Developing a Decision Support Tool for Ventenata IPM in the Inland Northwest

Study Region

We conducted laboratory and field experiments using V. dubia populations from a 300 km latitudinal gradient that encompassed parts of eastern Washington, northeastern Oregon, and northern Idaho (Figure 1). note: Tables and Figures are in attached Survey Results document. Though V. dubia is present throughout a greater extent of the PNW, producer surveys (unpublished data) and road surveys indicate that V. dubia invasion impacts have been most significant in the study region. The study region includes the intermediate (30-45 cm) and high (45-60 cm) precipitation zones of the Columbia Plateau that are dominated by dryland annual cropping. At a regional scale, these annual croplands interface with forested ecosystems to the northeast and canyon grassland ecosystems to the southwest (Figure 1). At a landscape scale, variation in topography and seasonal precipitation patterns contribute to a matrix of land use types in which perennial grass systems are embedded.

In Vitro Germination Trials

In order to inform our proceeding field studies on V. dubia life-history patterns, we replicated and extended previous research conducted in 1985 by Patton, Northam, and Callihan which determined the effects of after-ripening, cold stratification, seed-aging, and temperature on V. dubia populations located in the Palouse region.

Effect of After-ripening, Cold Stratification and Seed-Aging on Germination

In the fall of 2009, we initiated studies to evaluate the effects of cold stratification and seed-aging using seed collected from a natural infestation near Pullman, WA in mid-July. We initiated a study in 2013 to evaluate after-ripening requirements using seed collected on July 9 from the same area as the 2009 collections. In each case, seed were immediately separated from plants using a hammer mill and air screen seed cleaning machine, separated into 100-count batches, and stored at 21 °C. Germination of seed that was dry-stored for 0, 30, 60, and 120 days after harvest was evaluated in 2013 to characterize after-ripening requirements. To determine the effects of cold-stratification, we compared seed that was dry-stored for 4 months and exposed to 5 °C for three durations (0, 5, or 10 days). Germination of seed dry-stored for 4, 18, 30, and 42 months was evaluated to characterize the influence of seed-aging on seed-viability and germination.

Each germination trial used similar methodology. Experimental treatments were replicated four times and arranged in a randomized complete block design, using 100 viable seeds per replicate. Seeds were placed in a 150 mm diameter petri-dish lined with Anchor® seed paper, which was kept moist with 1% bleach (6% sodium hypochlorite) solution. Seed viability was evaluated using a forcep test. Seeds that collapsed under gentle pressure were discarded to avoid use of non-germinable seed in experiments. Germination was evaluated for 70 days in growth chambers set at 18°C and 14 hours of daylight (Patton et al. 1985), and seed germination (> 2 mm radicle) was recorded every 2 to 4 days. Germinated seed were removed from the petri-dish at the time of recording.

Effect of Temperature on Germination

In November 2010, we initiated a study to evaluate germination response of V. dubia seed, which was collected and dry-stored for 4 months, to a range of constant temperatures (5.0 to 29.2 °C). The study was conducted on V. dubia seed from populations located in five different ecosystems (forest, pasture, hay, Conservation Reserve Program (CRP) land, and rangeland) in the Intermountain PNW. In order to complete the study, three experiments were performed sequentially on a temperature gradient bar (Holt and Orcutt 1996). Each experiment included five test temperatures (EX1: 5.0, 7.8, 8.6, 9.4, 11.3 °C, EX2: 11.8, 12.2, 13.3, 15.8, 17.0 °C, EX3: 21.5, 23.3, 25.6, 27.3, 29.2 °C) that were monitored with insulated thermocouples. The experimental unit consisted of 25 V. dubia seeds placed in 40 mm petri-dishes that were filled with moistened filter paper and covered with plastic lids sealed with parafilm. Experimental units were replicated four times per temperature treatment. Test populations (n = 5) were arranged using a randomized complete block design at each test temperature across the gradient bar. Germination (> 2 mm radical visible) was recorded and seeds were removed every 2 to 4 days. Seed mortality within petri-dishes that resulted from pathogenic fungi was also recorded and similarly removed from petri-dishes at the time of recording; seeds lost to pathogenic mortality were considered to be non-germinable. Each experiment was terminated at 42 days.

Statistical Analyses

For each germination experiment, cumulative germination (%) was plotted against time (d) and nonlinear regression models were fitted to the data and analyzed using a model where g is cumulative germination expressed as a percentage, d is time in number of days , C is total cumulative germination expressed as a percentage, and b₀ and b₁ are estimated parameters, with b₀ estimating mean germination time (d to 50% germination of germinable seed) and b₁ as the rate (seeds d^-1) at which germination occurs. To determine if germination rates varied across a given set of treatments, we contrasted full and reduced models using the F_R test statistic from the residual sums of squares reduction test in PROC NLIN (Version 9.3, SAS Institute, Cary, NC). Total cumulative germination was specified in the model for temperature experiments because seed viability was not quantified across populations, which would confound curve fitting comparisons if parameterized. For temperature experiments, total cumulative germination was analyzed separately by fitting general linear mixed models in SAS (Version 9.3, SAS Institute, Cary, NC) to examine population main effects by temperature with block as a random factor. Total cumulative germination was transformed using arcsine square-root to achieve homogeneity of variance.

Seed Bank Persistence

In October 2009, we initiated a study to evaluate the effects of burial duration and depth on V. dubia seed bank persistence. The study was conducted on a south-facing slope with Palouse and Tucannon silt loam soil at the NRCS Pullman Plant Materials Center (PPMC) near Pullman, WA, and on a west-facing slope with Spokane loam soil on Paradise Ridge south of Moscow, ID. Both sites were infested with V. dubia and receive approximately 51 cm of annual precipitation. V. dubia seeds collected in mid-July near Pullman, WA were separated in 100-count batches and after-ripened for approximately 3 months. Seed batches were placed in 10 by 10 cm nylon mesh fabric packets containing 10 cm³ of sifted soil from the PPMC study site and buried on October 16 at the Paradise Ridge site and October 20 at the PPMC site. Burial depth (2 and 8 cm) and burial duration (1, 13, 25, 37, and 49 months) treatments were arranged in a randomized complete block design with four replications. Replicates were located in natural V. dubia infestation patches within approximately 0.015 ha of each other.

Seed packets were exhumed and seeds with visible radicles (> 2 mm) were counted as germinated. At the 1 month burial interval, non-germinated seed were subjected to a growth-chamber germination assay to determine the germinable fraction of the seed bank. Seeds were surface-sterilized using a 5% bleach solution (0.25% sodium hypochlorite) and placed on petri-dishes (150 mm) lined with Anchor® seed germination paper in a growth chamber set at 18 °C with 12 hours of fluorescent light. Germination was recorded every 3 days for 60 days. At subsequent burial intervals (13, 25, 37, and 49 mo), non-germinated seed were indiscernible from soil due to decay, eliminating the need for germination assays. Seed bank fate for each burial duration and depth was categorized as 1) germinated: seeds with visible radicles, 2) persistent: non-germinated seeds that germinated in the growth chamber, and 3) lost: seeds not found due to decay or non-germinated seed that did not germinate in the growth chamber.

The effect of seed burial depth on seed bank fate (germination, persistence, and loss) one month after burial and seed bank persistence in subsequent years were analyzed by fitting general linear mixed models to examine the main effect of burial depth by burial duration with block and site as a random factor. Seed bank fate data were arcsine square root transformed to achieve homogeneity of variance. Treatment means were separated using Fishers LSD (P < 0.05). Results are presented using untransformed treatment means.

Effects of Litter on Seedling Emergence

We conducted a garden experiment at the NRCS-Pullman Plant Materials Center (PPMC) in 2011-2012 and replicated it in 2012-2013 to evaluate the effect of V. dubia litter on V. dubia seedling emergence from an artificial seed bank. In August 2011, V. dubia plants were collected from a local population and were hand threshed. In addition, foliar cover (%) was visually estimated and V. dubia biomass (g) was collected in a random sample of 40 plots (400 cm²).   Linear regression was performed to estimate V. dubia litter fresh weight (g) that correlated to 33, 66, and 100% cover. Experimental treatments consisted of 100 V. dubia seeds sown on the soil surface and four litter treatments corresponding to 0, 33, 66, and 100% V. dubia cover (0, 98, 196, 392 g m^-2). In 2011, the experimental unit consisted of 0.05 m² pot. Tilled soil was hand-sifted in the field and used to fill the pots to within 5 cm of the pot edge, allowing for placement of litter within the pot and to protect litter from being windblown. Due to poor drainage within the pots following snow melt in late winter, wooden plot frames were used in replacement of pots in the 2012-2013 experiment.

The experiment was conducted with a 4 × 4 factorial treatment design arranged in a randomized complete block with four replications. Treatment levels included four litter levels (0, 98, 196, 392 g m^-2) and four harvest dates. Three harvests were scheduled approximately two weeks apart following germination in the fall growing season, and one harvest was scheduled for the spring growing season. Separate harvest dates allowed for destructive sampling of pots or frames to quantify seedling densities below the litter layer. Soil moisture (% VWC) and soil temperature (°C) at a 2 cm depth (5TM Soil Probe; Decagon Devices®) were monitored for 0, 98, and 392 g m^-2 litter treatments during both years of the study in an extra litter treatment replicate that was located between experimental blocks. In the 2011-2012 experiment, final densities in several 196 and 392 g m^-2 litter treatments exceeded 100 plants, raising concerns about seed contamination from the threshed litter. In 2012-2013, we included check plots with no sown seeds for each litter treatment. Litter treatments were rescaled (23, 44, 64% for 98, 196, 392 g m^-2, respectively) in both years to provide estimates of litter treatment effects for comparisons to the litter control. V. dubia seedling density was quantified at each fall harvest date. Plant density, tillers per plant, and per capita seed production were quantified at the spring harvest date.

The effect of litter level (g m^-2) on seedling emergence was analyzed by fitting general linear mixed models to examine the main effect of litter level by harvest date and year with block as a random factor. The effect of litter levels on plant density, tillers per plant, seed production were subjected to analysis of variance for 2012-2013 only due poor drainage in 2011-2012. Data were natural log transformed to achieve homogeneity of variance. Treatment means were separated using Fishers LSD (P < 0.05). Results are presented using untransformed treatment means.

Seedling Emergence and Phenological Development Patterns

We monitored V. dubia seedling emergence and phenological development in field plots through the 2011-2012 and 2012-2013 growing seasons at six sites within the Intermountain PNW (Table 1; Figure 1). Four sites were replicated across years. Two sites were relocated in the second growing season due to resource constraints (Table 1). In each year, however, two sites were located in timothy-hay, CRP, and rangeland. We established monitoring plots at each site within natural infestations of V. dubia. Ten plots (400 cm²) were permanently marked at 0.5 m intervals along a 6 m transect. A weather station was also established on the transect and a data logger (EM50; Decagon Devices®) was used to take hourly measurements of precipitation (ECH2O; Decagon Devices®) and ambient air temperature 1.33 m above the soil surface (ECH2O-ECT; decagon Devices®). Soil moisture (%vwc) and soil temperature (°C) were recorded with two 5TM soil moisture probes (Decagon Devices®) that were installed at a 2.5 cm soil depth and positioned upslope and downslope of the transect. Data were collected hourly and periodically downloaded at the site to be entered into a database and quality control checked using the nearest regional weather station prior to use for data analysis.

Field plots were established each year by early- to mid-September prior to the first fall precipitation and visited weekly to bi-weekly during the fall growing season and approximately bi-weekly during the spring growing season. V. dubia seedling emergence was quantified and then plants were removed with glyphosate at a 5% v/v spray solution in spray bottle. The most advanced growth stage of emerged V. dubia plants was recorded using the methods of Moore et al. (1991) in quadrats (400 cm²) positioned on the opposite side of the transect in relation to permanent plots used for monitoring seedling emergence. Moore et al. (1991) describe a comprehensive system for quantifying the phenological development of perennial forage and range grasses using a continuous numerical index which can be used to develop quantitative relationships. This system concentrates on transitions from five primary growth stages of individual grass shoots (germination, vegetative, elongation, reproductive, and seed ripening) and can be utilized for monitoring annual grass phenology. We confined our growth stage observations from the fall vegetative stage of V. dubia seedlings to the hard dough seed ripening stage.

We calculated cumulated growing degree days for each site and year using a model where daily maximum and minimum temperatures (°C) were measured at a 2.5 cm soil depth averaged over the two soil probes, and base temperature was 7 °C. The base temperature was derived from our experiments on temperature (5 to 29 °C ) on V. dubia germination described in detail in the proceeding results section; germination of V. dubia seed did not occur below a test temperature of 8.6 °C. Preliminary germination trials did not identify an upper temperature threshold. Consequently, an upper threshold of downy brome (30 °C) was utilized.

Cumulative seedling emergence (%) was plotted against cumulative growing degree days (GDD) and non-linear models were fitted to the data for each site by year and analyzed using a nonlinear regression model where E is relative cumulative germination expressed as a percentage, GDD is cumulative growing degree days [2] , C is a constant value of 1, and b₀ and b₁ are estimated parameters, with b₀ estimating mean seedling emergence time (d to 50% of total seedling emergence per year) and b₁ estimating the relative slope around the inflection point. Cumulative germination was numerically transformed to a scale of 0 to 100% by dividing cumulative germination by the total number of seedlings emerged over the entire growing season for a given population. As a result, cumulative germination percentages were adjusted to a common scale for comparison of seedling emergence rates among populations. Calculation of cumulative GDD began after measured soil volume water content (%VWC) was greater than the estimated permanent wilting point (PWP) for silt loam soils (11%) using the relationship between soil %VWC, water potential, and soil texture described in Saxton and Rawls (2006). Seedling emergence patterns among perennial grass systems (hay, CRP, rangeland), pooled across sites within system, were compared using the residual sums of squares reduction test for each year of the study.

We compared the effect of perennial grass system on the cumulative GDDs and Julian days at which stem elongation and anthesis occurred. Transitions to stem elongation and anthesis stages were identified as the cumulative GDD and Julian days at the mid-point between bi-weekly sampling periods at which the given transition occurred. Cumulative GDDs were started on the date at which seedling emergence first occurred in the fall for each adjacent seedling emergence plot. Stem elongation and anthesis transitions were subjected to analysis of variance using a mixed model with study system, pooled across sites, as a main effect and year as a random effect. Data were natural log transformed to achieve homogeneity of variance. Treatment means were separated using Fishers LSD (P < 0.05). Results are presented using untransformed treatment means.

IPM for Ventenata

Timothy Hay

Field experiments were conducted from 2012 to 2013 on four study sites in northeastern Washington and north central Idaho. Field sites were selected upon similar soil characteristics, aspect, and management objectives to enhance trial replication. All sites have been in agricultural production for at least five years. Both hay sites were near Cusick, WA (477539 N 5348457 E); the north timothy site was 0.34 km from the south timothy site. Soils at these sites were comprised of Cusick silty clay loam series which were moderately deep and poorly drained. Slopes were 0 to 3%. Average annual precipitation was 69 cm (Deer Park station; 40 km from field site). The timothy hay sites were predominately timothy (Phleum pratense L.), ventenata, meadow foxtail (Alopecurus pratensis L.), Kentucky bluegrass (Poa pratensis L.), and Canada thistle [Cirsium arvense (L.) Scop.].

CRP

The two CRP sites were located near Troy, ID. The south CRP site (5175159 N 520960 E) was approximately 4.3 km from the north CRP site (5178763 N 523363 E). Those sites were primarily comprised of Taney and Southwick silt loam series, which are moderately deep and well drained. Slopes ranged from 0 to 25%. Average annual precipitation was 59 cm (Moscow station; 20 km from field sites). Frequent plant species within these sites include ventenata, orchardgrass (Dactylis glomerata L.), Japanese brome (Bromus japonicas Thunb. Ex Murr.), meadow foxtail, and autumn willowherb (Epilobium brachycarpum), all of which are 75% of the total plant composition.

Experimental Design

The experimental design was a randomized complete block split-plot design. The perennial vegetation removal or nutrient addition treatments (mowing, burning, and fertilizing) were applied to the whole plot and the ventenata removal treatment (herbicide) was applied randomly as the split-plot within the whole plot. Each block was placed within each field where soil and plant communities were similar (Gotelli and Ellison 2004). We evaluated treatments within two infestation levels low (<25% foliar cover) and high (> 50% foliar cover), hereafter referred to low and high respectively. Each set of treatments was replicated three times within the field and repeated at another site in each system. Duration of the experiment spanned the fall of 2012 through the spring of 2013, with data collected before treatment in the summer of 2012 and data collected after treatments were applied at the peak of forage production during the summer of 2013.

Foliar cover and biomass were the two variables used to evaluate ventenata control and perennial vegetation response to treatments. All data were collected along permanent transectswithin the center of each treatment. Percent foliar cover estimates were determined by using the line-point intercept method (Herrick et al. 2005). Plants were recorded by species at each point along the transect and biomass samples were collected within a 25 cm by 50 cm frame (0.125 m²) for each treatment.

Biomass samples were split into two equal parts. One part was then sorted into ventenata biomass and forage biomass (all desirable vegetation) and the other half of the sample was left combined. All biomass samples were then oven-dried at 60 C for 72 hours and subsequently weighed with a digital scale to the nearest hundredth gram. The mass of the combined samples were weighted and applied to the ventenata and foliage samples to result in a final mass.

Treatments

Timothy Hay

Our treatments included 5 cm harvest height, 10 cm harvest height, and 5 cm harvest height with light post-harvest grazing (Fransen 2005). We applied the following treatments to each of those management strategies: fertilize only, fall-applied herbicide only, fertilize with herbicide, and a control treatment (Table 2) within the two infestation levels of ventenata. The differing timothy harvest heights allowed us to contrast hay production at the different heights in conjunction with fertilization and herbicide application. Fransen (2005) suggests that harvesting timothy at a minimum 10 cm harvest height will increase the plant’s ability to better compete through increased storage capability of carbohydrates. Each plot measured 4.9 m by 6 m. Foliar cover estimates were based off of one meter increments and one biomass sample was collected within each plot.

We used the herbicides flufenacet plus metribuzin1 (Axiom® DF) which may soon be labeled for ventenata control in timothy. We applied flufenacet plus metribuzin to the timothy plots at a rate of 0.58 L ai ha-1, based on prior research (Wallace and Prather 2010). The fertilizer amendments were applied as a split application in the fall and spring using recommendations provided by Shewmaker and Bohle (2010) and Mahler (2005a and 2005b). Soil samples were taken to a depth of 30 cm two weeks prior to fertilizer applications to determine recommended rates (Table 1). The selected fertilizer analysis (46-62-45) was applied in the form of a dry granular with phosphorus3 and potassium4 applied in the fall. Nitrogen5 was applied as a split application to the timothy sites as 11.3 kg in the fall and 11.3 kg in the spring. The fertilize treatment was to help promote perennial vegetation carbohydrate storage and increase next season growth by increasing plant competitiveness (Fransen 2005).

CRP

Treatments were selected from currently approved cost-share mid-contract management treatments outlined by the NRCS and the FSA (NRCS 2009; NRCS 2013). The following treatments were selected: fall burn, spring burn, sickle mow and remove, rotary mow, fertilize, and herbicide (Table 2). These treatments are common management techniques employed in north central Idaho. Each plot measured 5 m by 5 m. Foliar cover estimates were based off of half meter increments along the permanent transect. Two biomass samples were collected within each plot on alternate sides of the transect; collected one meter away from the transect.

Fall herbicide applications included the use of sulfosulfuron2 (Outrider®) 21.3 g ai ha-1 on the CRP sites. Previously we found that sulfosulfuron can achieve up to 100% control of ventenata nine months after treatment at a rate of 21.3 g ai ha-1. Fertilizer amendments were based on the same analyses used in the timothy experiments. Phosphorus and potassium rates were calculated using the 30 cm soil sample results. The CRP trials received 11.3 kg nitrogen in the fall but not in the spring.  Prescribed burning, whether fall or spring, can restore and rejuvenate native ecosystem processes within grasslands. Prescribed burning within this project was used to achieve three goals 1) remove leaf litter, 2) stimulate perennial grasses and 3) provide nutrients to the soil (DiTomaso 2000; Masters and Sheley 2001). Fall burning may have the additional impact of increasing herbicide to soil contact and subjecting any emerged ventenata seedlings to frost injury. Whereas, spring burning was used to remove all vegetation in spring which would subject seedlings to frost injury. Sickle mow and remove was used to reduce standing vegetation and remove most of the litter, which would allow for greater herbicide to soil contact. The rotary mow treatment left vegetation within the plot which could increase litter and promote ventenata survival.