Assessing pasture grasses, legumes and pasture blends for varying soil conditions in New England and Pennsylvania

Milestones:

The first and second milestones involved project leaders meeting in 2007 at the Northeast Pasture Consortium meeting, in conference calls, and by email to develop protocols for forage evaluations at research sites and farms and to select the blends and mixtures to be tested. Three field experiments were established in the fall of 2007 at the UMass Research Farm, the Vermont Technical College Dairy Farm (UVM site), and a Pennsylvania site (USDA-ARS Pasture Systems and Watershed Mgt Res. Unit).

The three research sites included 25 to 28 pasture blends varying in complexity from two species (a grass and a legume) to seven species with varied varieties in different blend from several commercial seed companies that market pasture blends and varieties to farmers in the New England and northeast region. Farmers and seed industry representatives (third milestone) at the pasture consortium, at extension meetings, and pasture walks gave input to project leaders on specie choices, varieties and mixes/blends. All studies were seeded in the fall of 2007. In 2008 these three studies were rotationally (mob stocked) grazed 4 to 7 times (7 times in MA, 6 times in PA and 4 times in VT) during the grazing season. Data on establishment was collected by project leader teams and cooperators at the three main field experimental sites. All plots were evaluated for establishment and winter survival in spring of 2008. Again in spring of 2009 plots were evaluated for winter survival.

Ten farmers, some indicating a desire to evaluate pasture forages, and others recruited were provided five seed mixes for planting in this first year, and assistance in planting was provided to some as needed. A further distribution of seed was made to 9 farmers who planted test plots (fourth milestone). The distributions by state were 7 in Massachusetts, 8 in Vermont, 2 in New Hampshire, and 1 in each of Connecticut and Maine. The Pennsylvania test site was also a local beef farm. Thus the project included one or more farms in almost all New England states planting three to five selected blends in on-farm trials for a total of 20 farms including Pennsylvania. Reports from farmers on establishment and winter survival in spring of 2008 and 2009 were positive. Farmer sites were evaluated under grazing in 2008 and in 2009, and farmers reported positive animal responses or preferences for many of the newly seeded pasture mixes.

The fifth and sixth milestones in part were met with field tours each year at Massachusetts and Vermont experimental sites, pasture walks and meetings at a participating farms in Connecticut, multiple farms in Massachusetts, and similar in other states. Multiple field tours were hosted at all three main experiment sites, and many farms hosted pasture walks where information and results were provided to attendees in 2008, 2009 and 2010.

All milestones have been more than completed and results of the project are detailed in this final report.

Results and Discussion:

Forage Yield:

Crop establishment in 2007 prior to winter after establishment was monitored and winter survival in spring was good and deemed to be adequate for all mixtures and blends at all three state locations. Total pasture yield (forage plus weeds) varied among the three to four years and among the three state locations (Fig. 1). Yields in 2008 and 2011 were higher in Massachusetts than in 2009 and 2010 while in Pennsylvania the yields in 2009 were higher than in 2008 and 2010. Yields and number of grazings were much lower for the upland Vermont location and each year the yield decreased compared to the previous year. These data also showed that more complex mixtures, those with more species, tended to have higher yield, especially in Massachusetts and Pennsylvania, than those with fewer species (Fig. 2). It was also evident that there was much variation in yield within each number of species group (Fig. 1), and that species in blends that included small amounts of a species in the mixtures (e.g. with less than 10%, of trefoil, timothy, or bluegrass) did not establish sufficient plants to contribute much to forage yield.

The differences in yield among the years was most likely related to better growing conditions 2008 in Massachusetts when seven grazing times were possible before freezing conditions while in 2009 and 2010 there was only six grazing times. Rainfall in 2008 was near normal with good rainfall in June through September while 2009 was wetter than average both in Massachusetts (+ 6 inches May through August) and in the northeast. Pennsylvania was dry in 2008 and had better rainfall in 2009 than the previous dry year in 2008, and was dryer than the excessive 2009 rainfall at the Massachusetts site. Vermont was cooler than Massachusetts and Pennsylvania and rainfall was variable among years as in the other states. The low yields in Vermont (Fig. 1) are a reflection of the cooler conditions but also were probably due to the poorer soil condition (wet, poor drainage) and greater competition from weeds. The low Vermont yield in 2010 was due to just second pre-grazing harvest being collected. The first harvest was lost due to grazing the spring growth before data collection. The third pre-grazing harvest was not taken because the plots were too variable with weed invasion.

At each location for each year the blends and mixtures were ranked based on yield. In the first year (2008) in Massachusetts, the top 15 of 28 blends contained perennial ryegrass as one of the components (Fig. 3). However, in 2009, 2010 and 2011, 6 of the top 7 blends contained orchardgrass as a component. After the winter of 2008/09 among the 2-species mixtures all orchardgrass mixtures yielded more than all perennial ryegrass mixtures. In the spring of 2009 there was noticeable damage to perennial ryegrass in the Massachusetts field plots. The highest yielding 2-specie mixtures contained Extend orchardgrass and one blend (#13) was among the top 1 or 2 each year. Blend #13 was a 6-species commercial blend and included Extend orchardgrass (see Table 1 for a listing of varieties in each blend), and in 2010, the top 5 blends all contained Extend orchardgrass attesting to its superiority as a orchardgrass variety. Figure 4 gives a comparison of Extend orchardgrass in Massachusetts with other orchardgrasses and other blends without orchardgrass. Most of the blends without orchardgrass contained perennial ryegrass, thus figure 4 indicates the superiority of all orchardgrass cultivars over perennial ryegrass in Massachusetts.

In Pennsylvania two mixtures (#’s 12 and 13) were ranked as the top 2 for highest yield and this was consistent over the three years (Fig. 5). While in Massachusetts blends with orchardgrass had greatest yield, in Pennsylvania blends with red clover, perennial ryegrass, orchardgrass, or tall fescue in Pennsylvania and with more than 2 species had greater yields. This reflects perhaps the less severe winter temperatures in Pennsylvania compared to Massachusetts and Vermont where perennial ryegrass suffered some winter damage reducing yield in the second and subsequent production years.

In Vermont plots contained significant amounts of weeds after winter thinning of susceptible species which may have affected the ranking of blends and mixtures (Fig. 6). There was no clear consistency for yield among blends and mixtures and orchardgrass or the mixes with the variety Extend had both high and low rankings each year. Weed content was likely a factor. In 2008 the average weed content was 13.6%, in 2009 34.7% and in 2010 (one harvest) 46.9% with grass weeds being 8 time more abundant than broadleaf weeds. The high and increasing content of weeds lead to abandonment of the experiment midway through 2010. In comparison to Vermont the content of weeds in Massachusetts was negligible (less than 2%) for all plots throughout the four years. Similarly, Pennsylvania had few weeds. Comparative results for forages and weeds for Vermont and Pennsylvania studies in 2009 are shown in Figure 7.

In Massachusetts we made an attempt to determine animal preference to different species and blend complexity by measuring the amount of pasture consumed during grazing periods in 2008 and 2009. When grazing pressure adequate all plots were grazed similarly regardless of the number of species in the pasture blend (Fig. 8). Perennial ryegrass and plots with much white clover were grazed preferentially with the whole plot grazed evenly while bordering plots were sometimes left, depending on species, for grazing later in the grazing period. Tall fescue and orchardgrass blends were not the first to be grazed but were grazed somewhat evenly if animals were restricted to the area long enough. These blends were less grazed if forage supply was over abundant for the animal grazing density.

The distribution of forage production over the season is shown for Massachusetts in Figure 9, Pennsylvania in Figure 10, and Vermont in Figure 11. These are included mainly to show seasonal changes in production which were variable among years within all locations. Variations likely occurred because timing of grazing, rainfall patterns and quantity, and temperature profiles in different years. For example, the Pennsylvania site had a summer slump in production in 2008 (a dry year) and 2010 but the opposite in 2009, a year with more adequate rainfall.

Forage Quality:

This project generated thousands of forage samples that ideally would be analyzed for forage quality. Because funding was not sufficient in the project this has not been completed for all samples. In Massachusetts we were fortunate enough to obtain funding in another project and purchased a NIRS forage analysis system. We have used this to complete the analysis for all samples collected in Massachusetts. The four year average for protein levels varied from a low of 17.6% to a high of 20.8% (Fig. 12). The blends containing perennial ryegrass had higher protein levels than most blends containing orchardgrass. Of the top 14 blends only two contained orchardgrass and all but one contained perennial ryegrass. Of the bottom half only two did not contain orchardgrass.

The NDF levels in blends were almost the reverse of the protein levels (Fig. 12). All 2-specie mixes with Alice white clover and a perennial ryegrass cultivar were the blends with the five lowest NDF levels. This was also true for ADF levels (Fig. 12). Blend 13, which was consistently high yielding in Massachusetts and Pennsylvania, had low protein (4th from the bottom) but highest estimated NDF and ADF values.

Variety/Cultivar Differences among the Pasture Blends:

Massachusetts -- Variety/cultivar performance within species in the pasture plot assessments must be made within the context of the other species within a blend. For example, there were 3 cultivars of alfalfa planted as part of three very different blends; Baralfa42-IQ was planted with Intensiv orchardgrass, Barfleo timothy, and Hakari bromegrass. Maxigraze GT was planted with Aubisque perennial ryegrass, Treasure timothy, Ginger Kentucky bluegrass, and Spring Green festulolium, and FSG 408DP alfalfa was planted with Boost perennial ryegrass, FSG 9601 red clover, Select tall fescue, Big Daddy annual ryegrass, and Summit timothy. All blends were assessed at multiple inventory dates in 2010 and 2011 for cultivar differences. While overall cultivar performances were compared, analysis of results is limited by the unknown effects of the different companion species.

The method used in Massachusetts to determine which species were present in a plot is as follows. A measuring tape was placed through a diagonal of a 3 m x 7 m plot. Over the central 3m of the diagonal, at 30 cm intervals, a pin was dropped at the tape. Every species found within 1 cm of the pin was recorded. If inventory was 3 m, this gave 10 inventory points and a species could get a maximum “score” of 10 if observed at each point. If five species were present, the “score” could theoretically be as high as 50. By allowing for more than one species per point to be counted, shorter species were not excluded, and fewer points were needed to avoid missing present species entirely. It would be expect to have seasonal differences in species. Some do well early in the season, in cool weather; others excel in hotter, drier weather. While the inventory timings and diversity of companion species present created some difficulty, there were clearly cultivars which performed better than others. Table 4 shows results of this assessment analysis.

The analysis for alfalfa shows significant differences of frequency of observation among the three varieties. Maxigraze GT performed best, but this could be because of its inherent superiority or it could be because alfalfa competes well with the rest of the blend of which Maxigraze GT was a part. Other species included in the pasture blends (table 4) were analyzed in a manner similar to that described for alfalfa. Birdsfoot trefoil, bromegrass, and meadow fescue were not included because only one cultivar of each was planted. Plots in with more than one cultivar within a species were not included in these analyses because cultivar identification was not possible. In plots of one planted cultivar within a species, it was assumed that if the species was present, it was the planted cultivar.

Cultivar differences in red clover performance were significant. However, for white clover the analysis may not reflect cultivar differences accurately. For most species, the species did not seem to have moved far from the plot in which it was planted. Thus we are fairly confident that the cultivar observed was the cultivar planted. In the case of white clover, however, all plots contained substantial amounts of white clover, so it is never clear which cultivar was being counted in the inventories or if the white clover observed was from the soil white clover seed bank.

Cultivar differences were not observed for festulolium or tall fescue in the pasture blends. Festulolium, a cross between meadow fescue and annual ryegrass, initially dominated mixtures when present the first year but declined over time as it did other sites. The data collected for timothy indicated three of the cultivars outperformed the other three. Plot to plot differences observed for Kentucky bluegrass made it impossible to confirm any difference in species presence based on either cultivar or time of observation. It appeared that Kentucky bluegrass had increased from 2010 to 2011, but statistical confirmation was not possible.

Orchardgrass was a major component of the pasture in pasture plots being included in 14 of the 28 blends. Over the four years grazed, Extend orchardgrass which was observed least in plots among the orchardgrass had the largest yield. Five blends contained Extend orchardgrass lending confidence to the determination. Perennial ryegrass was also abundant being the most planted species, included in 22 of the 28 blends. There were 12 different cultivar/cultivar combinations in the experiment. Cultivar differences ranged from a high of 88 to a low of 45 indicating significant differences occur within this species.

Vermont -- There was generally a good relationship between species richness and pasture yield in both years when only accounting for seeded grasses and legumes. In 2009, the more complex mixtures had fewer weeds than the 2-specie binary mixtures. By the middle to end of the 2009 season, there appeared to be a negative relationship between species/cultivar richness and the percentage of white clover in the stand. Perennial ryegrass, festulolium and Hakari brome declined rapidly over the first two years probably due to the elevation and latitude of this study and the poorly drained soil type. Some of the festulolium cultivars had severe winter injury particularly ‘Duo’ and ‘Lofa’. ‘Spring Green’ had the least injury and decline; however, it still dropped to about 20% of dry matter composition by the middle of the second harvest year. Amongst the orchardgrass cultivars, ‘Barandana’ appeared to be the most persistent under the conditions of this study, but it even declined over the two year period. ‘Tekapo’, a New Zealand low growing type orchardgrass showed the least persistence. Within the complex mixtures, tall fescue, meadow fescue and timothy appeared to slowly increase or stabilize. Kentucky bluegrass was only in a two of the mixtures and was slow to establish. All plots had grass weeds with the predominant grass weeds in the plots being reed canarygrass and quackgrass.

By the end of the first full harvest year (2008) in Vermont, seeded white clover made up approximately 30% of dry matter in the single grass cultivar plots and remained stable or increase slightly in the second year although most of the seeded grasses declined. Based on the treatments that were not seeded to any white clover, it appears the “baseline” for native white clover was between 8 and 15% by the middle to end of 2009. Red clover and alfalfa never became very prevalent in mixtures where it was included. They appeared to be losing persistence over time, perhaps not surprising considering the poor soil drainage class. Birdsfoot trefoil was slow to establish but showed higher amounts in the second year except toward the end of the grazing season.

On-Farm Pasture Evaluations:

A total of 20 farms agreed to establish with pasture blends. Seed of commercial blends was provided to participating farmers with a recommendation to establish test pastures in the late summer of the year (Table 5). Most seeded in the fall although some seeded in the spring the next year. However, not all farm tests were established. In each farm evaluation three to five pasture blends were selected as a subset of the larger evaluation at the research sites and were managed under grazing. Pasture blends established on farms were mostly planted in strips or larger fields but were not replicated on farms. Hence, data from these studies is not presented as research findings. Participating farmers included mostly beef and dairy with one or more sheep and poultry, and were from every New England state except Rhode Island. Project leaders worked with farmers to collect information on establishment and productivity of the pasture blends.

• Figure 2. Massachusetts annual yield with different number of species.
• Figure 3. MA yields ranked high to low.
• Figure 5. PA yields ranked high to low.
• Figure 7. Forage and weed composition PA and VT.
• Figure 8. Pasture available and consumed.
• Figure 9. MA forage distribution across harvests.
• Figure 10. PA forage distribution across harvests.
• Figure 11. VT forage distribution across harvests.
• Table 4. Observations on differing cultivars of each species.
• Figure 1. Total annual yields for MA, PA, and VT.
• Figure 4. Extend Orhardgrass vs other blends.
• Figure 6. VT yields ranked high to low.
• Figure 12. Pasture forage quality.
• Table 5. Pasture mixes used in on-farm trials.