Cover Crop Grazing: Optimal Seasonality for Soil and Livestock Benefit

View the project final report

• 2016 annual report

Commodities

• Agronomic: millet, peas (field, cowpeas), radish (oilseed, daikon, forage), sorghum (milo), wheat
• Animals: sheep

Practices

• Animal Production: feed/forage, grazing management
• Crop Production: cover crops, fallow, no-till
• Education and Training: extension, on-farm/ranch research
• Production Systems: integrated crop and livestock systems
• Soil Management: soil quality/health

We explored the effects of cool- and warm-season cover crop mixtures, with and without grazing, on soil properties and spring wheat yield in two plot-scale studies in southwestern Montana. Cover crop mixtures were terminated with herbicide (i.e. ‘brown manures’), grazing, or haying, and compared with chem-fallow. Based on equal growing degree days, cool-season biomass production (4.5 Mg/ha) was <10% greater (p = 0.03) than warm-season biomass at Fort Ellis (2015), but at the north Gallatin Valley (NGV) location in 2016, cool-season biomass (2.4 Mg/ha) was more than double (p <0.01) the warm-season biomass. Thus, warm-season cover crops may need to be sown earlier in the growing season (mid-May) in order to optimize growing season rainfall and achieve adequate biomass for grazing, especially in years with abnormally low June rain. Forage quality of the warm-season mixes were generally superior to cool-season cover crop mixes but forage quantity is of paramount interest in most Montana grazing systems.
When measured after termination at both sites, soil under cover crops contained 4-8 mm (0.15-0.3 in) less soil water and 24-30 kg/ha (21-27 lb/ac) less Nitrate-N than chem-fallow (P<0.01). When Fort Ellis was resampled Apr 13, 2016, prior to spring wheat seeding, soil water no longer differed among treatments, but soil Nitrate-N differences remained the same. Although differences in soil water and nitrate were detected between warm-season and cool-season cover crops, the differences were small. Some soil measures showed that retaining cover crop biomass on plots produced more biologically active soil compared with removal of crop biomass via 24 hr of mob-grazing. However, at both sites, acid phosphatase enzyme activity was 29% greater on graze-terminated vs spray-terminated cover crop treatments, averaged across cool- and warm-season crop mixtures when measured 9-10 months later prior to seeding the spring wheat response crop. Enzymatic activity was reduced on the chem fallow control plots at the Fort Ellis site, but not at the site in the northern Gallatin Valley. Cover crop treatments did not affect a spring wheat response crop differently at Fort Ellis, while the fallow control had slightly greater wheat grain yield and protein. Wheat data from the NGV site were not obligated within the time frame of this study but will be reported when those data are processed and analyzed, no later than March 1, 2018.

Objectives/Performance Targets

The purpose of this research was to investigate the effects of integrating livestock grazing in dryland wheat-cover crop systems.

The study objectives were:

Livestock Grazing

Objective 1. To evaluate both cool- and warm-season annual cover crop mixtures for potential forage in dryland wheat-cover crop systems soil and

Objective 2. To investigate how grazing cover crop mixtures as a termination method affects soil biochemical properties in comparison to using herbicide for crop termination.

Agronomic

Objective 3. To assess how changes in soil properties, catalyzed by grazing, affect subsequent wheat yields.

Educational

Objective 4. To increase local producer knowledge of cover crop mixtures performance as forages/graze and termination methods and potential benefits for soil quality.

Objective 5. To provide local producers with information regarding the potential benefits of ICLS and methods of ICLS adoption in Montana through scientific and extension publications.

Materials and Methods Site Characterization

The 2015-16 study site was the Montana State University – Fort Ellis Research Farm (N 45.667°, W 110.978°) located near Bozeman, MT. The site that was planted in barley during the 2014 growing season has been under no-till management since 2012. The predominant soil type at the Fort Ellis Research Farm is Blackmore silt loam and the mean annual rainfall is ~500 mm (19.7 inches). Fort Ellis was planted in cover crops in 2015, followed by a spring wheat response crop in in 2016 (cv. Duclair). In 2016, a second site in northern Gallatin Valley, MT (NGV) was established in partnership with local farmer, Jason Camp. The site was sown to winter wheat during the 2015 season and hayed alfalfa for several years prior. This field site experiences less annual rainfall (~360 mm) with a predominant soil type of Amsterdam silt loam (Tables 1 & 2).

Study Design The experimental design was a randomized complete block consisting of six cover crop/termination treatments (mainplot: 15.2 m x 7.6 m) with a split plot arrangement for response-year N fertizlizer treatments, with four replications. Each cover crop mixture consisted of cereal, legume, and brassica crops. Cereal species were oat (Avena sativa L. cv. Oatana) in the cool-season mixture and sorghum (Sorghum bicolor L. sp.- 2015) and millet (Panicum miliaceum L. sp.- 2016) in the warm-season mixture. Legume species used included forage pea (Pisum sativum L. cv. Arvika) in the cool-season mixture and soybean (Glycine max L. sp.) in the warm-season mixture. Radish (Raphanus sativus L. sp.) was the brassica used in both the cool-and warm-season mixes as its growth pattern is daylength-sensitive (radish bolts if planted prior to mid-June, otherwise remains vegetative and forms large diameter tap root). Cover crops species were seeded in the same row (spacing: 30 cm) to a depth of 2.5 cm with a low-disturbance no-till disk seeder. Seedling rates were calculated by dividing a recommended monoculture rate proportionally by the number of species in the mixture. Each mixture was seeded at a target rate of 120 m-2 with individual cover crop species added to a mixture at a target seed rate of 40 m-2. Termination of cover crop coincided with first-bloom of pea in the cool-season mixtures (McCauley et al. 2012; Table 3). Termination of warm-season cover crop mixtures was determined by the accumulation of growing degree days (GDD; Tbase = 0°C) approximately equal to the cool-season mixture.

Cover crop treatments are described in Table 4.

Chemical-fallow (fallow), served as the control treatment, representing a common grower standard practice in Montana. Lambs (Ovis aries) were used to ‘mob-graze’ the grazed cover crop treatments. However, after grazing, the cool-graze treatment was sprayed within a month of grazing to ensure termination of plant growth and preserve soil moisture and in consideration that re-growth is likely insufficient to economically warrant repeated grazing. The warm-graze and warm-season haying (warm-hay) were not sprayed after grazing/swathing as the first frost in the Bozeman area typically occurs mid-September and terminates re-growth of warm-season cover crop foliage. The last two years incurred unusually late first fall frosts (Table 3). Radish is frost-tolerant and can withstand temperatures as low as -4 °C (NRCS, 2009). Cool-season spraying (cool-spray) and warm-season spraying (warm-spray) were terminated with glyphosate and left as a ‘brown manure’ (Table 3). The cool-spray and warm-spray contrasted cool-graze and warm-graze for plant biomass removal. In the warm-hay, cover crops were swathed and then killed by frost in order to isolate the effect of cover crop biomass removal with livestock absent. However, in year-2, due to very limited biomass growth, warm-hay was swathed and the biomass was left on the plots to simulate mowing, since it would have been very difficult to remove scant cut biomass without excess soil surface disturbance.

In year-2, all main plots were sown to spring wheat (cv. Duclair) with a low-disturbance no-till disk drill to a depth of 2.5 cm, perpendicular to the drill direction used for year-1 cover crop seeding. Nitrogen fertilizer treatments subplots were banded >5 cm below and to the side of the seed furrow at rates of 0, 67.5, and 135 kg N ha-1 (0, 60, and 120 lb n ac-1). Cover Crop Sampling Cover crop population stand counts were conducted 4 wk after planting in four 1-m strips per plot totaling ~1 m2 area. Cover crop count plots were at least 0.5 m from any treatment border to avoid edge effects. Crop management treatments were implemented on the same day, within cool- and warm-season growth regimes, respectively. Above-ground cover crop biomass for all plots was sampled in the same areas used for seedling counts by hand-harvesting plant biomass to the soil surface within 24 hr of termination. Grazed and swathed plots, were resampled similarly within 48 hr after termination to measure post-grazing and post-haying residual biomass. All samples were dried at 50°C prior to obtaining dry weights. For warm-season cover crop treatments, enlarged radish roots below soil surface were dug, cleaned, and measured as biomass. After drying, cover crop samples were ground and analyzed for C and N using a LECO combustion analyzer (LECO Corp., St. Joseph, Michigan). Cover crop forage quality subsamples were taken from all grazed and swathed treatments pre- and post-termination (including cool-season re-growth prior to glyphosate application). Dried samples were sent to Midwest Laboratories (Omaha, NE) and analyzed for fiber, protein, nutrient and mineral content with NIR spectroscopy. Soil Sampling and Laboratory Procedures To test for grazing compaction, soil penetration resistance was measured in 2015 and 2016 with a hand-held static cone penetrometer (Durham Geo-Enterprises Inc.- Nova Metrix, Wakefield, MA). Six measurements were taken from each treatment within a week seeding at relatively uniform soil moisture content (required for accurate comparison). Additionally, soil bulk density was measured prior to seeding in both years, which may be a more reliable measure of soil compaction as it is not so sensitive to soil water content. Site bulk densities, by depth, were determined by averaging all values for each block and used to calculate Equivalent Depth (mm) of soil water to a depth of 0.91 m (Or and Wraith 1999). Soil water was measured prior to planting cover crops, after cover crop termination, and prior to wheat seeding the following spring to determine overwinter attenuation of post-termination soil water differences. Soils were weighed fresh (or frozen) from the field, and after being placed in an oven at 50oC for 72 hr to measure soil water content. Soil nitrate was measured using the same cores as soil water (depth of 0.91). Samples were oven dried at 50°C and sieved through a 5-mm sieve. After sieving, a 5-g subsample of soil was added to 25 mL of 1 M KCL, shaken for 30 min, filtered through a Whatman 5 filter, and then analyzed on a Lachat flow injection analyzer (Lachat Instr., Loveland, Colorado). Soil P was measured by collecting separately six soil cores to a depth of 15 cm in each plot and processing samples by adapting methods from Olsen and Sommers (1982). Samples were oven dried (50°C) and sieved (2-mm). After sieving, a 1.25-g subsample of soil was added to 25 mL of 0.5 M NaHCO3 (pH 8.5), shaken for 30 min, and then filtered through a Whatman 5 glass fiber filter. Then, 5 mL of soil solution was combined with H2SO4 acid as described in Kuo (1996) and analyzed by spectroscopy at 880 nm. Potentially mineralizable nitrogen (PMN) was measured by collecting six soil cores to a 15-cm depth from each plot. The soil probe was sterilized with isopropyl alcohol between plots to prevent sample contamination. PMN is calculated as the difference in plant available N at time zero and after incubation for 14 d at 30 oC (Keeney 1982). Time zero plant available was measured using 5-g oven-dried equivalent soil and the same method noted previously for soil nitrate extraction. To calculate plant available N after incubation, three 5-g oven-dried equivalent, field moist soils samples from each plot were added to 12.5 mL of double-deionized water and held under a flow of N2 gas for five seconds in order to create a nitrogen atmosphere. Next, 12.5 mL of 2 M KCl was added to each sample and incubated in a hot water bath at 30°C for 14 days. After incubation, samples were filtered through a Whatman 5 filter and analyzed on a Lachat flow injection analyzer. The three incubated samples from each plot were averaged prior to statistical analysis. PMN was reported as the amount of NH4 after incubation minus the amount measured prior to incubation. Microbial biomass was calculated from the same soil cores collected for PMN and using Soil Induced Respiration method modified from West and Sparling (1986). A 5-g field-moist soil sample was added to 10 mL of yeast solution and shaken horizontally at 20°C for 4 hr. Headspace CO2 concentration, caused by microbial respiration, were measured after 10 min of shaking (Time 0), after 2 hr of shaking (Time 2), and after 4 hr of shaking (Time 4). Headspace CO2 concentrations were measured using gas chromatography (Varian Inc., Palo Alto, CA). Soil enzyme activity was measured from the same soil cores as PMN and microbial biomass and analyzing samples using methods adapted from Parham and Deng (2000) and Dick (2011). Specific enzymes targeted were: β-glucosidase (EC 3.2.1.21; cleaves cellobiose from cellulose), β-Glucosaminidase (EC 3.1.2.52; cleaves N-acetyl glucosamine from chitin), and acid and alkaline phosphatases (EC 3.1.3.1, EC 3.1.3.2; cleaves phosphates from organic phosphorus compounds). One gram of field-moist soil was inoculated in duplicate with a p-nitrophenol (pNP)-labelled substrate for each enzyme substrates, incubated for 1 hr at 37°C, and then analyzed for pNP using a spectrophotometer (Parham and Deng, 2000). Wheat Sampling At Fort Ellis, spring wheat biomass samples were taken in two 1-m strips from each fertilization subplot at anthesis. Plants were cut at the soil surface, dried at 50°C, and weighed to obtain plant dry matter. At both sites, a plot combine (Wintersteiger, Salt Lake City, UT) was used to harvest wheat from each treatment plot and N fertilization subplot. Yield was determined by grain weight per harvested area. A NIR transmittance machine (Infratec GmbH, Dresden, Germany) was used to analyze wheat for protein and moisture content. Statistical Analyses All statistical analyses were performed with R statistical software (The R Foundation for Statistical Computing, Vienna, Austria; version 3.2.5). Data were examined for normality and homogeneity of variance using residual and Q-Q plots. Linear models built with treatment and block (rep) as independent variables and analyzed with ANOVA were used in all parameter analyses except for wheat parameters. For wheat parameters, a linear model was built with cover crop treatment (previous year), block, and fertility subplot as independent variables and analyzed with ANOVA. Statistical differences were determined using Fisher’s Protected Least Significant Difference (LSD) Test with no p-adjustment and α = 0.05 (package: agricolea). Pre-planned orthogonal contrasts were used to compare combinations of cover crop and fallow treatments.