibits little to no innate dormancy (Schwinghamer and Van Acker 2008), and therefore subsequent rounds of the grow-out procedure were not warranted. The percentage of seeds produced that were retained on each plant (seed retention %) was estimated using plant biomass DW (g), seed weight [number (#) of seeds g −1 ], and an average harvest index (HI; 0.31) for kochia in this region (Beckie et al. 2016) (eq. 1).
Retention % = # seeds plant −1 biomass DW × 0.31 HI × # seeds g −1 × 100 (1) The number and percentage of seeds retained were analyzed using the lme4 package in R v. 4.2.1 (R Core Team, Vienna, Austria). The Shapiro-Wilk test was used to assess normality, while visual inspection of the residuals and fitted values was used to assess homoscedasticity. A square root transformation was implemented to meet these assumptions. A linear mixed-effects model was used to test differences among the levels of site description, nearest crop, stem attachment, and plant status. Each of these were considered fixed factors, while county was considered a random factor. The emmeans package was used to calculate estimated marginal means and the multcomp package was used for compact letter display based on the Dunn-Sidak means comparison (α = 0.05). The cor.test function was used to test Pearson correlations between the number or percentage of seeds retained and plant biomass DW. The ggplot2 package was used for data visualization.

Results and discussion
Kochia plants retained 2091 ± 297 seeds plant −1 (mean ± SE) or an estimated 18.5 ± 1.9% of total seed production on average, 7-8 months after plant senescence and following emergence of subsequent cohorts from the soil seedbank the following spring (Fig. 2). Seed retention ranged widely among kochia plants from 15 to 24 041 seeds plant −1 or an estimated 0.1% to 86.5% of total seed production. Kochia biomass DW correlated positively with the number, but not the percentage, of seed retained (Pearson R = 0.41, P < 0.001 vs. Pearson R = −0.14, P = 0.140, respectively). Kochia tumbleweeds with detached stems had about half the number of seed retained on average (1150 seeds plant −1 ) compared with plants that had stems attached to the soil (2268 seeds plant −1 ) (P = 0.031) (Fig. 2). This was likely caused by seed dislodging due to the disturbance of tumbling (Beckie et al. 2016) over distances unquantified in our study. Minor differences in kochia seed retention were observed among the nearest preceding crop (P < 0.011) (Fig.  2). Sites where the crop was not specified (either no adjacent crop, or the nearest crop was unidentifiable due to tillage operations) had a lower number (416 seeds plant −1 ) and percentage (5.2%) of seeds retained compared with cereal (2003 seeds plant −1 ) or cereal (19.3%) and oilseed (35.0%) crops, respectively. Site description and plant status were not found to influence kochia seed retention.
The current study documented that serotiny can result in persistence of a large amount of kochia seed in aerial seedbanks over winter. The short-lived (1-2 years) longevity of kochia seed in soil seedbanks has been the focus of several past studies (Schwinghamer and Van Acker 2008;Dille et al. 2017;Beckie et al. 2018). About 50% of kochia seeds succumb to processes leading to mortality (decay, predation, or lethal germination) by May, 7 months after entering the soil seedbank (Beckie et al. 2018). However, kochia seed in the soil seedbank rapidly diminished to >90% mortality by June (8 months after burial). Our research suggests that kochia seed can persist in aerial seedbanks over winter contributing further to the persistence of this species while evading decay, predation, or lethal germination in the soil seedbank.
Kochia seed serotiny likely evolved as a mechanism to facilitate longer distance seed dispersal via prolonged seed retention on tumbleweeds, aiding in spread of kochia populations across the landscape. Beckie et al. (2016) showed that kochia tumbleweeds lose about 90% of their seed within 1 km of travel from their origin, but about 10% of seed were retained beyond this distance. Serotiny could represent also a mechanism to aid in asynchronous recruitment of kochia cohorts from seed persisting in simultaneous soil and aerial seedbanks. A similar adaptation was reported for other halophytic, albeit perennial, Amaranthaceae species like glasswort (Seidlitzia rosmarinus Bunge ex Boiss.) and saltbush (Halothamnus iraqensis Botsch.) (Bhatt et al. 2019). This adaptation could facilitate recruitment of later emerging cohorts, or dislodg-ing of seed from kochia plants through the disturbance of crop seeding operations, thereby seeding kochia with the crop and allowing it to escape non-residual pre-plant burndown herbicides. Gao et al. (2014) documented that simultaneous formation of aerial and soil seedbanks enabled temporal heterogeneity in establishment of the annual Amaranthaceae species sand rice [Agriophyllum squarrosum (L.) Moq.] which inhabits harsh desert sand dune ecosystems. In the current study, we report a similar mechanism aiding in persistence of kochia, another annual Amaranthaceae species of significant implication to semi-arid cropping systems of western North America.
The current study showed that kochia exhibits seed serotiny, which could enable asynchronous recruitment of seedlings from seed persisting in either aerial or soil seedbanks. This could represent an adaptive bet-hedging strategy to aid in successful kochia establishment in high-stress and drought-prone environments where the species tends to thrive. Future research should aim to determine the relative contribution of aerial and soil seedbanks to kochia population persistence, and whether these simultaneous seedbanks result in asynchronous seedling emergence events. Further understanding of the contribution of aerial and soil seedbanks to kochia persistence will aid in the development of management strategies targeting this problematic tumbleweed, thereby helping to curb its unfettered spread within and among farmlands.