| 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. Center, Blackstone, VA 23824
* Corresponding author (jfike{at}vt.edu)
Received for publication July 1, 2005.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: ADF, acid detergent fiber • ADL, acid detergent lignin • CP, crude protein • NDF, neutral detergent fiber • PAR, photosynthetically active radiation • TNC, total nonstructural carbodydrate
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Improved nutritive value is one potential benefit that may be realized with silvopastures. Increased nutritive value or digestibility in response to trees and/or shade has been reported for both cool- (Krueger, 1981; Garrett and Kurtz, 1983) and warm-season species (Eriksen and Whitney, 1981). Nutritive value can be influenced by the presence of trees in silvopastures via morphological and physiological adaptations (Eriksen and Whitney, 1981; Allard et al., 1991, Kephart et al., 1992; Kephart and Buxton, 1993; Sharrow, 1999), in addition to changes in botanical composition (Brooks, 1951; Burner and Brauer, 2003) in response to micro-environmental conditions.
Black walnut and honey locust trees have potential for Appalachian agroforestry because of their phenological suitability and potential to generate additional products. Both species leaf out late in spring, maintain sparse, open canopies during summer, and drop leaves early in fall. Black walnut produces both high value wood and generates an annual nut crop and can be managed for either or both outputs (Williams et al., 1997). Selected varieties of honey locust (e.g., ‘Millwood’) are of interest for their potential to produce high-energy pods that might 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 area of oat (Avena sativa L.; Smith, 1950). Millwood stems are also highly palatable, thus these trees need extra protection from wildlife or livestock before maturity; in a pasture setting new seedlings would likely be killed by grazing.
Despite the potential benefits of silvopastoral practices, very little research has been conducted with forages grown under temperate hardwoods. Comparisons of the effects of different deciduous tree species are lacking in North America. Furthermore, silvopastoral research needs to be conducted in regionally appropriate field studies to account for differences in climate and differences in tree and forage production among other factors. Our objective for this study was to evaluate nutritive value of cool-season pastures in response to tree species, tree densities, and slope positions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
In 1995, black walnut and honey locust trees were planted in existing cool-season pastures. Three replicates each contained both a black walnut and a honey locust plot, and each tree plot contained four rows of trees. The four tree rows were planted in parallel and down the face of the 12% slope. Tree rows had a southwest to northeast orientation. For each tree plot, the distances between rows and between trees within rows were successively decreased from east to west, creating a tree density gradient both across and up the slope. Within each plot, the four rows of trees were spaced 14.6, 7.3, and 3.7 m apart; trees within rows were spaced 14.6, 7.3, 3.7, and 1.8 m apart.
Pastures were predominantly tall fescue, but contained orchardgrass (Dactylis glomerata L.), bluegrass (Poa pratensis L.), and clovers (Trifolium spp.) among other forages. Buergler et al. (2005) reported differences in botanical composition due to season and treatment, with decreased fescue and increased clover observed under conditions of high rainfall. After tree establishment, pastures were maintained by clipping two or three times per season and broadcast applications of maintenance-level fertilization (N or complete fertilizer, <25 kg of nutrient) each fall. In October 2001, pastures were fertilized with 39 kg N ha–1 (as urea). In October 2002, a blend of N, P and K was applied at rates of 45, 78, and 22 kg ha–1, respectively, in addition to 3.4 Mg ha–1 of lime according to soil test.
Field Layout and Sampling Procedures
Sampling sites (n = 54) were located at points across the combination of slope and tree density factors to determine the influence on forage nutritive value. Nine permanent sites were created across the combination of low-, medium- and high-shade environments at shoulder- (top, nearly level), mid- (on the slope) and toe-slope (base of the slope, nearly level) positions within each tree plot (n = 18 sites per replicate). Sampling-site locations were selected based on tree densities designed to create 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 mid-morning, low shade environment or low tree density. The sampling sites, 0.53 by 2.44 m, were placed long-side parallel and approximately centered between the tree rows.
Soils were sampled in August 2002 to characterize pH and nutrient concentrations of the top 5 cm of the soil profile. Three, 2.5-cm wide cores were taken randomly from each site, mixed, dried, and ground. Soil Ca, P, Mg, and K analyses were conducted at the Virginia Tech Soil Testing Laboratory. Soil N and C were determined using a PerkinElmer CNS analyzer (Norwalk, CT).
Soil temperatures were determined at 3-h intervals from 0600 to 1800 h before harvest in July 2002 and again in August. In 2003, temperatures were recorded from 0900 to 1800 h from April through July. On the same dates that soil temperatures were recorded, soil moisture was measured at 1800 h by time domain reflectometry.
Photosynthetically active radiation data were collected in late summer of each year. Measurements were taken in the third block at all densities but only at shoulder- and toe-slope positions using LI-COR LI-191-SB line quantum sensors (LI-COR, Lincoln, NE) mounted parallel to tree rows. Under high-density tree plantings, sensors were placed in the middle of the first two tree rows (1.8 m from either row). For readings under medium-density tree plantings, sensors were placed under the drip line, 1.8 m from the second row of trees. Sensors at low-density sites were midway between trees in the 14.6-m-spaced tree row. Sensors were placed about 25 cm above the ground to prevent shading by forages. Sensor data was collected with Campbell Scientific 21X (Campbell Scientific, Logan, UT) data loggers.
Measurements of PAR were collected every 10 s and averaged hourly. These data also were screened to minimize confounding due to cloud cover within and among weeklong sampling periods. For each hour, the highest average value for open-site PAR within the sampling period was determined, and PAR measures from all densities at these time points were analyzed.
Plot herbage was cut and removed in September 2001 and November 2002, before the start of the study in the following season. Forage sampling began each season when average forage canopy was about 25 cm tall (7–9 May). Sampling sites were cut to 7.5-cm-stubble height with a push mower with bag attachment, and harvests began after 1400 h to minimize effects of diurnal variation (Burner and Belesky, 2004). After harvest, entire mower-strip samples were dried at 60°C for 48 h and ground to pass a 1-mm screen with a hammer mill before analysis.
Fibers, Lignin, Nonstructural Carbohydrate, and Crude Protein Analyses
Neutral detergent fiber, ADF, ADL, CP, and TNC were determined by near infrared reflectance spectroscopy (NIRS, Foss NIRSystem 6500M, Silver Spring, MD). Samples were scanned with near infrared radiation from 1100 to 2500 nm, and log (1/reflectance) was recorded. A stepwise multiple regression equation was generated for each forage constituent using the program SUBSET. Optimum equations were selected based on low standard errors of calibration and validation, and large coefficients of determination (r2) for calibration and performance. These were derived by regressing predicted data against actual data using a subset of forage samples. Samples for calibration subsets for each assay were selected by WIN ISI Winscan software version 1.5 (Infrasoft International LLC, Port Matilda, PA). Prediction equations for each analyte were based on all harvests over both growing seasons.
For the calibration set, a subset of samples was analyzed by wet chemistry. Concentrations of NDF, ADF, and ADL were determined sequentially with an ANKOM fiber analysis system (ANKOM Technology, Macedon, NY). Samples were analyzed in duplicate. A 3% coefficient of variation between samples was the critical limit for repeating the analysis. Forage CP concentrations were determined at the Virginia Tech Forage Testing Laboratory using Kjeldahl 2400 with a Foss Tecator (AN 300, AN 3001, Sweden). Procedures of Dension et al. (1990) were used for determination of TNC.
Mineral Analyses
For determination of Ca, P, Mg, and K, ground forage samples were prepared by drying to constant weight at 60°C for 24 h. A 0.5-g subsample was then weighed into ignition tubes and ashed at 500°C in a muffle furnace for 24 h. Ash was dissolved in 10 mL of 6 M HCl, vortexed, and allowed to sit for 1 h before dilution to 50-mL final volume with distilled water. Samples were then refrigerated in scintillation vials before determination by atomic emission with an inductively coupled plasma spectrometer at the Virginia Tech Soil Testing Laboratory.
Statistical Analysis
The two tree species were arranged in a randomized complete block design. Slope and density treatments were arranged as split block factors, creating a 3 x 2 x 3 randomized complete split-split block design. Data were analyzed as repeated measures in time using the General Linear Model procedure of SAS (SAS Institute, 2004). Replicate and slope position were main factors in the whole plot, with tree species and tree stand density as sub-, and sub-sub-plots, respectively. Years were analyzed separately due to differences in weather and botanical composition (Buergler et al., 2005). Harvests were treated as fixed effects and all main effects and their interactions were tested across and within harvest dates within years (Table 1). Means were separated by LSD, and treatments were considered significant at P < 0.05. Where appropriate, trends (P < 0.10) are reported but LSD is not noted.
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
By the end of 2003, trees in the study were 10 yr old and approaching full canopy closure within the closest row spacings. Walnut trees had greater diameter at breast height at toe-slope (5.6 cm) than at mid- (3.5 cm) or shoulder-slope (2.9 cm) positions, but this measure was unaffected by slope in locust trees (mean = 5.6 cm) (species x slope interaction, P < 0.01). Walnut trees were shorter (P < 0.01) than locust trees (5.1 vs. 3.2 m) and both species tended (P = 0.07) to be taller at toe- (4.7 m) than at mid- (3.9 m) and shoulder-slope positions (3.8 m).
At the black walnut toe-slope position, 2-yr mean total PAR (determined during periods of highest insolation) was 49.3, 21.8, and 15.7 mol m–2 d–1 for the low-, medium-, and high-tree density sites. Honey locust canopies were less dense than walnut at toe slopes and received PAR of 52.9, 37.8, and 31.3 mol m–2 d–1. At the shoulder-slope position, mean PAR values for plots under low, medium, and high black walnut tree densities were 49.4, 38.6, and 39.9 mol m–2 d–1, indicative of smaller, more uniform tree size on the upslope. Honey locust trees were less affected by slope position, and PAR values declined with density similar to reductions at toe slopes (53.1, 33.3, and 29.3 mol m–2 d–1).
Soil temperatures at low-density sites were often greater than 24°C, the upper critical optimum temperature for some cool-season forages (Sprague, 1943). In 2002, the mean temperatures (measured at 3-h intervals from 900 to 1800 h in July and August) from low- to high-tree density were 31.5, 29.8, and 29.3°C. In cooler 2003, data were collected daily at 3-h intervals from 23 May through 31 July. During this period few temperature readings above 24°C were observed at 900 or 2100 h. From 1200 to 1800 h (representing 210 readings over the 70-d measurement period), 149, 124, and 103 hourly measures of soil surface temperature were greater than 24°C at low, medium- and high-density sites. During the 70-d period, mean temperatures across these three measurement times (1200, 1500, and 1800 h) were 27.6, 25.5, 24.4°C. Thus, shade from trees likely modified temperatures to the benefit of cool-season grasses.
Soil moisture measures during June and July 2002 averaged 169, 137, and 136 g kg–1 soil for toe-, mid-, and shoulder-slope positions. In 2003, measures were taken April through July, and differences among slope positions were smaller (296, 279, and 287 g kg–1 soil) for these respective slope positions given the season's ample precipitation. Soil moisture under black walnuts was lower than under honey locust both in dry 2002 (133 and 155 g kg–1 soil) and wet 2003 (274 vs. 301 g kg–1 soil). Tree density appeared to have the least effect on soil moisture in either year, though values were generally greater at low-density sites (153, 142, and 147 g kg–1 soil in 2002; and 293, 282, and 287 g kg–1 soil in 2003). Although patterns were consistent across sampling times, significant treatment differences were rarely observed.
Neutral Detergent Fiber
In 2002, NDF concentrations changed little at toe-slope positions while those at mid- and shoulder-slope positions were greater in June 2002 (slope x date interaction, P < 0.05, Tables 1 and 2). Slope effects (P < 0.05) were observed again in May 2003 and for the mean of the 2003 harvests. Toe-slope positions appeared buffered from environmental extremes, in part as a function of cooler soil temperatures. Trees were smaller at mid- and shoulder slopes, allowing more radiant energy the reach the forage or soil surface, and greater soil temperatures were thus observed.
|
Density x date interactions (P < 0.05) for NDF were observed each season (Table 1), primarily due to date effects in 2002. Concentrations of NDF tended (P < 0.10) to be lower at high tree density sites over the 2003 season, driven primarily by the large reduction (P < 0.01) with density at the July 2003 harvest (Table 2). The small reductions in forage NDF with increasing tree densities are similar to changes reported for fescue grown under reduced irradiance (Kephart and Buxton, 1993; Lin et al., 2001). However, it is not clear if this effect is strictly from reduced light or whether reduced temperature also plays a role. Elevated ambient temperatures increase forage fiber concentrations in growing plants (Fales, 1986), and trees affected both temperature and available light in the growing environment.
Acid Detergent Fiber
Forage ADF concentrations were much greater at mid- and shoulder-slope positions than at toe-slope positions in June 2002 (slope x date interaction, P < 0.01) (Tables 1 and 3). Slope effects (P < 0.05) were observed each year (Table 1). Both in 2002 and in 2003, ADF concentrations were about 15 g kg–1 lower (P < 0.05) at toe-slope positions (Table 3).
|
No response to species was observed in 2003 (Table 1).
Tree density effects on ADF varied by date (density x date interaction, P < 0.05) (Table 1), were less than 10 g kg–1, and were driven by slight changes in the magnitude of the response. Limited ADF response to tree density appears to agree with results of Lin et al. (2001), who reported that ADF levels were unaffected or slightly increased in potted forages grown under shade.
Acid Detergent Lignin
Slope effects on ADL were obscured by a complex three-way interaction (P < 0.01) in 2002 and two-way interactions (P < 0.05) in 2003 (Table 1). In May 2002, ADL concentrations were similar between species at toe- and mid-slope positions but were greater for locusts at shoulder-slope positions (Table 4). For locusts, ADL concentrations increased with each successive harvest and were similar by slope position in 2002. In contrast, ADL concentrations under walnut were greatest in June and intermediate in July 2002, and for these two harvests ADL concentrations were lower at toe-slope positions. Slope x date interaction (P < 0.05) in 2003 was driven by lower ADL concentrations at mid- and shoulder-slope positions in June.
|
Forage ADL under honey locusts varied little (23.7 g kg–1) across density treatments but increased with increasing tree density (from 18.7–23.6 g kg–1) in forage grown under walnut (species x density interaction, P < 0.05). Increases in ADL with shade have been reported by others (Lewis et al., 1983; Samarakoon et al., 1990; Kephart and Buxton, 1993). Morphological adaptations to low light environments generally include increased internodal length and reduced specific leaf weight (Allard et al., 1991), with internodal tissues typically containing elevated levels of lignin.
Total Nonstructural Carbohydrate
The response of forage TNC concentrations to slope position varied with tree species within each year (slope x species interaction, P < 0.05). In 2002, TNC concentrations in forage from under honey locust trees declined from toe- to shoulder-slope positions at May and June harvests, but TNC levels were lower only at mid-slope under walnut trees (Fig. 1
). This pattern was reversed in July 2002, when forages at mid-slopes had TNC concentrations equal or greater to TNC in forage from toe slopes, regardless of tree species. The TNC levels in forage from toe-slope positions in June 2002 reflect the dynamic effect of the tree canopy on forage TNC. A mid-May frost in 2002 settled on the toe slopes and froze back tree leaves at those sites. The tendency (P < 0.10) of greater TNC at toe-slope positions in June 2002 indicates that toe slope trees had not fully recovered. However, differences due to slope or species were not apparent by July 2002 as the tree canopies developed. In 2003, concentrations of TNC under honey locust tended (P < 0.10) to be greater at mid-slope positions in May and were greater (P < 0.05) in June. Forage from under walnut trees had greater TNC at the shoulder-slope positions over the season. Walnut trees at shoulder slopes were smaller than honey locust trees and allowed more light to reach the forage canopy. This may have boosted TNC concentrations at those sites given the high rainfall and cloudy skies of the 2003 growing season.
|
|
The potential impact of changes in TNC to grazing animals is unclear. Animals prefer forages with greater TNC (Ciavarella et al., 2000; Mayland et al., 2000; Fisher et al., 2002), and even changes as small as 10 g kg–1 of TNC between forages are sufficient to affect animal selection (Burritt et al., 2005). However, performance of animals in maturing silvopastures has received little research attention. Furthermore, while reduced TNC with shading is to be expected, this may be offset by increased forage digestibility (Garrett and Kurtz, 1983). As pastures are typically energy-limiting for livestock, TNC levels (relative to that of open pasture) may have potential to serve as a guide for pruning and thinning and supplementation decisions.
Crude Protein
Concentrations of CP were greater (P < 0.05) in forage from toe-slope positions in 2002 (Table 6). This was largely due to the strong effect of slope in June 2002 (slope x date interaction, P < 0.05) and likely reflects the better moisture status at toe slopes given a persistent water deficit in 2002. Slope position did not affect CP levels in 2003, perhaps due to the greater precipitation and cloudy conditions from that year.
|
Crude protein concentrations tended (P < 0. 10) to increase with tree density in each year, although the change with each increase in tree density was often less than 5 g kg–1. Shaded plants often have greater concentrations of CP (Kephart and Buxton, 1993; Wilson, 1995, 1996), thus greater CP concentrations in forage would generally be expected with increasing tree density. Strongest (P < 0.05) responses to density were observed in June 2002 and May 2003. Between the two tree species, changes in CP concentration were similar with density at the May and June 2002 harvests, but in July 2002, CP concentrations decreased with density under honey locusts (species x density and species x density x date interactions, P < 0.05).
Minerals
Calcium, Phosphorus, Magnesium, and Potassium
Forage Ca, P, and K concentrations were greater (P < 0.05) at toe-slope positions in 2002, likely due to greater moisture availability (Fig. 2
). Forage Mg was unaffected by slope. No differences by slope position were observed in 2003, likely due to more uniform soil moisture and to limestone and fertilizer applications in October 2002.
|
Over each season, forage mineral concentrations generally increased (P < 0.05) with tree density (Fig. 3 ). These increases were not simply due to changes in nutrient concentration, because forage grown under medium density sites typically had both greater mineral nutrient concentrations and greater biomass yield (Buergler et al., 2005). Concentrations of both Ca and P increased in dry matter with successive harvests in 2003, but the increase was greater at toe slopes for Ca and at upper slopes for P in 2003 (slope x data interaction, P < 0.01) (Buergler et al., data not shown).
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 | |||
, tree species
, tree density
, and harvest date in an emulated silvopasture.
