| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Oregon State Univ., Corvallis, OR 97331
b Virginia Tech, Blacksburg, VA 24061
c Usda-Ars Asfrc, Beaver, Wv 25813
d Virginia Tech, Southern Piedmont Agric. Res. and Ext. Cent., Blackstone, VA 23824
* Corresponding author (jfike{at}vt.edu)
Received for publication December 14, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: PAR, photosynthetically active radiation
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Agroforestry may serve as an alternative approach 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 (Matson et al., 1997; Garrett and McGraw, 2000).
Silvopastoralism is one such agroforestry practice that intentionally integrates trees, forage crops, and livestock into a structural practice of planned interactions (Clason and Sharrow, 2000). Greater forage production, nutritive value, and digestibility are reported for pastures grown under trees (Smith, 1942; Garrett and Kurtz, 1983; Burner and Brauer, 2003) relative to open sites, and this may reflect changes in botanical composition. For example, Brooks (1951) noted reduced species of poor forage merit such as broomsedge (Andropogon virginicus L.) and poverty grass (Danthonian spicata L.) when observing swards under black walnuts vs. open pastures.
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 (Williams et al., 1997). Honey locust is of interest because selected varieties (e.g., ‘Millwood’) can produce high-energy pods that potentially can serve as a valuable source of livestock feed (Wilson, 1991). The pulpy pods contain up to 350 g kg–1 sugar, and yields are similar to an equivalent acreage of oat (Avena sativa L.) (Smith, 1950). Stems of Millwood are also highly palatable; thus, these trees need extra protection from wildlife before maturity: In a pasture setting, new seedlings would likely be grazed out. Phenologically, both species are suited to use in cool-season pastures. Both species produce leaves late in spring; have sparse, open canopies, which allow penetration of some light to the forage understory; and release leaves early in fall.
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; therefore, it is important to test tree impacts on forage production systems on a regional scale before making widespread recommendations to farmers. Our study objective was to measure forage production and botanical composition in response to tree species, tree density, and slope position in an emulated silvopasture (i.e., there was no animal component in the system) in southern Appalachia.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
|
To examine the interactions of slope and shading environment on botanical composition and forage production, sampling sites were located at points across the combination of slope and tree density gradients. At shoulder-, mid-, and toe-slope positions, three 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 per replicate). Sampling site locations were selected based on tree densities that created three shade classes: (i) full to partial shading all day; high shade environment or high tree density; (ii) morning sun exposure with shading events after solar noon; medium shade environment or medium tree density; and (iii) full exposure to sunlight from midmorning; low shade environment or low tree density. The sampling sites, 0.53 by 2.44 m, were placed with the long dimension parallel to the tree rows.
Botanical Composition
Each season, sampling began when average forage canopy height was about 25 cm although some inflorescences had begun to appear. 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.
Botanical composition of each site was determined just before 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 by 0.5 m) were randomly placed within each 1.3-m2 sampling site. Herbage within the quadrats was clipped to 7.5 cm 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 Mass
Herbage within each sampling site was cut to a height of 7.5 cm with a push mower with bag attachment. Immediately after cutting, samples were weighed so that moisture concentration could be determined. Samples were dried at 60°C for 48 h and then weighed for determination of dry matter, and values were used to calculate estimates of yield per unit land area. Sites were harvested in the afternoon (after 1500 h) to limit differences due to diurnal variation. In 2003, sampling sites for Rep 2 and 3 of black walnut at both high and medium tree densities on mid- and shoulder slope were appropriately relocated before the 7 May 2003 harvest due to groundhog (Marmota monax L.) damage and tree death.
Comparison of "forage-only" production (i.e., excluding weeds and dead herbage) among treatments was assessed for those dates at which botanical composition was available. Adjusted forage mass was calculated as dry matter yield minus that portion of yield represented by weeds and dead herbage (including tree leaves) within corresponding harvests.
Photosynthetically active radiation (PAR) was measured in the third replicate at shoulder- and toe-slope positions using LI-COR LI-191-SB line quantum sensors (LI-COR, Lincoln, NE). Sensors were mounted parallel to tree rows about 25 cm above the ground to prevent shading by forages. Data from the sensors were collected using Campbell Scientific 21x (Campbell Scientific, Inc., Logan, UT) data loggers with measurements made every 10 s and averaged hourly. Measurements were made in late summer of both years because honey locust has an indeterminate growth pattern and the canopy continues to develop until that time.
Statistical Analysis
Tree species were arranged in a randomized complete block design. Density and slope treatments were arranged as split block factors, creating a 2 x 3 x 3 randomized complete split-split block design. Data were analyzed using the General Linear Model procedure of SAS (SAS Inst., 2004). Field replicate, slope position, tree species, and tree stand density were the whole, main, sub, and sub-subplots, respectively. All main effects and interactions were tested across both growing seasons, by year, and by harvest date. Means were separated by LSD, and treatments were considered significant at P < 0.05. Where appropriate, trends (P < 0.10) are reported.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Trees in the study were 10 yr old and approaching full canopy closure in the closest row spacings by the end of 2003. Average height across density and slope treatments was 3.2 m for walnut and 5.1 m for locust trees.
Botanical Composition
Botanical composition values are reported across seasons within years due to the high variability between the two growing seasons (Table 1). For example, legumes were negligible in 2002 but abundant in 2003.
|
0.2) due to tree species were observed for other forages or weeds. In 2003, grass fractions were not affected (P 0.2) by tree species, but the legume fraction was greater (P = 0.01) under honey locust than under black walnut trees. Plots under honey locust also had a lower percentage of (P < 0.05) dead herbage.
Numeric (nonsignificant) reductions in percentage tall fescue in the stand with increasing tree density were observed both years. Other cool-season grasses tended (P 
0.10) to make up a greater fraction of the sward under medium-density sites in 2002 and with a numerically greater proportion at both medium- and high-density sites in 2003. Percentages of warm-season grasses did not differ with tree densities in 2002 but were greater (P = 0.01) at low-density sites in 2003.
Weed percentages were similar across densities in the dry 2002 growing season but were much lower (P = 0.03) at low-density sites in 2003. In 2002, the fraction of dead herbage in the plots increased numerically with density and was significantly greater (P < 0.001) at high-tree-density sites in 2003. Over seasons, the percentage of dead herbage increased most with tree density under black walnut trees (species x density interaction; P = 0.001).
Slope did not affect botanical composition in 2002. In 2003, weeds were more common in swards at the midslope and lowest at the shoulder slope. Percentage dead herbage was greatest at toe slopes, with least and intermediate percentages at mid- and shoulder-slope positions.
Forage Mass
In 2002, forage mass from plots under black walnut trees was 22% greater (P = 0.01) than from plots under honey locust trees, but tree species had no effect (P = 0.2) on yield in 2003 (tree species x year interaction; P = 0.02). Forage yields were lower (P < 0.001) in 2002 compared with 2003 (4660 vs. 6260 kg ha–1) across all treatments (Table 2) due to environmental conditions.
|
0.08) to be lower under honey locust at spring harvests (May and June) and were significantly reduced (P 0.03) at summer and fall harvests (July, August, and November). The magnitude of these differences influenced the observed species x date interaction (P < 0.001). In 2003, lower (P 0.01) yields under honey locust were observed later in the season (August and October harvests; species x date interaction; P < 0.001).
Over years, forage mass was 20% greater (P < 0.001) from plots under medium-density trees compared with the average of yields from plots under low- and high-density trees. With the exception of the August 2003 harvest, forage mass under medium-density trees was numerically or significantly greater than forage mass from low- and high-density sites at all harvests. The magnitude of these differences resulted in significant date x density interactions (P 0.04). Yields were similar between high- and low-density treatments at all but the June 2002 and July and August 2003 harvests when yields were lower (P 0.01) under high-density sites (density x date interaction; P = 0.003). Forage mass reductions at high-density sites were observed under honey locust trees in 2002 (species x density interaction; P = 0.01) but not in 2003 (Fig. 4)
.
|
Adjusting forage mass for weeds and dead herbage did not change the relationships between forage mass and field treatments (Table 2). Across all five sampling dates, the coefficient of determination (r2) for adjusted yield regressed on yield was 0.933 with slope approaching unity (0.997). Adjusted forage mass from plots at medium-density sites was greater than for either high- or low-density plots in 2002. In 2003, adjusted forage mass was again lowest at high-density sites but not different between low- and medium-density plots. Slope had no effect on forage production for the times sampled.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Dead herbage was negligible at the spring and summer sampling events each year. This was likely due to the high frequency of defoliation of actively growing herbage. The 35-d cutting frequency allowed little time for senescence of forage plants. Increased dead herbage observed at the September 2003 harvest was largely due to leaves deposited on the canopy surface, especially under black walnut trees. This likely reflects both later leaf drop and slower rates of leaf decomposition for that species. Greater percentage of dead herbage in the sward may have implications for grazing animals given dead forage is an antiquality component relative to intake by grazing animals (L'Huillier et al., 1984; Montossi et al., 2001).
Similar percentages of warm-season grass across densities in 2002 likely reflect the poor germination and growing conditions for those species due to lack of soil moisture. The increased warm-season grass component at low tree densities in 2003 likely reflects more suitable growing conditions for those species. At low-density sites, solar exposure was continuous from or before about 0900 h; thus, these sites were likely very similar to truly open pasture. Limited shading in the more open environment would allow for higher temperatures needed for warm-season grass growth (Gardener et al., 1985). These data agree with other research (Kephart et al., 1992; Lin et al., 1999) showing that reduced irradiance has a greater impact on productivity of warm-season grasses than cool-season grasses. Brooks (1951) reported warm-season grasses such as broomsedge and poverty grass made up less of the sward under black walnut trees compared with adjacent open pastures.
The percentage of forage (i.e., all grasses and legumes) at medium- and high-density sites was similar both years. Apparent reductions in tall fescue from 2002 to 2003 appear to be a function of differences in rainfall between the two growing seasons (Fig. 2). Fescue at these sites is highly endophyte infected, and the endophyte would have provided competitive advantage under the moisture-limiting conditions of 2002. Poor conditions for legumes and less-competitive cool-season grasses, coupled with poor conditions for germination and growth of warm-season grasses, would also support greater percentage of tall fescue in the sward in 2002. It is also possible that the loss of May 2002 samples biased the results in favor of tall fescue in that other cool-season grasses that are more productive earlier in the year may have been underrepresented by loss of the spring 2002 composition data.
Greater percentage of forage components at low tree density sites in 2003 (compared with 2002) was driven by increased amounts of warm-season grasses and a concomitant reduction in weeds. Similar levels of fescue across tree densities are in contrast to a report of open pastures with 50% tall fescue vs. 95% tall fescue under 35-yr-old black walnut trees (Garrett and Kurtz, 1983). However, pastures in that study were newly established, and the trees were 25 yr older than ours.
High levels of weeds were common across densities in 2002, with common blue violet (Viola papilionacea L.) typical to medium- and high-density sites in both years. As with dead herbage, increased weed species may have implications for animal production in silvopastures if animals must spend more time in selecting preferred species.
The increase in forage mass at medium density relative to open- and high-density sites suggests that alteration of the microclimate by trees boosts forage productivity. At the black walnut toe-slope position, total sunny-day PAR for the medium- and heavy-shade sites was 45 and 32% of the low-shade sites (5 mol m–2 d–1). Honey locust, with a less dense canopy, received 70 and 58% of PAR intercepted at low-shade sites. At the shoulder-slope position, black walnut trees were very small, and PAR for the light- and heavy-shade sites was 79 and 82% of the open-site values. Honey locust was similar at the shoulder-slope position compared with toe slope with daily PAR at 63 and 56% of open sites. Less light was available at high-density sites while at low-density sites, soil temperatures were often higher than desirable for cool-season grass production (data not shown).
Competition with trees may have limited the resources available for forage production at high-density sites, but this would not have been the case at low-density sites. Our yield results agree with those of Garrett and Kurtz (1983), who reported 40% yield increases for newly established fescue growing under black walnut in Missouri, USA, compared with open pasture. Yield increases were not as large in our study (about 20%). This smaller response may reflect more moderate climatic conditions at our site, lower sensitivity of mature pastures to more open environments, or differences in response of pure stands (Missouri) vs. mixed pasture (Virginia).
Forage mass was numerically greater at low-density sites in August 2003, the only occurrence over 11 harvest dates. This may be due to an increase in warm-season grasses at sites receiving limited amounts of shade. At low-density sites, warm-season grasses made up about 14 and 12% of the sward when botanical composition was estimated in July and Sept of 2003. This compares with averages of about 6 and 2% warm-season grasses under medium- and high-density sites in July and September 2003, respectively.
Weeds and dead herbage inflated the estimate of forage mass by about 15 to 20% for each set of main effects within seasons. When the relationship between unadjusted and adjusted forage mass was tested with simple linear regression, high coefficients of determination (R2 > 0.90) were obtained for nearly all sets of main effects within seasons (data not shown). These results indicate that weeds and dead fractions (herbage and tree leaves) in the sward created little bias with respect to differences among treatments.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
While both tree species appear compatible with forage production systems in southern Appalachia, honey locust may not be as effective in promoting increased forage production, particularly at sites with lower rainfall. 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.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. L. Buergler, J. H. Fike, J. A. Burger, C. M. Feldhake, J. R. McKenna, and C. D. Teutsch Forage Nutritive Value in an Emulated Silvopasture Agron. J., September 5, 2006; 98(5): 1265 - 1273. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Crop Science | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
