Effect of Light on Pasture Productivity and Quality in A Silvopastoral System

Reader Impact Factor Score
[Total: 2 Average: 3]

Published on International Journal of Forestry & Plantation
Publication Date: June, 2020

Michael Jide Nworji

Department of Forestry and Wildlife Management, Faculty of Agriculture, Chukwuemeka Odumegwu Ojukwu University, Igbariam Campus, Anambra State, Nigeria

Journal Full Text PDF: Effect of Light on Pasture Productivity and Quality in A Silvopastoral System.

This study evaluated the influence of solar radiation on the production of dry matter and nutritive value of understory pasture in thinned red alder blocks in a silvopastoral system with particular attention on how the current tree density treatment of 100-stem ha-1 affected the light at ground-level and how this in turn affected the understory pasture production and quality. This was achieved by estimating seasonal pasture biomass production and forage quality under varying solar radiation intensities using pasture clippings from 0.25 m2 grazing exclusion cages systematically placed in varying canopy gap levels created as a result of the thinning, and hemispherical view colour images of the overstory canopy gaps taken with the use of a Nikon Coolpix 990 digital camera fitted with a Nikon FC-E8 fisheye converter (lens) pointed upward and mounted on a tripod vertically above each exclusion cage. Forage quality was quantified in terms of protein content, fibre content, and energy potential in dried and milled forages using the Near Infrared Reflectance Spectroscopy standard procedures of forage quality analysis. A strongly significant (p < 0.001) correlation was established between solar transmission at ground level and the observed pasture dry matter production, pasture quality parameters, and distance from each grazing exclusion cage to the nearest tree. Pasture dry matter production was found to increase significantly (p < 0.05) with increasing solar transmission (decreasing shade), and with increasing distance from each grazing exclusion cage to the nearest tree. The study also linked observed seasonal variation in the pasture yield and quality parameters to leaf fall (deciduous nature of red alder tree). Mean pasture yield and quality was shown to be greater in spring, summer and early autumn when the alder trees have leaves, and lower from mid-autumn to late winter when the trees are without leaves. The study concluded that thinning can be used to increase understory pasture production and nutritive value in a silvopastoral system by increasing transmitted light. These findings can help livestock managers make appropriate decisions of tree stand spacing optimization, stocking rate, pasture grazing and wood quality improvement.

Keywords: Agroforestry, silvopasture, pasture, forage, red alder, understory, canopy, solar radiation.

Agroforestry system has been defined as a land use system and technologies where woody perennials are deliberately integrated with agricultural crops and/or animals on the same land management unit, in either a spatial arrangement, or a temporal sequence, there being both ecological and economic interactions between the different components. In a general sense, agroforestry is a term covering all farm practices that deliberately combine the production of trees and/or shrubs with other crops and/or livestock in a manner that may be collectively beneficial. Classification of agroforestry systems has been based on their structure, function, socioeconomic aspects and ecological spread (Devkota et al. 2001). On the basis of structure, Nair (1987) specified three common basic classification types: a) agro-silvicultural, a combination of arable crops and trees, b) silvo-pastoral, which includes trees plus pasture/livestock, and c) agrosilvo-pastoral, which covers crops, pasture/livestock and trees. Within the agroforestry concept, silvopastoral systems are those where trees are combined with forage and livestock production on the same land management unit. Based on the context of the definitions, agroforestry can further be regarded as an intervention and silvopastoral systems as the link between the system components of trees and livestock (Devendra and Ibrahim, 2004). Integrating trees, forage, and livestock creates a land management system that can produce marketable products while at the same time maintain long-term productivity.
Through the intentional integration of trees with livestock, silvopastoral practices strive to simultaneously optimize economic, environmental and social benefits, thereby ensuring the attainment of multiple objectives. There are four main components of silvopastoral practices that can be readily manipulated at any given site – trees, pastures, animals and the soil (Mead (2009). The interactions between these major components are dynamic and their understanding is important in the development of comprehensive management practices. Understorey pasture production in silvopastoral system is influenced by tree density and light penetration, with light penetration decreasing with increasing tree density. Light usually becomes the dominant competitive factor with time as the trees shade the pasture. Trees, because of their growth habit, will shade pastures, the degree of shade being related to the density of the tree canopy. How quickly this happens depends on tree species, their temporal and spatial arrangement, age, and factors that influence tree vigour.
The influence of trees on the understory pasture is contingent upon the degree to which they modify the microclimate and soil properties (Benavides et al 2009). The performance of a pasture is mainly influenced by humidity, temperature and radiation (Mc Calla and Bishop-Hurley, 2003) and, in temperate regions, the best relationship between these factors are generated in the period of spring (Skinner et al., 2009). The tree–pasture-animal interaction affects not only pasture production but also pasture quality and through that animal productivity and plant nutrient status (Mead, 2009). However, other interactions related to shelter and animal health can occur, of which reducing stresses on animals can often be very important. Thus, agroforestry is much more complex than either pastures or forests on their own and the mixing of these two major components can result in some unexpected interactions, both positive and negative.
Pasture production beneath trees is normally governed by the degree of competition between trees and pasture for light, moisture, and nutrients (Mead 2009; Dodd et al., 2005). Again, understory pasture dry matter (DM) yield and quality (nutritive value of herbage) are known to be strongly influenced by tree shading, which is a function of the degree of overstory canopy closure and available light (Kephart and Buxton, 1993; Sibbald et al., 1994; Knowles et al., 1999; Devkota et al., 2001; Lin et al., 2001; Sigurdsson et al., 2005; Peri et al., 2007; Benavides et al. 2009). Their amount and values are useful indicators of the sustainability of silvopastoral farms as they have significant impact on both resource status and economic performance (Lambert et al., 1996).
Benefits of silvopastoral systems are numerous and have been summarized by Mosquera-Losada et al. (2005). Extension of the growing season of herbage via protection of swards from environmental extremes and overall increases in forage production have been shown by Sibbald (1999). Kephart and Buxton (1993) found that shade tended to decrease secondary cell-wall development and proposed that morphological changes in herbage grown under reduced light, e.g. under a tree canopy or in areas with prolonged cloudiness, would very likely increase the nutritive value of herbage, estimated in terms of its crude protein (CP) concentration. Peri et al. (2007) also found increased CP concentrations with increasing shade for herbage of Dactylis glomerata.
Conversely, low radiation levels have been shown to reduce forage production and nutritive value. Research in Scotland, UK (Sibbald et al., 1994) showed that herbage production decreased with increased shading (or attenuation of full sunlight) when precipitation and temperature favour herbage growth. Belesky (2005) and Peri et al. (2007) found that herbage plants grown in areas with lower light levels were smaller, had fewer numbers of tillers and produced less dry matter (DM) compared with treatments with higher levels of radiation. Shade-grown grasses of cool-temperate origin increase allocation of N to leaves to maximize light acquisition. Lin et al. (2001) found that, in general, acid– detergent fibre (ADF) concentration was either unaffected or increased because of shading. The high nitrate concentrations, along with depressed levels of total non-structural carbohydrates (TNC), found in shade-grown herbage (Deinum et al., 1968; Chiavarella et al., 2000) could compromise nutritive value. Concentration of TNC in herbage has been positively associated with improved dietary protein utilization in the rumen, and increased selection and intake by grazers (Chiavarella et al., 2000; Mayland et al., 2000). High levels of N in herbage have also been associated with off-flavours in meat from pasture-raised beef cattle (Lane and Fraser, 1999).
The measurement of pasture productivity and quality in grazing systems is a complex issue that relates the variations in climate to seasons and the soil – plant-animal interaction, dynamics of water and nutrients from soil, changes in botanical composition, and seasonal variations in the stocking rate, grazing intensity and frequency. The performance of a pasture is mainly influenced by humidity, temperature and radiation (Mc Calla and Bishop-Hurley, 2003) and, in temperate regions, the best relationship between these factors are generated in the period of spring (Skinner et al., 2009). Pasture biomass productivity and quality values are crucial in management of grazing lands and livestock. More accurate and timely estimation of pasture biomass production and forage quality during the grazing season can help livestock managers make appropriate decisions of pasture fertilization and stocking rate. Laboratory analyses of the composition of feed or forage are used to assess their nutritive value. A typical feed analysis includes measurements of some important quality attributes or parameters (e.g., crude protein, fibre, digestibility, etc.) used to define nutritive value. Conventional laboratory chemical “wet chemistry” methods of laboratory chemical analysis have long been used for assessment of forage quality (Kellems and Church 1998). Neutral detergent fiber (NDF), acid detergent fiber (ADF) and crude protein (CP) concentrations are commonly used forage quality variables (Ball et al. 2001). These three quality variables are closely associated with intake potential, digestibility, and nutritive values of forage (Ball et al., 2001). Conventional “wet chemistry” methods used to determine these quality variables are time consuming and costly, and also require personnel with special skills. Additionally, the hazardous waste generated from laboratory processes must be disposed of in order to reduce the risk of environmental pollution.
Another technique for the assessment of forage is the use of the near-infrared reflectance spectroscopy (NIRS) method. This is a rapid and inexpensive computerized approach to quantify the nutritive values of forage and grain crops (Marten et al., 1989; Shenk and Westerhaus 1994). Studies indicate that strong correlations exist (r > 0.95 between NIRS and the various components of forage nutritive value (Norris et al., 1976; Shenk et al., 1979; Counts and Radloff, 1979; Ward, 1980). NIRS uses near-infrared light instead of chemicals as in conventional methods, to determine protein, fibre, mineral, energy and other variables of interest. In NIRS method, air dried and finely ground samples are exposed to infrared light in a spectrophotometer. The reflected infrared radiation is converted to electrical energy and fed to a computer for the determination of the quantity of these components in the feed. It is based on the fact that each major organic component of forage and grain will absorb and reflect near-infrared light in a different way (Norris et al., 1976; Stermer et al., 1977; Shenk et al., 1979). Though NIRS analysis is fast and very precise but its accuracy depends on appropriate calibration with adequate number of “wet chemistry” samples similar to those being analysed, and therefore requires a period of time for the preparation of the samples.
Understory pasture productivity is mainly affected by the intensity of solar radiation reaching the forest floor, which in turn depends upon the degree of canopy closure and characteristics of the tree canopy. Citing Knowles et al. (1997), Devkota et al. (2001) observed that canopy closure and available light provide measures of the potential shading by trees that are independent of tree stems per hectare and the height to which trees have been pruned. The authors concluded that canopy closure is the critical factor to manage if pasture production is to be maintained at an economic level and suggested a critical threshold range of 40-50% as the percentage of tree canopy closure that would be required to maintain pastoral enterprises under deciduous tree based silvopastoral systems.
Understanding of the relationship between canopy closure and understory pasture DM production is crucial in the development of comprehensive management practice for deciduous tree based silvopastoral systems. Since there is inverse relation between understorey pasture productivity and tree density, thinning of trees for greater light penetration will be necessary to achieve maximum forage yield. In winter of 2012, the three red alder 200-stem ha-1 blocks at Henfaes SNNE were thinned down to 100-stem ha-1. Understorey pasture DM production has been strongly linked to overstory canopy closure (Knowles et al. 1999) and, unlike tree spacing, this index also accounts for both the arrangement and size of the trees. Hence, there is the need to develop a relationship between red alder canopy closure and understory pasture DM production, which can be used by farmers to optimise the spacing of alder-stands for multiple objectives, such as soil conservation, pasture grazing, stocking rate, and wood fuel. The objective of this study was to evaluate the influence of solar radiation on the production of DM and the nutritive value of understory pasture in thinned red alder blocks in a silvopastoral system with particular attention on how the current tree density treatment of 100-stem ha-1 affected the light at ground-level and how this in turn affected the understory pasture production and quality.

2.1 Study area description
The study was conducted in the years 2012 to 2014 at the United Kingdom’s Silvopastoral National Network Experiment (SNNE), Henfaes in North Wales, which is one of six National Network Experiments established across the country with trees planted at different configurations and densities to investigate the potential of silvopastoral agroforestry on UK farms (Sibbald and Sinclair, 1990). The site was established in 1992 on 14.47 ha of agricultural land at the Bangor University’s Henfaes Silvopastoral Systems Experimental Farm (SSEF) (53°14′N 4°01′W), Abergwyngregyn, Gwynedd, North Wales (Figure 1).

Figure 1. Location and aerial photograph of the Henfaes study site

The local climate in Henfaes is hyperoceanic, cool and temperate. Mean monthly temperature over the course of this study period was 10.6 oC, and temperatures of the warmest and coldest month was 20.0 oC in July and 3.4 oC in January, respectively. Average monthly precipitation ranged from a minimum of 25 mm in April to a maximum of 114 mm in December. Soil is a fine loamy brown earth over gravel (Rheidol series) classified as a Dystric Cambisol (Teklehaimanot and Mmolotsi, 2007). The parent material consists of postglacial alluvial deposits from the Aber River with a water table that is between 1 and 6 metres deep (Teklehaimanot and Sinclair, 1993). Further details of the site topography, climatic conditions, soil geology and hydrology etc. can be found in Teklehaimanot et al. 2002 and Sibbald & Sinclair, 1990.
Sycamore (Acer pseudoplatanus) and red alder (Alnus rubra) were planted on the site at establishment in 1992 at different configurations to investigate their use in agroforestry systems (Sibbald et al., 2001). The blocks were 4,225 m2 (0.42 ha) each and sown to a mixture of perennial ryegrass (Lolium perenne L.) and white clover (Trifolium repens L.) at establishment of the farm. All treatments and controls are replicated three times in a complete randomised block design. The experimental area was rotationally grazed by sheep throughout the period of the study at average stocking rate of 0.5 to 1.0 AU per ha (Teklehaimanot et al., 2002). For the purpose of this study, due to limited time and resources, only the three-red alder (Alnus rubra) blocks were studied. The red alder that was originally planted at 400 stems ha-1 across three blocks were selectively thinned to 200 stems ha-1 in 2000 and subsequently to 100 stems ha-1 in the winter of 2012, respectively, primarily to improve the health and productivity of both the trees and the understory pasture as well as to provide data for the construction of biomass allometric equations for open-grown red alder trees (Nworji, 2017, 2019). A detailed description of the experimental design and site characteristics is given on the Henfaes agroforestry website.

2.2 Determination of the seasons
Data on the climatic condition of the study area were collected from the on-site automatic weather station. Data for three years (2012-2014) were averaged for precipitation, solar radiation, temperature and relative humidity at the study site.

2.3 Assessing pasture biomass
All trees in the three red alder blocks were inventoried prior to and following random thinning operations in winter of 2012. In each of the three thinned alder blocks, seven 0.25 m2 grazing exclusion cages were systematically placed in varying canopy gap levels created as a result of the thinning. The perimeters of these 7 canopy gaps were not delimited, however, the distance of each cage from the nearest tree was measured in metres with a 50 m measuring tape. For consistency in data collection, pastures in all exclusion cages were clipped to a residual level of 2.5 cm at the start of the setup on June 30, 2013. Thereafter, forages in the grazing exclusion cages were harvested to a residual sward height of 2.5 cm on the last day of every month from July 2013 to June 2014. The grazing exclusion cages were permanently positioned and maintained within the predetermined locations all through the study period.
To determine their dry matter content, the harvested forages were taken immediately to the Henfaes Research Centre laboratory, weighed with a 0.005g precision balance, and sub-samples were extracted for the evaluation of the dry matter content, weighed and dried in a forced air oven at 60oC (140oF) for 24 hours or until constant weight was achieved. This temperature is sufficiently high to decrease the water considerably and low enough not to significantly modify sample chemistry. The monthly production of dry matter per hectare was obtained by multiplying the green matter production value by the dry matter content and dividing by 100. The daily growth rate per hectare was obtained by dividing the monthly production by the number of days in each month. To express production in kg DM ha-1, the daily production rate per hectare was multiplied by a conversion factor of 0.1.

2.4 Assessing forage quality
To assess pasture quality, the monthly oven-dried forage was then milled to pass a 1-mm sieve. The 12 months (July 2013 to June 2014) collection of dried and milled forages were pooled based on the seasonal differences in the condition of the tree canopy, that is, with-leaves condition (months of March to September when the trees are with foliage) and without-leaves condition (months of October to February when the tress are without foliage). The pooled samples were stored in sealed plastic bags and taken to the feed and forage analysis laboratory of Bioparametrics Ltd, Peter Wilson Building, West Mains Road, Edinburgh, Scotland, EH9 3JG, United Kingdom in early August 2014 for quality analyses. Forage quality was quantified in terms of protein content (Crude Protein), fibre content (Acid Detergent Fibre and Neutral Detergent Fibre), and energy potential (Metabolisable Energy). The laboratory applied the Near Infrared Reflectance Spectroscopy (NIRS) (Model FOSS 6500, NIRSystems Inc., Silver spring, MD, USA) standard procedures of forage quality analysis as outlined by Stuth et al (2003) to determine the concentrations of crude protein (CP), acid detergent fibre (ADF), neutral detergent fibre (NDF), and metabolisable energy (ME) in the dried and milled forages. The amount of pasture CP, ADF, NDF, and ME availability in g m-2 d-1 were estimated by multiplying pasture biomass by the CP, ADF, NDF and ME concentrations, respectively.

2.5 Light measurement
A total of 42 hemispherical photographs taken looking upward were used to estimate solar radiation penetration to the canopy gaps of varying sizes created by the thinning of the three red alder blocks. These comprised of 21 photos taken above the 21 grazing exclusion cages in August 2013, when the leaves were fully expanded and repeated in December 2013, after the trees had completely shed their leaves (Figure 2). Light transmission through the canopy gaps was measured with the use of a Nikon Coolpix 990 digital camera (Nikon Corporation, Tokyo, Japan) fitted with a Nikon FC-E8 fisheye converter (lens) pointed upward (Nikon Corporation, Tokyo, Japan). The digital camera was mounted on a tripod vertically above each exclusion cage in the gaps at a height of approximately 1.5 m above the ground and hemispherical view (HemiView) colour images of the overstory canopy gaps were taken. The top of the camera was oriented relative to magnetic north and positioned horizontally with the aid of a spirit level, and adjustment for magnetic declination was made during the photo analysis (Rich, 1990). Automatic settings were selected for aperture width and shutter speed (Inoue et al., 2004). The digital images were downloaded directly to a personal computer and analysed with image-processing software, HemiView Version 2.1 (Delta-T Devices, Cambridge, UK). The images were processed following the approach of Brunner (2002). This comprised the manual setting of a threshold value to separate canopy and sky elements into a binary black and white image. The lens distortion was corrected using the Coolpix 900 option (Hale and Edwards 2002).
The hemiView photos were used to calculate absolute amount of radiation beneath the tree canopies. Hemispherical photographs were analysed to derive a variety of solar radiation indices such as gap fraction (GF), direct site factor (DSF), indirect site factor (ISF), and global site factor (GSF).

Figure 2: Hemispherical view (Hemi-View) images of red alder overstorey canopy gaps of trees with leaves and trees without leaves.

The primary indicator of canopy gaps was canopy gap fraction (GF). Gap fraction is the proportion of open area within a canopy, the fraction of view looking up from beneath the canopy that is not blocked by wood and foliage. Gap fraction measurements hinge on dividing the sky into several sectors and calculating the gap fraction for each sky sector (Awal 2008). A gap fraction of zero (0) means the sky is completely blocked (obscured) in the particular sky sector, whereas a gap fraction of one (1) means the sky is completely visible (not obscured). The gap fraction algorithm used in the instrument assumes a diffusely lit sky, so wherever possible, images were captured under calm and completely overcast sky conditions to maximize image contrast and minimise interference by direct sunlight (e.g. sunflecks).
Global (or total) solar radiation (also called global site factor) is the amount of solar radiation actually reaching a particular location and is influenced mainly by cloudiness, time of year, latitude, and surface geometry (Igbal, 1983). Global solar radiation (global site factor) is the sum of the direct irradiance (direct site factor) in the sunlight and the diffuse solar radiation (indirect site factor) scattered from sunlight as it passes through the atmosphere (Rich 1990).
Data for photosynthetically active radiation (PAR), taken hourly over the study period (2012-2014), were procured from the automatic weather recording station in the study area (Henfaes research centre). Analyses of radiation flux was applied for photosynthetically active radiation (400-700 nm) photon flux density in order to determine radiation available for photosynthesis.
The relative contribution of direct sunlight and diffuse skylight to global radiation flux was expressed as direct and indirect site factors, respectively. Direct site factor is the proportion of direct sunlight and indirect site factor is the proportion of diffuse skylight under the canopy relative to that outside the canopy (Rich 1990).
Radiation flux density was calculated as the sum of direct sunlight and diffuse skylight that passed unimpeded through canopy openings (gaps) and expressed in absolute units as mole per square metre per day (mol m-2 d-1). Calculation of percent solar radiation values was therefore based on actual measurement of global site factor that passed through the canopy gaps.

2.6 Data analysis
Statistical analysis was carried out by checking assumptions of normality, homogeneity of variance and multicollinearity and transforming the data to logarithms as appropriate. For log-transformed variables, the mean of the untransformed data was used to express central tendency and the standard error derived from log-transformed data was used to express precision.
An ANOVA was used to determine the impact of thinning on pasture productivity and quality parameters. The productivity and quality data with corresponding radiation flux densities were pooled over pasture blocks, gap locations, months and seasonal conditions (n = 252). Where appropriate, a linear mixed model was fitted using the method of restricted maximum likelihood (REML) to take account of the variance components of the random effects of month, season, and canopy gap level (location) and their interactions on pasture productivity.
A subset of data comprising the condition of alder trees in two seasons (with- and without-leaves) were used for comparison. Herbage harvested from March to September corresponded to 7 months of production under the with-leaves condition, referred to here as with-leaves condition and harvests from October to February corresponded to 5 months, referred to as without-leaves condition.
Linear regression analysis was used to explore the functional relationships between solar radiation transmission and pasture productivity and quality. Differences were assessed at the significance level of p <0.05. All analysis was conducted using the SPSS version 22 software.

3.1 Pasture production
Results of the monthly, seasonal and locational pasture productions parameters obtained from 21 grazing exclusion cages in the canopy gaps in three alder blocks in the 2013/2014 growing seasons are presented in Figure 3, to Figure 6. Significant (p < 0.05) differences were observed between the monthly, seasonal and locational pasture dry matter (DM) yield for the amount of solar radiation transmitted through the tree canopy of red alder.
Linear fitted curves were attempted to predict the trend of monthly, seasonal and locational averages of pasture dry matter yield. The results of this study show that all coefficients were statistically significant (p ≤ 0.0001). Analysis of the mixed model determined by the variance of components, indicated that about 45% of the total variation in pasture dry matter yield can be attributed to seasonality, 40% to month, and 15% to location (grazing exclusion cage locations). Only 0.6% of the variation was attributable to the blocks, which did not show significant values (p = 0.05).
Strong significant relationship was obtained between solar transmission and monthly pasture production (R2 = 0.90, p < 0.05). Pasture productivity increased with increasing solar radiation intensity (Figure 3). Solar radiation reaching the various grazing exclusion cages ranged from a minimum of 4.99 mol m-2 d-1 in December to a maximum of 64.13 mol m-2 d-1 in June (Figure 3).

Figure 3: Mean monthly levels of solar radiation in relation to pasture productivity.

Pasture yield was significantly different (p < 0.05) between months (Figure 3). Mean pasture yield ranged from a daily minimum of 0.70 g DM m-2 d-1 in December to a daily maximum of 5.49 g DM m-2 d-1 in June. Higher mean pasture yield was recorded in the months of April, May, June, July, August and September with a peak in June (5.49 g DM m-2 d-1) for the three alder blocks, which corresponded to the seasons of spring, summer and early autumn, and the months when the alder tree leaves are in bloom, and temperature and radiation levels are the highest and precipitation is close to moderate level (Figure 3). Lower mean pasture yield was obtained from mid-autumn (October) to late winter (February) with a deep in December (0.70 g DM m-2 d-1), which corresponded to the seasons when the trees are without leaves and temperature and radiation levels are lower.
Again, pasture yield differed significantly (p < 0.05) between seasons. Seasonal herbage productivity model showed that 25.85% of seasonal herbage yield occurred in spring, about 47.71% were obtained in summer, 16.62% in autumn and only 9.82% in winter (Table 1) .

Furthermore, significant (p < 0.05) relationship was found to exist between pasture production and distance of each grazing exclusion cage to the nearest tree. Cage distance to trees varied from a minimum of 5 m to a maximum of 11.25 m with a mean distance of 8.17 m and standard deviation of 1.90 m (Table 2). The proportion of incident solar radiation reaching the cages increased significantly (R2 = 0.79, p < 0.05) as the distance between each cage and the nearest tree increased (Figure 4). Similarly, proportion of pasture production increased significantly (R2 = 0.81, p < 0.05) with increasing distance from each cage to the nearest tree (Figure 5).

3.2 Forage quality (nutritive value)
Descriptive statistics for light transmission and the pooled dried and milled forage parameters are presented in Tables 2 & 3. Results show that solar radiation, DM yield and pasture quality parameters (nutritive values) varied significantly (p < 0.05) between the with-leaves condition (months of March to September when the trees are with foliage) and the without-leaves condition (months of October to February when the trees are without foliage).
Solar radiation reaching the various grazing exclusion cages varied between 12.03 and 41.27 mol m-2 d-1 in with-leaves condition and from 7.46 to 11.57 mol m-2 d-1 in without-leaves condition (Table 2). Mean concentrations of DM, CP, ADF and NDF were 94.43%, 20.07%, 29.15% and 57.23% of dry matter, respectively, in with-leaves condition compared to 95.29%, 18.85%, 28.19% and 51.37%, respectively, in without-leaves condition. Mean metabolisable energy varied from 8.33 MJ kg DM in without-leaves condition to 8.60 MJ kg DM in with-leaves condition (Table 2).
Table 2: Descriptive statistics of light incidence, percentage dry matter (DM), crude protein (CP), acid detergent fibre (ADF), neutral detergent fibre (NDF) and metabolisable energy (ME)

In with-leaves condition, variance for the pasture parameters was highest in CP (CV = 9.36%) and lowest in DM (CV = 0.31%) (Table 2). In without-leaves condition, CP and DM again showed the highest (CV = 8.38%) and the lowest (CV = 0.67%) variance, respectively. Between seasons, variance in solar radiation was higher in with-leaves (CV = 38.21%) than in without-leaves (CV = 9.95%). Pastures in without-leaves growing conditions had 56% lower solar transmission, <1% higher DM % content, 6% lower crude protein concentration, 3% lower acid detergent fibre concentration, 10% lower neutral detergent fibre concentration, and 3% lower metabolisable energy compared with the results in with-leaves growing conditions.
Similarly, DM yield and available CP, NDF, ADF and ME obtained by multiplying their respective concentration values by the DM yield differed significantly (p < 0.05) between the with-leaves and the without-leaves conditions (Table 3). Mean daily DM yield was 3.03 g DM m-2 d-1 in with-leaves conditions and 0.98 g DM m-2 d-1 in without-leaves conditions. Mean daily available CP, ADF, NDF, and ME were 0.62 g m-2 d-1, 0.90 g m-2 d-1, 1.76 g m-2 d-1, and 0.26 MJ g m-2 d-1, respectively, for the with-leaves conditions compared to the without-leaves conditions. Generally, the pasture parameters show greater variability in without-leaves conditions compared to with-leaves conditions while light transmission exhibited greater variability in with-leaves (CV = 38.21%) than in without-leaves (CV = 9.95%).

Table 3: Descriptive statistics of dry matter yield (DM), and available crude protein (CP), acid detergent fibre (ADF), neutral detergent fibre (NDF) and metabolisable energy (ME).

A Pearson product-moment correlation coefficient was computed to assess the relationship between the light transmission, DM yield and available CP, ADF, NDF and ME. All parameters were found to be significantly correlated with each other (p < 0.001) (Table 3). Again, there was positive correlation between all the variables in both with-leaves and without-leaves conditions, r ≥ 0.657, n = 21, p < 0.001. Based on the range in the light transmission and forage variables, a linear relationship was established between light and the forage parameters in with-leaves (R2 ≥ 0.76) and in without-leaves (R2 ≤ 0.62) by adjusting the data to a linear function (Tables 3 & 4). A scatterplot summarizes the results (Figures 6 a-j). Overall, light transmission was strongly correlated with DM yield, CP, ADF, NDF and ME in with-leaves (r ≥ 0.872) but only moderately correlated with these variables in without-leaves (r ≤ 0.657).

The light availability at ground level, measured at varying canopy gap levels created by purposive thinning of red alder blocks, correlated significantly (P < 0.001) with the observed changes in pasture productivity in the present study. Results show that pasture DM yield increased significantly (p < 0.05) with increasing solar transmission (Figures 4), and with increasing distance from each grazing exclusion cage to the nearest tree (Figures 5). Furthermore, pasture DM yield and pasture quality parameters varied significantly (p < 0.05) between the with-leaves condition and the without-leaves condition (Figure 6). Productivity of pasture have most often been related to competition with the overstory trees for light, soil water or nutrients. Studies have shown that, among those environmental factors, light obstruction by tree canopy has been the main driving factor in most temperate forest ecosystems (Sigurdsson et al., 2005; Anderson et al., 2001; Engelmark et al., 2001; Peterken, 2001; Nygaard and Odygaard, 1999; Stone and Wolfe, 1996). Those authors have observed that more open canopy (less shade) would lead to both increase in pasture productivity and diversity. Again, light availability at ground level, measured as canopy gap fraction, has been found to be the factor that correlated best with the observed changes in ground vegetation biomass and composition (Gonzalez-Hernandez et al., 1998; Sigurdsson et al., 2005). Therefore, it can be said that level of exposure to solar radiation (the level of shade) is a significant factor determining the productivity of pastures. This result is in agreement with the findings of previous research studies. This result is comparable to that of Knowles et al. (1999) where a strong relationship (R2 = 0.89) was shown to occur between measured pasture yield and predicted canopy closure. In a research studies in Scotland, UK Sibbald et al. (1994) found that herbage production decreased with increased shading (or attenuation of full sunlight) when precipitation and temperature did not limit herbage growth. Lin et al. (2001) also showed a reduction of yield because of increased shade for orchard grass, ryegrass and white clover. This is also in agreement with Peri et al. (2007) and Neel et al. (2008). Tree shade limits pasture photosynthesis (Rao et al., 1998; DeMontard et al., 1999; Sharrow, 1999; Esquivel-Mimenza et al., 2013) particularly in C4 species, such as Brachiaria brizantha where the rate of photosynthesis tails off at about 1500 µmol m-2 s-1, which is a bright but not fully sunny day in summer. Pasture grown in areas with lower light levels have been found to be shorter, had fewer numbers of tillers and produced less dry matter compared with treatments with higher levels of radiation (Belesky, 2005; Peri et al. (2007). Ehrenreich and Crosby (1960) reported higher production for understory plants within a hardwood forest as crown cover decreased. Production greatly decreased when canopy cover increased. However, the extent of biomass reduction observed largely depends on the interception of solar radiation caused particularly by the tree species. Decreasing irradiance reduces the growth of pasture species (Smith and Whiteman 1983; Shelton et al., 1987) and influences the outcome of relationships. It has also been noted that light interception by the trees is one of the main factors affecting the productivity of pasture in silvopastoral systems, especially when water and soil nutrients are freely available (Ong et al., 1996; Power et al., 2001). Furthermore, this result agrees with tree-removal studies that report that herbaceous forage production increases as trees are eliminated (Pratchett 1978; Walker et al., 1989; Harrington and Johns 1990). Thinning trees generally increases biomass productivity of understorey plants especially when pre-treatment stand density is high (Uresk & Severson 1998; Brockway et al., 2002). Ducherer et al. (2013) reported that total understory biomass increased up to 80% within 3 to 4 years after thinning in the Ponderosa pine and Douglas fir forests. Depending on sites and years, biomass production of one or more functional groups, such as forb, shrub, or graminoids, may increase. Uresk and Severson (1998) also reported that eliminating or reducing the overstory in ponderosa pine forests increases understory biomass production. Studies of thinning forested environments to create desirable levels of available light have been conducted. Garrett et al. (2004) described two popular harvesting practices for creating light levels favourable for the growth of forages in hardwood forests: the group selection and the shelterwood methods. The group selection method creates patches of high light intensity, while the shelterwood method is designed to create a more even distribution of light throughout the forested understory. They further recounted that in young immature stands, release thinning such as timber stand improvement, crop tree, and deferment cuts, will provide increased light levels for forage production while at the same time improving the growth of trees identified for retention (Garrett et al., 2004). Dey and Parker (1997) reported that with a shelterwood harvest in an oak stand, removing 43 and 77% of the basal area increased light intensities to 35% and 65%, respectively. Other studies have discovered that up to 50% of the basal area of hardwood forests may need to be cut to increase light levels to 35- 50% of that found in the open (Sander 1979, Marquis 1988, Dey and Parker 1997). The issue of whether thinning improve or degrade understory pastures productivity and quality is sometimes contentious. Studies have indicated that forage production is often reduced by trees that compete with understory herbaceous species for water, nutrients, and light (Kay and Leonard 1980; Monk and Gabrielson 1985; Burrows et al., 1990; Belsky 1992). As a result, pasturelands are cleared of trees by expensive mechanical and chemical techniques (Beisky 1992). However, other studies have indicated that trees may increase forage production in areas of low tree density, moderate or high soil fertility, and low rainfall (Belsky et al., 1993; McClaran and Bartolome 1989; Burrows et al., 1990; Wilson et al., 1990; Belsky and Amundson 1992). In some instances, full sunlight may not be required to maximize pasture growth. Many understory pastures will need only about 10% of full sunlight to reach a state of growth where photosynthesis exceeds respiration, and will reach light saturation at 50% (C3 plants) and 85% (C4 plants) of full sunlight, respectively (Gardner et al., 1985). The light intensity within mature hardwood forests is typically lower than 20% and may be as low as 1% (Dey and MacDonald 2001). While the opening up of some proportion of the canopy can increase the intensity and duration of light reaching the forest floor, and thereby improve the growth of forage crops, the relationship of the canopy opening to available light is not linear, with an estimated residual stocking density of 30% required to provide light levels of 50% of open values (Sanders 1979). The variability in the pasture productivity and quality parameters observed in this study has also been associated with solar transmission intensity and duration (Photoperiod), and leaf fall (deciduous nature of red alder tree), and the distance of each grazing exclusion cage from the nearest tree. The with-leaf period corresponded to the summer months (April to October) when the alder trees are with leaves, and temperature and radiation levels are the highest and precipitation is close to moderate level while the without-leaf period corresponded to the winter months (November to March) when the trees are without leaves and temperature and radiation levels are lower (Figures 3 & 4). Total annual precipitation at the experimental location in 2013 (892 mm) was much higher than that in 2014 (607 mm). Again, both light intensity and length of day are of importance. In the United Kingdom, photoperiod (the duration of sunshine) varies broadly with the time of the year, with long periods of daylight during the with-leaves condition and short periods of daylight during the without-leaves condition. The length of day is longer in the 7 months of with-leaves condition than in the 5 months of without-leaves condition. Thus, the rate of daily growth is greater in with-leaves condition with longer hours of sunlight, and slower in without-leaves conditions with shorter hours of sunlight. Variation in photoperiod can influence induction of reproductive development of many forage species, which affects forage quality indirectly by decreasing leaf production and increasing production of more stem material in grasses, resulting in higher NDF concentrations (Buxton, 1995). Photoperiod can signal the appropriate time for the transition from vegetative growth to reproductive development, modify the rate of reproductive growth once established, and trigger changes in the rate of leaf area expansion and of dry-matter production which are not necessarily related to reproduction (Hay, 1990; Buxton et al.,1995). Buxton et al., (1995) noted that both day length and solar intensity affect morphology, growth, flowering, and maturity of forages, and that at the appropriate photoperiod level, plant development changes from vegetative growth to reproductive phase. The reproductive expression is enhanced as the photoperiod length is increased. The author gave example by citing Heide (1985) who noted that average stem height of timothy grass increased three-fold as photoperiod increased from 8 to 24 hours. Apart from the effects on flowering, long photoperiods cause high forage quality because of greater photosynthetic activity, which in turn increases soluble sugars that dilute the NDF. Again, long photoperiods usually alter plant morphology, increase yield, increase shoot/root ratios, decrease leaf/stem ratio, and dilute CP in herbage (Buxton et al., 1995). It is apparent that environmental conditions in spring and summer enhanced pasture productivity and quality. Development of a balanced diet of silvopastoral systems requires sound understanding of the trends in herbage DM yield and forage quality throughout the year. Trend similar to the model presented in Figure 4 of the present study was obtained in other studies conducted in other temperate regions (Demanet et al., 2015), though the scale of the changes may differ because of differences in site conditions. Benavides et al. (2009) observed that several studies conducted under deciduous trees have shown seasonal variation in pasture yield because of leaf fall. A study conducted by Douglas et al. (2006) over a 3-year period estimated that the average biomass accumulation of swards beneath a stand of Populus spp. at 25–100 stems ha-1 and aged 8–11 years was 23% less than open pasture. Benavides et al. (2009) noted that this figure nevertheless varied seasonally with differences being greater in spring, summer and early autumn and more similar during the leafless period in late autumn and winter (Benavides et al., 2009). In a trial conducted in north-west Spain, with similar climatic conditions to the United Kingdom, Rozados-Lorenzo et al. (2007) compared pasture production under three evergreen species with that under three broadleaved species. They reported that higher yields were achieved under the broadleaved species because of their lack of foliage at the beginning of spring promoting increases in pasture production. Furthermore, the result of the distance from each grazing exclusion cage to the nearest tree in the present study agrees with studies of canopy gaps in which direct relationship has been found to exist between gap size/canopy gap level and amount of light penetration (Dey and MacDonald, 2001; Garrett et al., 2004; Minckler 1961). Dey and MacDonald (2001) reported that gaps have little effect on available light when openings are smaller than 0.04 ha and larger than 0.4 ha in size. Similarly, most studies of light intensity measurement that emphasised the reading of light intensity at the centre of the gap have recognised the decreased availability of light when moving from gap centre towards the edge (Minckler 1961). Garrett et al. (2004) observed that for the silvopasture practice and growth of forages, the uneven distribution of light, such as may be created with gaps, is not desirable, that it is crucial to understand the important role slope and aspect may play in determining appropriate residual densities when thinning a forest stand. They further noted that south-facing slopes that naturally receive greater solar exposure should, logically, have higher densities of trees than north-facing slopes that are predisposed to less direct sunlight (Garrett et al., 2004). Light intensity may not be the only limiting factor influencing pasture yield and quality in this study. The addition of extra N from red alder, an N-fixing tree, would certainly have exerted significant influence on the pastures. Other influences such as moisture and nutrient competition, allelopathic effects and smothering are also at work under the trees. In a trial to compare the impacts of shade duration on pasture production with deciduous and evergreen tree species Power et al. (2001) reported that at low levels of shade (<40%), pasture relative yields under a nitrogen-fixing tree, Acacia melanoxylon, were greater than relative yields under corresponding levels of artificial shade. They further noted that for a deciduous tree species to be effective in modifying pasture yield during the leaf-free period, they must be leaf-free for longer than 4 months. The usual practice is to present forage overstory/understory models as linear relationships (Joyce and Mitchell, 1989; Mitchell and Bartling, 1991). However, some studies indicate that the true relationship over the life of a silvopasture is most likely curvilinear with little effect until tree canopy exceeds 30–50% coverage (Krueger, 1981; Joyce and Mitchell, 1989), followed by a fast decline in understory production as tree canopies merge. The linear model employed in this study made it possible to quantify the trend of daily, monthly, seasonal and locational averages of pasture growth and dry matter yield over time with high level of significance (p ≤ 0.0001). The frequency of grazing is determined by variations in the seasonal growth of pasture (Holmes et al., 2002). Demanet et al., 2015 and Holmes et al., 2002 observed that under optimal conditions of use, rotation lengths may vary between 15 and 35 days in spring and 25 and 90 days, in autumn and winter. Mathematical expression of forage production enables the prediction of periodic growth of pastures, which are useful in the determination of stocking rate and the development of feed balances, adjustment of diets to regulate demand and supply processes, conservation programs and surpluses. General information on the feed quality of a range of typical forages and animal dietary requirements for ruminant animals are presented in Appendices 1 and 2 (AFRC 1993; NRC, 1996), respectively. Forage quality is a direct reflection of the ability of a given forage to meet the nutrient needs of the consuming animal. The concentrations of the mixed forage quality parameters and the dietary maintenance requirements for sheep are within the acceptable range. Again, forage quality is a direct reflection of an animal’s ability to consume, digest, and assimilate essential nutrients contained within the feed. It has been estimated that about 50-75% of this ability of a given forage to meet the nutrient needs of the consuming animal is related to intake, 25-50% is related to digestibility, and 5-15% is related to metabolic efficiency. Dry matter represents everything contained in a feed sample except water; this includes protein, fibre, fat, minerals, etc. In practice, it is the percentage of the feed that is not water (moisture). DM increases with plant age when harvested. The lower the DM, the more moisture is present, and the lower is the nutrient density in the fresh feed. Also, high moisture may decrease the storing quality of a feed (through moulding) unless it is made into silage. When fresh forages and grasses make up the bulk of the diet, a large amount of water is consumed, which could limit intake of energy and protein sources. The crude protein content of a feed sample represents the total nitrogen (N) in the diet, which includes a mixture of true proteins, amino acids, nitrate, and non-protein nitrogen, such as urea and ammonia in a forage. Because N is an integral part of any amino acid, non-protein nitrogen has the potential to be utilized for protein synthesis by rumen microorganisms. The protein in a forage is important since protein contributes energy, and provides essential amino acids for rumen microbes as well as the animal itself. Some protein fractions are more digestible than others, but in general the higher the protein level, the more digestible is the feed. The more protein that comes from forage, the less supplemental protein is needed. However, most nutritionists consider energy value of forages to be more important than CP. The dietary CP requirement of sheep and lamb are 9 -12% and 11 – 14%, respectively (NRC, 1996) (Appendix 2). Across all treatments and seasons, the concentration of CP in herbage in this study exceeds the requirements of growing, finishing and lactating sheep (Table 2). The ADF value refers to the percent of the least digestible parts of cell wall components of cellulose, lignin, silica, insoluble CP, and ash. ADF content increases as the plant matures and is generally higher in legumes than grasses of the same age. Lignin is indigestible, whereas cellulose can be digested by the rumen microbes or bugs. Generally, ADF has been used to predict digestibility and thus energy content of a forage. The lower the ADF content, the higher the digestibility and the higher the energy value of a forage. Forages with higher ADF are lower in digestible energy than forages with lower ADF, which means that as the ADF level increases, digestible energy levels decrease. For any given sample, ADF will be lower than NDF content and the difference between the two reflects the amount of hemicellulose present. In the present study, mean dietary ADF values exceeded recommended levels of 20 – 25% for sheep (Table 2 and Appendix 2) indicating that voluntary feed intake was not limited by low fibre content. The NDF value is the percent of total fibre in the feed containing all cell wall components including cellulose, hemicellulose, lignin, silica, insoluble CP, and ash. It is the fibre in the diet that stimulates rumination, chewing, and saliva production. The NDF of a forage is inversely related to the amount that animals are able to consume; thus, as the NDF content of a forage increases, the amount of the forage a ruminant will consume generally decreases. As a result, NDF is often used in formulas to predict the dry matter intake. In the present study, mean dietary NDF values exceeded recommended levels of 25 – 35% for sheep (Table 2 and Appendix 2) indicating that voluntary feed intake was not limited by low fibre content. The ME content is the energy in feed, minus energy in faeces, urine and methane that arise from digestion (Waghorn 2007). In ruminants, the ME of a diet is calculated from the heat of combustion of feed eaten, minus the heat of combustion of faeces, urine and methane, derived from the feed eaten. It is the useful energy made available through the process of digestion and the value is expressed as a proportion of the dry matter (MJ/kg). Metabolisable energy of forage is a superior measure to dry matter for estimating animal production potential, because it represents the energy available to the animal for maintenance and production. For a given feed source, the ME declines as the level of feeding increases, due to variation in the amounts of energy lost in faeces, urine and methane. The dietary ME requirement of sheep is 8 -10% (Appendix 2). Across all treatments and seasons, the concentration of ME in herbage in this study exceeds the requirements of growing, finishing and lactating sheep (Table 2). From the results of this study, reducing tree stocking density to 100-stems ha-1 or below will likely maximise the amount of solar radiation reaching the understory pasture, allow pasture to persist most of the rotation and enhance sward botanical composition. 5. CONCLUSION This study established that thinning opened the canopy, and decreased canopy area of alder in a silvopastoral system resulting in increased pasture production by increasing transmitted light. Pasture DM production increased with increasing solar transmission as well as with increasing distance between each grazing exclusion cage and the nearest tree (Figures 4, 5 & 6). Changes in the concentration and availability of CP, ADF, NDF and ME with solar transmission can be explained by changes in pasture production. The variability in the pasture productivity and quality parameters observed may also have been influenced by climatic conditions, especially precipitation and air temperature, solar transmission intensity and duration (photoperiod), and leaf fall. Light intensity and climatic conditions may not be the only limiting factors influencing pasture yield and quality parameters in this study. The addition of extra N from red alder, an N-fixing tree, would certainly have exerted significant influence on the pastures. Other influences such as moisture and nutrient competition, allelopathic effects and smothering are also at work under the trees. The forage quality parameters, CP, ADF, NDF and ME followed similar trend as the DM yield, deceasing with reduction in solar transmission, canopy gap size and temperature. In addition, concentrations and availability of these parameters were greater in with-leaves than in without-leaves growing seasons in response to variation of photoperiod (the duration of sunshine/day length) in the United Kingdom. These results demonstrated that level of exposure to solar radiation (the level of shade) was a significant factor determining the productivity and quality (nutritive values) of pastures, and that canopy closure is the critical factor to manage if pasture production is to be maintained at an economic level. These findings can be used by farmers to optimise the spacing of alder-stands for multiple objectives, such as soil conservation, pasture grazing, stocking rate and wood fuel. Both pruning trees or removing trees can decrease canopy closure, reduce light competition, improve the grazing productivity of the understorey pasture, allow the tree leaves to be used as livestock fodder, and improve wood quality. Deciduous tree species with suitable architecture and tolerance to specific site limitations will allow more light to the pasture, particularly when leafless. From the results of this study, reducing tree stocking density to 100-stems ha-1 or below will likely maximise the amount of solar radiation reaching the understorey pasture, allow pasture to persist most of the rotation and enhance sward botanical composition. In conclusion, the use of thinning to reduce the density of red alder to 100-stem ha-1 in a silvopastoral system increased understory pasture production and nutritive value by increasing transmitted light. The practical significance of these results is that reduced competition for light after thinning the alder trees could improve the grazing productivity of the understorey pasture. . 6. ACKNOWLEDGEMENTS The completion of this research study was made possible with the assistance, encouragement and guidance of the following people and organisations to whom I owe my sincerest gratitude: Foremost, I owe my deepest gratitude to my supervisors, Dr. James Walmsley and Dr. Mark Rayment, for their sustained guidance, wisdom, thoroughness, support, patience and commitment over the many years of my field research and during the writing up process despite their many other academic commitments. Both supervisors were instrumental in the development of concepts and understanding of silvopastoral agroforestry systems that enabled this study. I am also very grateful to other staff and fellow students of the School of Environment, Natural Resources and Geography (SENRGY) at Bangor University who played significant roles during my studies. My special thanks will go to Ian Harris, Mrs. Llinos Hughes, Mark Hughes, and all other staff at the Bangor University research farm at Henfaes (the site of the Silvopastoral National Network Experiment) for providing me with useful information on the management of the research farm and for their support, assistance and patience during my data collection process and laboratory works. My greatest thanks go to the Farm Woodland Forum for awarding me the Lynton Incoll Memorial Scholarship that enabled me to present some of my PhD research chapters at the 2014 Farm Woodland Forum annual meeting in Devon, UK. I am equally grateful to all the researchers, staff, students and volunteers who have contributed to the huge body of knowledge that has been generated by the various UK’s Silvopastoral National Network Experiments. I am deeply indebted to the Nigerian Tertiary Education Trust Fund (TETFUND) for sponsoring my PhD research programme in agroforestry under its Academic Staff Training and Development Programme as well as to the Management of Chukwuemeka Odumegwu Ojukwu University (COOU) (Former Anambra State University), Anambra State, Nigeria for granting me the leave to undertake this research programme. Last but certainly not least, the encouragement I received from my friends and family was phenomenal, particularly my spouse and children, for their love, support, understanding, and extreme patience over the period of my research programme. Finally, I give gratitude to God for giving me enough spirit and strength to surmount the various obstacles encountered in making the completion of my research studies a reality. 7. REFERENCES 1. AFRC (1993). ‘Energy and protein requirements of ruminants. An advisory manual prepared by the AFRC Technical Committee on Responses to Nutrients. CAB International: Wallingford, UK, 159 pp. 2. Anderson, L. J., Brumbaugh, M. S., & Jackson, R. B. (2001). Water and tree–understory interactions: a natural experiment in a savanna with oak wilt. Ecology, 82(1), 33-49. 3. Awal, M. A. (2008). Assessment of gap fraction by hemispheri-cal photograph in oil palm plantation. Suranaree Journal of Science and Technology, 233-241. 4. Ball, D. M., Collins, M., Lacefield, G. D., Martin, N. P., Mertens, D. A., Olson, K. E., & Wolf, M. W. (2001). Understanding forage quality. American Farm Bureau Federation Publication, 1(01). 5. Belesky, D. P. (2005). Growth of Dactylis glomerata along a light gradient in the central Appalachian region of the eastern USA: I. Dry matter production and partitioning. Agroforestry systems, 65(2), 81-90. 6. Belsky, A. J. (1992). Effects of grazing, competition, disturbance and fire on species composition and diversity in grassland communities. Journal of Vegetation Science, 3(2), 187-200. 7. Belsky, A. J., & Amundson, R. G. (1992). Effects of trees on understorey vegetation and soils at forest-savanna boundaries in East Africa. Nature and dynamics of forest-savanna boundaries. Chapman and Hall, London, 353-367. 8. Belsky, A. J., Mwonga, S. M., & Duxbury, J. M. (1993). Effects of widely spaced trees and livestock grazing on understory environments in tropical savannas. Agroforestry systems, 24(1), 1-20. 9. Benavides, R., Douglas, G. B., & Osoro, K. (2009). Silvopastoralism in New Zealand: review of effects of evergreen and deciduous trees on pasture dynamics. Agroforestry Systems, 76:327–350. 10. Brockway, D. G., Outcalt, K. W., Tomczak, D. J., & Johnson, E. E. (2002). Restoring longleaf pine forest ecosystems in the Southern United States. Restoration of boreal and temperate forests, 445. 11. Brunner, A. (2002). Hemispherical photography and image analysis with hemIMAGE and Adobe Photoshop, Horsholm, Denmark. 12. Burrows, W. H., Carter, J. O., Scanlan, J. C., & Anderson, E. R. (1990). Management of savannas for livestock production in north-east Australia: contrasts across the tree-grass continuum. Journal of Biogeography, 503-512. 13. Buxton, D. R., Mertens, D. R., Moore, K. J., Boyd, L. J., & Oldfield, J. E. (1995). Forage Quality for Ruminants: Plant and Animal Considerations11This paper includes data presented by the senior author to the 28th Pacific Northwest Animal Nutrition Conference, Boise, ID, October 26–28, 1993. The Professional Animal Scientist, 11(3), 121-131. 14. Chiavarella, T. A., Dove, H., Leury, B. J., & Simpson, R. J. (2000). Diet selection by sheep grazing Phalaris aquatica L. pastures of differing water-soluble carbohydrate content. Crop and Pasture Science, 51(6), 757-764. 15. Counts, G. E., & Radloff, H. D. (1979). The potential of infrared analysis in forage evaluation. In Proceedings of the Annual Meeting. American Society for Animal Science Western Section. 16. De Montard, F.X., H. Rapey, R. Delpy and P. Massey. (1999). Competition for light, water and nitrogen in an association of hazel (Corylus avelana L.) and cocksfoot (Dactylis glomerata L.). Agrofor. Sys. 43:135–150 17. Deinum, B., Van Es, A. J. H., & Van Soest, P. J. (1968). Climate, nitrogen and grass. 2. The influence of light intensity, temperature and nitrogen on vivo digestibility of grass and the prediction of these effects from some chemical procedures. Netherlands Journal of Agricultural Science, 16, 217-223. 18. Demanet, R., Mora, M. L., Herrera, M. Á., Miranda, H., & Barea, J. M. (2015). Seasonal variation of the productivity and quality of permanent pastures in Andisols of temperate regions. Journal of soil science and plant nutrition, 15(1), 111-128. 19. Devendra, C., & Ibrahim, M. (2004). Silvopastoral systems as a strategy for diversification and productivity enhancement from livestock in the tropics. In International Symposium on Silvopastoral Systems (2004, Yucatán, México). The Importance of Silvopastoral System in Rural Livelihoods to Provide Ecosystem Services. Mannetje, Lt (pp. 10-24). 20. Devkota, N. R,, Wall, A. J,, Kemp P. D., Hodgson J (2001). Relationship between canopy closure and pasture production in deciduous tree based temperate silvopastoral systems. Proceedings of the XIX International Grasslands Congress, Brazil, 11–21 February 2001, Pp 652–653 21. Dey, D. C., & MacDonald, G. B. (2001). Overstorey manipulation. Regenerating the Canadian Forest: Principles and Practice for Ontario. Fitzhenry & Whiteside Limited, Markham, Ontario, 157-175. 22. Dey, D. C., & Parker, W. C. (1997). Overstory density affects field performance of underplanted red oak (Quercus rubra L.) in Ontario. Northern Journal of Applied Forestry, 14(3), 120-125. 23. Dodd, M. B., McGowan, A. W., Power, I. L., & Thorrold, B. S. (2005). Effects of variation in shade level, shade duration and light quality on perennial pastures. New Zealand Journal of Agricultural Research, 48(4), 531-543. 24. Douglas, G. B., Walcroft, A. S., Hurst, S. E., Potter, J. F., Foote, A. G., Fung, L. E., … & Van Den Dijssel, C. (2006). Interactions between widely spaced young poplars (Populus spp.) and introduced pasture mixtures. Agroforestry Systems, 66(2), 165-178. 25. Ducherer, K., Bai, Y., Thompson, D., & Broersma, K. (2013). Thinning of a ponderosa pine/Douglas-fir forest in south-central BC: impacts on understorey vegetation. Journal of Ecosystems and Management, 14(1). 26. Ehrenreich, J. H., & Crosby, J. S. (1960). Forage production on sprayed and burned areas in the Missouri Ozarks. Journal of Range Management, 13(2), 68-70. 27. Engelmark, O., Sjöberg, K., Andersson, B., Rosvall, O., Ågren, G. I., Baker, W. L., … & Ennos, R. A. (2001). Ecological effects and management aspects of an exotic tree species: the case of lodgepole pine in Sweden. Forest Ecology and Management, 141(1), 3-13. 28. Esquivel-Mimenza, H., Ibrahim, M., Harvey, C. A., Benjamin, T., & Sinclair, F. L. (2013). Standing herbage biomass under different tree species dispersed in pastures of cattle farms. Tropical and Subtropical Agroecosystems, 16(2). 29. Gardner, F. P., Pearce, R. B., & Mitchell, R. L. (1985). Physiology of Crop Plants. Iowa State Univ. USA. Int. Rice Res. Notes, 27, 11-12. 30. Garrett, H. E., Kerley, M. S., Ladyman, K. P., Walter, W. D., Godsey, L. D., Van Sambeek, J. W., & Brauer, D. K. (2004). Hardwood silvopasture management in North America. In New Vistas in Agroforestry (pp. 21-33). Springer Netherlands. 31. González-Hernández, M. P., Silva-Pando, F. J., & Jiménez, M. C. (1998). Production patterns of understory layers in several Galician (NW Spain) woodlands: seasonality, net productivity and renewal rates. Forest Ecology and Management, 109(1), 251-259. 32. Hale, S. E., & Edwards, C. (2002). Comparison of film and digital hemispherical photography across a wide range of canopy densities. Agricultural and Forest Meteorology, 112(1), 51-56. 33. Harrington, G. N., & John, G. G. (1990). Herbaceous biomass in a Eucalyptus savanna woodland after removing trees and/or shrubs. Journal of Applied Ecology, 775-787. 34. Hay, R. K. M. (1990). The influence of photoperiod on the dry-matter production of grasses and cereals. New Phytologist 116, 233-254. 35. Henfaes Silvopastoral System Experimental Site website: Accessed 22nd March, 2012. http://pages.bangor.ac.uk/~afs032/Henfaes/Henfaes.htm. 36. Heide, O. M. (1985). Physiological aspects of climatic adaptation in plants with special reference to high-latitude environments. In Plant production in the North: proceedings from Plant Adaptation Workshop, Tromso, Norway, September 4-9, 1983/edited by Ase Kaurin, Olavi Juntilla and Jarle Nilsen. Tromso: Norwegian University Press, c1985. 37. Holmes, C.W., Brookes, I. M., Garrick, D. J., MacKenzie, D. D. S., Parkinson, T. J., & Wilson, G. F. (2002) ‘Milk production from pasture: principles and practices.’ (Massey University: Palmerston North, New Zealand). 38. Inoue, A., Yamamoto, K., Mizoue, N., & Kawahara, Y. (2004). Effects of image quality, size and camera type on forest light environment estimates using digital hemispherical photography. Agricultural and Forest Meteorology, 126(1), 89-97. 39. Iqbal, M. (2012). An introduction to solar radiation. Academic Press, New York, New York, USA 40. Joyce, L. A., & Mitchell, J. E. (1989). Understory cover/biomass relationships in Alabama forest types. Agroforestry systems, 9(3), 205-210. 41. Kay, B.L., & O.A. Leonard. (1988). Effect of blue oak removal on herbaceous forage production in the north Sierra foothills. p. 323-328. In: T.R. Plumb (tech. coord.) Proc. Symp. Ecology, Management and Utilization of California oaks. Gen. Tech. Reu. PSW-44. Pacific SW Forest and Range Exp. %a., USDA Forest Sch., Berkeley, Calif. 42. Kellems, R.O. and Church, D.C. (1998) In: Kellems R.O. and Church D.C. (eds) Roughages, livestock feeds and feeding, 4th edn, pp. 59–84. Englewood Cliffs, NJ: Prentice-Hall, Inc. 43. Kephart, K. D., & Buxton, D. R. (1993). Forage quality responses of C3 and C4 perennial grasses to shade. Crop Science, 33(4), 831-837. 44. Knowles, R. L., Hawke, M. F., & Maclaren, P. (1997). Agroforestry research at Tikitere. NZ Forest Research Institute and AgResearch., 26 pp. 45. Knowles, S. O., Grace, N. D., Wurms, K., & Lee, J. (1999). Significance of amount and form of dietary selenium on blood, milk, and casein selenium concentrations in grazing cows. Journal of Dairy Science, 82(2), 429-437. 46. Krueger, W. C. (1981). How a forest affects a forage crop. Rangelands, 3(2), 70-71. 47. Lambert, M. G., Barker, D. J., Mackay, A. D., & Springett, J. (1996). Biophysical indicators of sustainabillty of North Island hill pasture systems. In PROCEEDINGS OF THE CONFERENCE-NEW ZEALAND GRASSLAND ASSOCIATION (pp. 31-36). 48. Lane, G. A., & Fraser, K. (1999). A comparison of phenol and indole flavour compounds in fat, and of phenols in urine of cattle fed pasture or grain. New Zealand journal of agricultural research, 42(3), 289-296. 49. Lin, C. H., McGraw, M. L., George, M. F., & Garrett, H. E. (2001). Nutritive quality and morphological development under partial shade of some forage species with agroforestry potential. Agroforestry Systems, 53(3), 269-281. 50. Marquis, R. J. (1988). Phenological variation in the neotropical understory shrub Piper arielanum: causes and consequences. Ecology, 69(5), 1552-1565. 51. Marten, G. C., Shenk, J. S., & Barton, F. E. (1989). Near infrared reflectance spectroscopy (NIRS): Analysis of forage quality. Agriculture handbook (USA). 52. Mayland, H. F., Shewmaker, G. E., Harrison, P. A., & Chatterton, N. J. (2000). Nonstructural carbohydrates in tall fescue cultivars: Relationship to animal preference. Agronomy Journal, 92(6), 1203-1206. 53. McCall, D. G., & Bishop-Hurley, G. J. (2003). A pasture growth model for use in a whole-farm dairy production model. Agricultural Systems, 76(3), 1183-1205. 54. McClaran, M. P., & Bartolome, J. W. (1989). Effect of Quercus douglasii (Fagaceae) on herbaceous understory along a rainfall gradient. Madrono, 141-153. 55. Mead, D. J. (2009). Biophysical interactions in silvopastoral systems: a New Zealand perspective. In Actas del 1º Congreso Nacional de Sistemas Silvopastoriles (pp. 14-16). 56. Minckler, L. S. (1961). Measuring light in uneven-aged hardwood stands. U.S. Forest Service Central States Experimental Station Technical Paper Number 184. 57. Mitchell, J. E., & Bartling, P. N. S. (1991). Comparison of linear and nonlinear overstory-understory models for ponderosa pine. Forest Ecology and Management, 42(3), 195-204. 58. Monk, C.D. & Gabrielson, F.C.J. (1985). Effects of shade, lit¬ter, and root competition on old-field vegetation in South Carolina. The Bulletin of the Torrey Botanical Club 112: 383-392. 59. Mosquera-Losada, M. R., McAdam, J. H., Romero-Franco, R., Santiago-Freijanes, J. J., & Rigueiro-Rodróguez, A. (2009). Definitions and components of agroforestry practices in Europe. In Agroforestry in Europe (pp. 3-19). Springer Netherlands. 60. Nair, P. K. R. (1987). Agroforestry Systems in major ecological zones of the tropical and subtropics. Paper prepared for the ICRAF-WMO International Workshop on the Application of Meteorology to Agroforestry Systems Planning and Management, Nairobi, 9-13, February, 1987. 61. National Research Council. (1996). Nutrient Requirements of Beef Cattle. 7th rev. ed. Natl. Acad. Press, Washington, DC. 62. Neel, J. P. S., Feldhake, C. M., & Belesky, D. P. (2008). Influence of solar radiation on the productivity and nutritive value of herbage of cool‐season species of an understorey sward in a mature conifer woodland. Grass and Forage Science, 63(1), 38-47. 63. Norris, K. H., Barnes, R. F., Moore, J. E., & Shenk, J. S. (1976). Predicting forage quality by infrared replectance spectroscopy. Journal of animal science, 43(4), 889-897. 64. Nworji, M. J. (2017). Physical and Bioeconomic Analysis of Ecosystem Services from a 65. Silvopasture System. PhD thesis, SENRGY, Bangor University, Wales. 66. Nygaard, P. H., & Ødegaard, T. (1999). Sixty years of vegetation dynamics in a south boreal coniferous forest in southern Norway. Journal of Vegetation Science, 10(1), 5-16. 67. Ong CK, Black CR, Marshall FM and Corlett JE (1996). Principles of resource capture and utilization of light and water. In: Ong CK and Huxley PA (eds) Tree-Crop Interactions: A Physiological Approach. CAB International, pp 73–158. Wallingford, UK 68. Peri, P. L., Lucas, R. J., & Moot, D. J. (2007). Dry matter production, morphology and nutritive value of Dactylis glomerata growing under different light regimes. Agroforestry systems, 70(1), 63-79. 69. Peterken, G. F. (2001). Ecological effects of introduced tree species in Britain. Forest Ecology and Management, 141(1), 31-42. 70. Power, I. L., Dodd, M. B., & Thorrold, B. S. (2001). Deciduous or evergreen: does it make a difference to understorey pasture yield and riparian zone management? In Proceedings of The Conference-New Zealand Grassland Association (pp. 121-126). 71. Pratchett, D. A. V. I. D. (1978, August). Effects of bush clearing on grasslands in Botswana. In Proceedings of the First International Rangelands Congress. Society for Range Management, Denver, Colorado (pp. 153-159). 72. Rao, M. R., P.K.R. Nair and C.K. Ong. (1998). Biophysical interactions in tropical agroforestry systems. Agroforestry Systems 38. p. 4-49. 73. Rich, P. M. (1990). Characterizing plant canopies with hemispherical photographs. Remote sensing reviews, 5(1), 13-29.7 74. Rozados-Lorenzo, M. & Gonzalez, Pilar & Silva-Pando, Francisco. (2007). Pasture production under different tree species and densities in an Atlantic silvopastoral system. Agroforestry Systems. 70. 53-62. 10.1007/s10457-007-9032-2. 75. Sander I.L. (1979). Regenerating oaks with the shelterwood system. pp. 54–60. In: Proceedings Regenerating Oaks in Upland Hard- wood Forests, The 1979 J.S. Wright For. Conf., Purdue Univ., Lafayette, IN. 76. Sharrow, S.H. (1999). Silvopastoralism: Competition and facilitation between trees, livestock, and improved grass-clover pastures on temperate rainfed lands. pp.111-130. In: Buck, Louise, J.P. Lassoie, and E.C.M. Fernandes (Eds.). Agroforestry in sustainable agricultural systems. CRC Press/Lewis Publishers, Boca Raton, FL. 77. Shelton, H. M., Humphreys, L. R., & Betello, C. (1987). Pastures in the plantations of Asia and the pacific. Tropical Grasslands, 21(4), 159. 78. Shenk, J. S., & Westerhaus, M. O. (1994). The application of near infrared reflectance spectroscopy (NIRS) to forage analysis. Forage quality, evaluation, and utilization, (foragequalityev), 406-449. 79. Shenk, J. S., Westerhaus, M. O., & Hoover, M. R. (1979). Analysis of Forages by Infrared Reflectance1. Journal of Dairy Science, 62(5), 807-812. 80. Sibbald, A. R. (1999). Silvopastoral agroforestry: soil–plant–animal interactions in the establishment phase. Grassland science in Europe, 4, 133-144. 81. Sibbald, A. R., Griffiths, J. H., & Elston, D. A. (1994). Herbage yield in agroforestry systems as a function of easily measured attributes of the tree canopy. Forest ecology and Management, 65(2-3), 195-200. 82. Sibbald, A.R. and Sinclair, F.L., (1990). A review of agroforestry research in progress in the UK. In Agroforestry Abstracts CABI (Vol. 3, No. 4, pp. 149-164). 83. Sibbald, A.R., Eason, W.R., McAdam, J.H. and Hislop, A.M., (2001). The establishment phase of a silvopastoral national network experiment in the UK. Agroforestry systems, 53(1), pp.39-53. 84. Sigurdsson, B. D., Magnusson, B., Elmarsdottir, A., & Bjarnadottir, B. (2005). Biomass and composition of understory vegetation and the forest floor carbon stock across Siberian larch and mountain birch chrono sequences in Iceland. Annals of Forest Science, 62(8), 881-888. 85. Skinner, R. H., Corson, M. S., & Gilmanov, T. G. (2009). Simulating gross primary productivity of humid-temperate pastures. Agronomy journal, 100(3), 801-807. 86. Smith, M. A., & Whiteman, P. C. (1983). Evaluation of tropical grasses in increasing shade under coconut canopies. Experimental Agriculture, 19(02), 153-161. 87. Stone, W. E., & Wolfe, M. L. (1996). Response of understory vegetation to variable tree mortality following a mountain pine beetle epidemic in lodgepole pine stands in northern Utah. Plant Ecology, 122(1), 1-12. 88. Stuth, J., Jama, A., & Tolleson, D. (2003). Direct and indirect means of predicting forage quality through near infrared reflectance spectroscopy. Field Crops Research, 84(1), 45-56. 89. Teklehaimanot, Z. and Mmolotsi, R.M. (2007). Contribution of red alder to soil nitrogen input in a silvopastoral system. Biology and Fertility of Soils, 43: 843-848. 90. Teklehaimanot, Z., & Sinclair, F. L. (1993). Establishment of the silvopastoral network experiment site, Henfaes, Bangor. In Agroforestry Forum (Vol. 4, No. 2, pp. 18-21). 91. Teklehaimanot, Z., Jones, M. and Sinclair, F.L., (2002). Tree and livestock productivity in relation to tree planting configuration in a silvopastoral system in North Wales, UK. Agroforestry Systems, 56(1), pp.47-55. 92. Uresk, D. W., & Severson, K. E. (1998). Response of understory species to changes in ponderosa pine stocking levels in the Black Hills. The Great Basin Naturalist, 312-327. 93. Waghorn, G. C. (2007). What is dietary metabolisable energy? In Proceedings of The Conference-New Zealand Grassland Association Vol. 69, p. 153. 94. Walker, J. W., Heitschmidt, R. K., De Moraes, E. A., Kothmann, M. M., & Dowhower, S. L. (1989). Quality and botanical composition of cattle diets under rotational and continuous grazing treatments. Journal of Range Management, 239-242. 95. Ward, R. G., Smith, G. S., Wallace, J. D., Urquhart, N. S., & Shenk, J. S. (1982). Estimates of intake and quality of grazed range forage by near infrared reflectance spectroscopy. Journal of Animal Science, 54(2), 399-402. 96. Wilson, J. R., Hill, K., Cameron, D. M., & Shelton, H. M. (1990). The growth of Paspalum notatum under the shade of a Eucalyptus grandis plantation canopy or in full sun. Tropical grasslands, 24(1), 24-28.