Optimizing Forage Production and Quality Within a temperate Silvopasture System

Final Report for GS03-024

Project Type: Graduate Student
Funds awarded in 2003: $9,959.00
Projected End Date: 12/31/2004
Region: Southern
State: Virginia
Graduate Student:
Expand All

Project Information


Sustainable farms are rare in Appalachia due to severe topography, land use potential, farm size, and lack of diversification. Incorporating trees into pasture production systems may improve the economic viability of small, pasture-based farms by increasing their productivity and diversifying the product base, but their impacts on forage production and nutritive value are not well known.
In a two-year study, we have tested the impacts of trees on pasture performance with respect to forage production and nutritional value. Microclimate measures were also made for interpretation of our results.
In 1995, black walnut and honey locust seedlings were planted within 3 replicate plots of mixed pasture. A total of six plots were arranged in a completely randomized block design. Within each plot, four rows of trees were planted down a 12% slope to create increasing shade gradients both across and up the slope. Fifty-four sampling sites were established on points across the combination of tree density and slope gradients.
Sample sites were harvested about every 35 days starting from May in 2002 and 2003. Samples were dried for yield per land area, and were analyzed for acid and neutral detergent fibers, non-structural carbohydrates, crude protein, and Ca, P, Mg, and K. Microclimate measures included photosynthetically active radiation (PAR), soil temperature, and soil moisture.
Yield increases of 15% were observed with moderately spaced trees both in dry (2002) and wet (2003) growing seasons. Fiber levels were typically decreased with shade, but the levels of difference were sometimes too small to be of biological relevance. Levels of TNC were usually lower with greater tree density, while CP concentrations often increased with shade. Lower CP in forage from open sites would be due to dilution effect. Concentrations of Ca typically increased with high tree density, but other mineral responses to treatments were complex and less consistent.
Reasons for yield and nutritional responses appear strongly linked to changes in forage microclimate. While shade from trees reduces incident light, the reduction in PAR under under medium shade was also accompanied by reduced soil temperatures, creating a better growing environment for the forage crop. Light resources were likely the limiting factor under high density trees, however. At sampling sites under low density trees, soil temperatures appear the limiting factor for forage production as temperatures were often above Toptimum for cool season forages. Soil moisture levels generally were not different across tree density or slope position treatments.
Our efforts show that appropriately spaced trees can benefit forage production in temperate pastures. The results provide an initial basis for selecting tree spacing at which forage production and quality is optimized. Future efforts will be needed to determine impact to animal health and performance, and to environmental and economic impacts.


Producers and researchers continue to strive for higher productivity levels by developing and improving agricultural technology including genetics, machinery, fertilization, and pesticides. However, such strategies may not address the needs for long-term productivity or sustainability of our production systems and in some cases may mask or even contribute to environmental contamination and degradation resulting from agrochemical pollution, soil erosion, pest problems, and loss of biological diversity. In short, the current methods of increasing productivity may come at high environmental cost and may not be sustainable.
Agroforestry may provide alternatives to using capital inputs for increasing production. Agroforestry practices have potential to optimize positive biological interactions between crop components and emphasize species diversity rather than only crop yield.
Silvopastoral practitioners intentionally integrate trees, forage crops, and livestock to reap the benefits of their interactions. Greater forage production, nutritive value, and digestibility are reported for pastures grown under trees relative to open sites, and this may reflect changes in botanical composition.
Potential tree species for the Appalachian region include black walnut and honey locust. Black walnut produces both high value wood and generates an annual nut crop; management for either or both outputs is possible. Honey locust is of interest because selected varieties (e.g. ‘Millwood’) can produce high-energy pods that potentially can serve as a valuable fodder source. The pulpy pods may contain up to 35%, and yields are similar to an equivalent acreage of oats.
Despite the potential benefits of silvopastoral practices, very little research has been conducted in the humid, temperate regions of eastern North America. Design and management of silvopasture systems will vary by location and, it is important to test tree impacts on forage production systems on a regional scale before making widespread recommendations to farmers.

Project Objectives:

Our objectives were to determine the interrelationships among tree species, tree spacing, and slope on yield, botanical composition, and nutritive value of cool-season pasture. Further, we wanted to describe the silvopasture microclimate and its effects on forages within the tree species-spacing-slope complex. This specifically entailed measurements of light intensity and quality, soil temperature, soil moisture, and soil nutrient profile. Third, we wanted to determine the extent to which trees and pasture interact to increase total production, yield, and value from the land resource.


Materials and methods:

This research was conducted at Kentland Farm, just south of Blacksburg, Virginia. Site elevation is approximately 540 m (1772 ft), with annual temperatures ranging from -24 °C to 38 °C (-12 °F to 100 °F) with a mean annual temperature of 10 °C (50 °F). Average annual precipitation is 102 cm (40 inches). Soils, Typic Hapludults, are well-drained on gently to steeply sloping topography.
In 1995, Black walnut (Juglans nigra) and honey locust (Gleditsia triacanthos) trees were planted in existing pastures. Each replicate (n = 3) contained both a black walnut and a honey locust plot with plots arranged in a completely randomized block design. Within each tree plot, four rows of trees were planted down the face of the 12% slope with 1.8, 3.7, 7.3, and 14.6 m (6, 12, 24, and 48 feet) within the rows and 3.7, 7.3, and 14.6 m (12, 24, and 48 feet) between rows. Spacings were designed to create increasing shade gradients both across and up the slope.
Pastures are predominantly tall fescue (Festuca arundinaceae), but contain orchardgrass (Dactylis glomerata) and bluegrass (Poa pratensis) among others.
Sampling sites (n=54) were located on points across the combination of tree density and slope gradients to examine the interactions of species, density, and slope on forage production. At shoulder-, mid- and toe-slope positions, permanent sites were created within low-, medium- and high-shade environments for a total of nine sampling sites within each tree plot (n = 18 sites/replicate). Sampling site locations were selected based on tree densities that created three shade classes: 1) full to partial shading all day, 2) morning sun exposure with shading events after solar noon and 3) full exposure to sunlight. The sampling sites were 0.5 m X 2.4 m (1.75 ft. X 8.0 ft) and ran parallel to the tree rows.
Each season, sampling began when average forage canopy height was about 25 cm. Harvests were scheduled at approximately 35-d intervals. In 2002, harvests occurred 9-10 May, 12 June, 17-20 July, 21 August, and 11-13 November. The planned September 2002 harvest was postponed due to drought. During 2003, sites were harvested on 7 May, 10 June, 16 July, 20 August, 24 September, and 29 October. Plots were harvested after 3 PM on all sampling dates in order to determine potential shade effect on non-structural carbohydrates. Samples were dried at 60 °C (140 °F) for 48 hours, then weighed for calculation of yield per land area.
Botanical composition of each site was determined just prior to harvests on 17 July and 11 November, 2002. An earlier May sampling was lost due to handling error. In 2003, samples were collected on 6 May, 16 July, and 24 September. Two quadrats (0.3 m X 0.45 m; 12 in X 18 in) were randomly placed within each 1.3-m2 (14 ft2) sampling site. Herbage within the quadrats was clipped to 7.5 cm (3 in) and separated into the following components: tall fescue, other cool season grasses, warm season grasses, legumes, broadleaf weeds, and dead herbage (including tree leaves). Separated herbage was dried and weighed and botanical composition was calculated as percentage of the dry matter for each component at a site. Weights of botanical composition components were summed and added to herbage mass values from their respective plots.
Forage samples were analyzed for neutral detergent fiber (NDF), acid detergent fiber (ADF), acid detergent lignin (ADL), crude protein (CP) and total non-structural carbohydrate (TNC) using near infrared reflectance spectroscopy. A subset of samples was analyzed by wet chemistry for calibration. Concentrations of NDF, ADF, and ADL were determined sequentially with an ANKOM fiber analysis system. Forage CP concentrations were determined using Kjeldahl procedures.
In 2002, forage canopy and soil surface temperatures were recorded at 3-h intervals from 0600 h for a 12- (16 July) or 24-h period (15-16 August) at each sampling site. Forage canopy temperatures were measured with an infrared thermometer and soil surface temperatures were measured with a thermocouple thermometer.
No differences in temperature at 0600 were observed in 2002, thus that sampling time was eliminated in 2003. Forage canopy and soil surface temperatures were recorded from April through October in 2003 to more fully characterize microclimate responses over the growing season. Measurement dates were 12 and 27 April, 20 May, 10 June, 8, 11, and 18 July, 29 August, 12 September, and 3 and 30 October 2003.
In addition to thermocouple readings, temperature dataloggers were installed at each sampling site in the top 5cm of soil in 2003. Data loggers measured soil surface temperature every 3 h from 1200 h on 22 May until 1200 h on 25 October.
Volumetric soil moisture to a depth of 15 cm (6 in) was determined by time domain reflectometry. TDR rods were inserted to a at the bottom edge of each sampling site and left in place for the duration of the study. Measurements were taken in June and July of 2002 and about every two weeks for the duration of the season (mid-April through October) in 2003.
Photosynthetically active radiation (PAR) was measured in the third replicate at shoulder- and toe-slope positions using quantum sensors mounted parallel to tree rows about 25 cm above the ground.

Research results and discussion:

Botanical composition was affected by tree species and density, but responses differed in wet and dry years. Pasture under walnut tended to have a lesser proportion of tall fescue in a dry year, in part as a function of greater levels of other cool season grasses under medium density trees. In the subsequent wet year, lower levels of legumes were observed under black walnut. This may indicate an allelopathic effect of that walnut on clover germination and growth.
Percentages of warm season grasses and weeds did not differ with tree densities in the dry year, likely because of the limited moisture through the season. Warm season grasses were greater at low density sites in the subsequent wet growing season, while conversely, weed percentages were much lower. Dead material (herbage and tree leaves) typically was greater with increasing density, especially under walnut, but substantial amounts were only observed in late summer/early fall harvest
Forage yield responses to tree species also differed by season. In a dry year, forage mass from plots under black walnut trees was 22% greater than from plots under honey locust trees. But in a wet year, tree species had no effect on yield
Over the two years of study, forage mass was 15% greater from plots under medium density trees compared with yields from open pasture. The data clearly show that forage yields are benefited with appropriately spaced trees.
Reasons for the forage yield responses appear tied primarily to microclimate changes due to trees. However, clear relationships between microclimate changes and forage yield were not always observed.
Averaged over a season of measures, forage canopy temperatures at low tree density sites were 2.3 °C (4 °F) warmer (27.4 vs. 25.1 °C; 81.3 vs. 77.2 °F) than at medium and high density sites. This difference among sites was twice as large (4.6 °C; 8 °F) for the average of measures taken in July or August of each year [38.5 vs. 33.9 °C (101.3 vs. 93.0 °F) for low vs. medium and high tree densities].
Forage at toe slope positions was cooler during spring and summer. Measures taken in April 2003 indicate mid-slope positions warm up first, but by May, canopy temperatures were similar at mid and shoulder slope positions. Temperature differences due to slope were not apparent in fall measurements.
As with canopies, soil temperatures differed little between tree species but were cooler with increasing levels of shade. However, soil surface temperatures were cooler with low density when night air temperatures were below 15 °C (59.0 °F). This suggests plants under trees grow under more buffered microenvironmental conditions.
Soil moisture levels were usually greater under honey locust trees, likely reflecting the lower levels of forage growth (and hence evapotranspiration) observed between the two tree species. Surprisingly, tree density had little significant effect on soil moisture levels at most sampling dates but was sometimes greater at toe slope positions.
Averaged over days, light levels at the forage canopy did not differ between tree species, but photosynthetically active radiation (PAR) tended to be greater under honey locust trees in early afternoon because of the trees’ sparser canopy.
As anticipated, PAR and tree density were negatively correlated, with a greater difference between low and medium tree density than between medium and high density trees. Light levels also tended to be greater under locusts at toe slope positions, but higher under black walnuts at shoulder slopes. This is because locusts had fairly uniform canopies across the slope while walnuts were much larger toward the bottom of the slope.
Our data indicate that with appropriate spacing, incorporation of black walnut or honey locust trees into pastures can alter botanical composition and forage production in southern Appalachia. Given that forage yields peaked at medium tree densities and that light levels under low tree densities were not dissimilar to open areas, our data suggest that incorporating walnut or locust trees could boost forage production over that of open pastures. Positive yield response is dependent on the maintenance of appropriate tree density, however.
While both tree species appear compatible with forage production systems in southern Appalachia, honey locusts may not be as effective in promoting increased forage production. However, greater amounts of legume and lesser amounts of dead herbage and tree leaves in swards under honey locust trees suggests this species may have benefits for forage nutritive value in mixed pastures.

Participation Summary

Educational & Outreach Activities

Participation Summary

Education/outreach description:

Data from our findings were presented at three national meetings in 2003 and three presentations were made at national meetings in 2004. This project supported a Master’s thesis. A refereed journal article from this work is in print, another has been accepted, and a third is soon to be submitted. In addition, a refereed review article about establishment of silvopastures has been published and was derived in part from the MS thesis.
Presentation and discussion of agroforestry practices was made with extension agents at the Virginia Cooperative Extension In-Service-Training sessions in March of 2004. Several new and challenging ideas were presented.
The silvopasture research site has been and continues to be a stop of interest for professional and producer tours of Virginia Tech’s Kentland Research Farm

Project Outcomes

Project outcomes:

Farmer Adoption

We have recruited several producers to establish tree plantations on farm, but this effort will be delayed until we secure funds for purchase and installation of the trees.


Areas needing additional study

Silvopastoral management techniques need further research into the interactions of trees, forages, and animals. The animal performance component is a key question for these systems. While shade from trees benefited forage yield, it may have mixed effects on nutritive value. A consistent, and potentially negative result, is the reduction of non-structural carbohydrates (NSC), an important energy source, in shaded forages. NSC provide energy to the grazing animal, and energy is typically the most limiting factor for animal production on pasture. Thus, decreases in forage TNC may be detrimental. This may, however, be offset by the additional yield, or greater digestibility of fibers from plants grown under shade. Further, the potential positive effects of shade on animal comfort and grazing behavior need to be determined. Potential allelopathic effects of walnuts on clovers and weeds also warrant exploration.
At a systems level, it is clear that adding trees to pastures can increase forage yield while also generating an additional crop. However, a net present value analysis is needed to model the potential benefits to the system as trees produce goods and services such as nut and fodder crops. Environmental impacts (and credits) should also be considered in such an analysis.
Additional research is needed on producer adoption of these methods. General producer willingness to incorporate trees into pastures and specific hurdles to adoption (such as costs of establishment or need for tree management) need further assessment. Attention should also be given to markets for tree products in order that producers wanting to sell these goods can be sure they have an outlet for them.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.