Temporal Variations in Chemical Composition, In Vitro Digestibility, and Metabolizable Energy of Plant Species Browsed by Goats in Southern Mediterranean Forest Rangeland

Simple Summary Mediterranean forest rangelands constitute essential feed resources for grazing goats. The objective of this study was to evaluate the temporal variations in chemical composition, in vitro digestibility, and metabolizable energy of browsed plant species by goats on forest rangelands of the Southern Mediterranean of northern Morocco. Overall, the nutritive value of the selected plant species was highest in spring and then steadily decreased through the summer and autumn. Most of the selected plant species present high levels of crude protein than the minimum required level for maintenance. This study provides a valuable and useful database to elaborate the seasonal grazing and feeding management plan for goat herds. Abstract Forest rangelands contribute largely to goat diets in the Mediterranean area. Information about browsed plant quality is essential for adequate feeding management. The purpose of this study was to evaluate the temporal changes in chemical composition and in vitro digestibility of the main plant species selected by goats in the Southern Mediterranean forest rangeland during two consecutive years; these were very contrasted (dry and wet). The browsed species were composed of herbaceous, eleven shrubs, and four tree species. Overall, large variability in chemical composition, in vitro organic matter digestibility (IVOMD), and metabolizable energy (ME) was observed among species, grazing season (spring, summer, and autumn), and years within each species. Crude protein (CP) content varied from 60 to 240 g/kg dry matter (DM). The fiber fractions, except for Quercus suber, increased significantly by advancing maturity. Due to the water stress, the lignin level presented a higher value during the spring of the dry year. Condensed tannin (CT) content varied from 2 to 184 g/kg DM. CP, IVOMD, and ME showed a negative correlation with lignin and CT. Based on the results presented herein, it is concluded that the nutritive value of the browsed plant species was highest in the spring and lowest during the summer and autumn of both studied years. With a good grazing management strategy, the selected plant species by goats could guarantee high-quality feeding resources throughout the year.


Introduction
Mediterranean forests are composite landscapes of shrubs and trees, which constitute essential dietary resources for domestic ruminants. They also play a very important role in sustaining biodiversity [1] and provide multiple ecosystem services to local people for millennia [2]. These woodlands are characterized by heterogeneous and diversified flora [3].
Lavandula stoechas L., Myrtus communis L., Phillyrea media L., Pistacia lentiscus L., and Rubus ulmifolius Schott.), trees (Quercus spp. (inclusive of Q. canariensis L., Q. ilex L., and Q. suber L.), and Olea europaea var. sylvestris (Mill) Lehr), and herbaceous (mainly Anthemis cotula L., Brachypodium distachyon L., Bromus rigidus Roth, Calamintha nepeta (L.) Kuntze, Cynodon dactylon (L.) Pers., Eryngium tricuspidatum L., Lythrum junceum Banks and Sol., Rumex bucephalophorus L.). According to Chebli et al. [11], these plant species are listed as the main dietary components of goats in Southern Mediterranean forest rangelands. Grazing in the forest rangelands of northern Morocco is practiced only over three seasons (spring, summer, and autumn). For the winter, pasture access is very limited; goats do not browse in forest pastures and graze only in fallow land around the goat shed, which explains the exclusion of this season from the study. The present research studied the chemical composition, in vitro digestibility, and metabolizable energy of all browsed species by goats. Samples were collected by hand-plucked simulation of each ingested part of the plant species similar to those consumed by goats. Diet composition and hand-plucked simulation are briefly summarized here and described fully in Chebli et al. [11]. The study concerned the botanical composition of each consumed part of plant by goats. The sampling was undertaken in the last month of each studied season (May, August, and November). Representative hand-plucked samples per plant species (a mixture of leaves and green tender stems) and herbaceous, similar to those consumed by goats, were imitated seasonally. For the thorny species, we used scissors to clip the selected parts. For herbaceous species, they were mixed into a single group because of difficulty to identify all ingested species by goats during grazing and their low selectivity. For shrubs and trees, the samples were harvested per species in special bags, with three replications, and transported to the laboratory for analysis.

Laboratory Analysis
Chemical analyses and in vitro digestibility studies were performed on three independent samples of the hand-plucked forage of each ingested plant species by goats during each grazing season of two consecutive years.

Chemical Analysis
Collected samples were dried at 40 • C in a ventilated oven to minimize changes in tannins content and activity until reaching constant weight [24], and then milled with a sieve mesh size of 1 mm for analysis. Dry matter (DM), organic matter (OM), crude protein (CP), and ether extract (EE) were analyzed according to the Association of Official Analytical Chemists [25]. The neutral detergent fiber (NDF) was estimated using the Mertens [26] method with α-amylase and sodium sulfite. Acid detergent fiber (ADF) was determined according to method 973.18 of AOAC [27]. Acid detergent lignin (ADL) was determined by the solubilization of cellulose with sulfuric acid, according to Robertson and Van Soest [28]. All fiber extractions were performed using ANKOM 200 Fiber Analyzer®(ANKOM Technology, Fairport, NY, USA). The NDF, ADF, and ADL values were expressed inclusive of residual ash. Condensed tannins (CT) were predicted by Porter et al. [29] method using butanol-HCl, and ferric reagents.

In Vitro Digestibility and Metabolizable Energy
In vitro dry matter (IVDMD) and organic matter (IVOMD) digestibility were performed using DAISYII Incubator®(ANKOM Technology, Fairport, NY, USA) as described by Mabjeesh et al. [30]. This device is essentially based on the in vivo simulation of digestion [31]. The rumen liquor for incubation was collected from five goats at a communal slaughterhouse, as described by El Otmani et al. [32]. These goats grazed in similar forest rangeland of the study area. The collected ruminal fluid was maintained in a thermos at 39 • C to keep rumen microflora alive. A weight of 0.5 g of each sample was placed in ANKOM filter bags (F57) and was put in jars (24 bags/jar). The inoculum, mixture containing 4/5 volume of artificial saliva, and 1/5 of rumen liquor was added in jars and incubated at 39.5 • C for 48 h. IVDMD and IVOMD were estimated by quantifying residuals DM and OM comparing to incubated initial quantities.
The metabolizable energy (ME; MJ/kg DM) of each consumed plant species was calculated using the equation [27]: where DMD is the dry matter digestibility in percentage.

Statistical Analysis
Data were analyzed using SAS software®(SAS Inst. Cary, NC, USA). Chemical composition, digestibility, and ME of each plant species (n = 15) and herbaceous were analyzed using a general linear model (GLM) procedure of SAS in a factorial structure. Data were compared between seasons (i.e., spring, summer, and autumn), years (i.e., 2016 and 2017), and their interactions. Simple correlation analysis was used to establish the relationships between the chemical composition, IVOMD, and ME. The correlation plot was obtained by utilizing the "corrplot" library in the R-package [33]. For all analyses, the significance level was declared at p < 0.05. In case of significant effect, means were compared using the Tukey's test.

Results
The chemical composition, IVOMD, and ME of the browsed plant species by goats at different sampling seasons and years are given in Table 1 (shrubs) and Table 2 (trees and  herbaceous). Overall, these parameters of shrubs (n = 11), trees (n = 4), and herbaceous species varied seasonally in each studied year.
Across shrub species, all of them presented a higher DM content in summer, except for C. salviifoluis and M. communis, with a higher DM content in spring, and for R. ulmiformis, with a higher DM content (also) in autumn. The higher water content was observed either in spring (for five shrubs) or autumn (for six shrubs). This parameter was significantly affected by both studied factors (season and year) except for C. crispus, which was not affected by year. Their effects on OM of the studied shrub species were variable, A. unedo, E. arborea, L. stoechas, M. communis, and P. media, having the same OM throughout the year. The CP content varied significantly among seasons of both years (p < 0.05), except for the season effect of the dry year (2016) on C. villosa and the season effect of the wet year (2017) on A. unedo, C. crispus, C. salviifoluis, E. arborea, and P. media (p > 0.05). During both years, the highest and lowest CP concentrations were recorded in C. villosa (about 240 g/kg DM in the autumn) and A. unedo (about 60 g/kg DM in summer), respectively. The CT content ranged from 1.97 g/kg DM (summer 2017) in C. villosa to 191 g/kg DM (summer 2016) in P. lentiscus. The EE content ranged from 15.8 g/kg DM in C. crispus, to 90-101 g/kg DM in C. monspeliensis (summer and autumn) and E. arborea (spring and summer). The highest NDF and ADF levels of both years were observed in C. villosa, with 629 and 482 g/kg DM, respectively. Overall, the ADL contents showed a significant increase from spring to summer-autumn (except for P. media) with a range from 62.3 g/kg DM in R. ulmifolius (2016) to 324 g/kg DM in E. arborea (2017). All studied shrub species presented a higher IVOMD in spring of both consecutive years, except for R. ulmifolius and M. communis, with the highest IVOMD in summer-autumn and summer, respectively. The lower IVOMD were found in summer (for five shrubs) or autumn (for two shrubs), or there was no significant difference between summer and autumn (for two shrubs). The ME results showed the same trend as the IVOMD. The highest ME content was observed in L. stoechas (about 10 MJ/kg DM) browsed during spring and the lowest one (about 4.5 MJ/kg DM) in C. villosa (autumn) and E. arborea (summer) during both dry and wet years. The most notable changes due to advancing maturity were found in the CP, CT, ADL, and ME contents. Generally, the CP and ME contents decreased, and CT and ADL contents increased during spring to summer-autumn of both years.  Across trees species and for the two years, DM content was higher in summer and lower in spring, except for O. europaea, where it was the opposite. The DM content in autumn was the same as in spring for Q. suber, the same as in summer for O. europaea and Q. ilex, and significantly different from the two other seasons for Q. canariensis. Each oak tree species showed no variation of its OM content during a year, except for Q. suber in 2016, which presented a decrease over time. Q. ilex recorded the higher CP concentration in spring of the wet and dry years (99.7 and 114 g/kg DM, respectively). Q. suber had the highest, and O. europaea the lowest CT content during all studied seasons. Among tree species, the oak trees showed low EE content (about 24 g/kg DM). The high EE content was recorded in O. europaea during the summer of the wet year (131 g/kg DM). The highest NDF content was recorded for Q. suber and Q. ilex (about 550 g/kg DM), and the lowest for O. europaea (about 410 g/kg DM). The highest and lowest ADL levels were observed during autumn and spring of the wet year in Q. ilex (about 190 g/kg DM) and in Q. canariensis (103 g/kg DM), respectively. Quercus spp. had a high IVOMD significantly during spring and a low one in summer. The IVOMD of O. europaea was similar in all studied seasons during the dry year. Nevertheless, this similarity was not observed in the wet year, with a slight increase over time. The ME levels of the studied tree species varied slightly, being particularly low in Quercus spp. (about 6 MJ/kg DM) during the summer of the dry and wet years and highest in Q. canariensis (8.8 MJ/kg DM) during the two springs.
Comparatively to the two other groups, herbaceous also had a higher DM content in summer and a lower content in spring. The OM content was variable according to the year and season. It was higher in the spring and similar and lower in the other seasons. The CP concentrations recorded the highest value in the spring of 2016 and 2017 (156 and 142 g/kg DM, respectively; p < 0.001). The CT contents recorded the highest values during the summer of 2016 (4.17 g/kg DM; p < 0.05) but were similar among seasons of 2017 (with about 2.7 g/kg DM; p = 0.116). The NDF concentrations were similar among seasons of both years (p > 0.05). The ADF content was higher in summer than spring and autumn of both years (p < 0.01). The ADL concentrations were similar among seasons of 2016 (about 70 g/kg DM; p = 0.265) but increased in 2017 from spring to summer (p < 0.05). The IVOMD and ME contents were higher in spring than in autumn and summer (p < 0.001). Overall, the highest CP and ME contents were recorded in the herbaceous and the lowest in shrub and tree species. An opposite trend was recorded for CT and ADL levels.
The correlation values among the chemical composition, IVOMD, and ME from the studied forage species are presented in Figure 1. The CP showed a negative correlation with ADL, CT, and EE (p < 0.001). The ME was strongly correlated with IVOMD (p < 0.001). The NDF, ADF, and ADL contents were positively correlated with each other (p < 0.001). A negative correlation was observed between IVOMD and ME with CT, ADL, and ADF (p < 0.001), and with NDF (p < 0.05).
The correlation values among the chemical composition, IVOMD, and ME from the studied forage species are presented in Figure 1. The CP showed a negative correlation with ADL, CT, and EE (p < 0.001). The ME was strongly correlated with IVOMD (p < 0.001). The NDF, ADF, and ADL contents were positively correlated with each other (p < 0.001). A negative correlation was observed between IVOMD and ME with CT, ADL, and ADF (p < 0.001), and with NDF (p < 0.05).

Discussion
The aim of this study was to assess the nutritive value of the plant species browsed by goats and their variations throughout three grazing seasons of two years. These years appeared very contrasted regarding the mean annual rainfall, with a dry year in 2016 and a wet one in 2017. According to Papachristou et al. [34], the bulk of the grazing goats' diet includes few ligneous and herbaceous species, representing less than ten species. Ligneous species A. unedo, C. villosa, E. arborea, M. communis, P. lentiscus, and Q. suber are considered the most widespread species in the Southern Mediterranean rangelands [15,23]. As observed and described by Chebli et al. [11], plant species analyzed herein represent the all-selected diet by grazing goats in Southern Mediterranean forest rangelands. During spring of 2016, the contribution of C. monspeliensis (28.8%), C. crispus (19.8%), and C. salviifolius (17.6%) was the highest followed by L. stoechas (17.3%) and herbaceous (7%). These species contributed lowly to the diet during autumn and summer (< 3%). In the autumn and summer, the diet proportion of Quercus spp. (3-20%), M. communis (14-19.4%), P. lentiscus (8-13%), A. unedo (11-13%), E. arborea (9.5-11%), and O. europaea (2-7%) was largely significant. During spring of 2017, the contribution of C. crispus was significantly increased by 42% with the decreased rate of C. salviifolius and L. stoechas by 10 and 15%, respectively. In the autumn, the greatest increase in contribution to the diet was observed for O. europaea followed by P. lentiscus, and E. arborea. The opposite trend was observed with the diet proportion of Q. canariensis and C. villosa. In the summer, the contribution of P. lentiscus Animals 2021, 11, 1441 13 of 17 and P. media increased by 93 and 17%, respectively. On the other hand, diet contribution of A. unedo and E. arborea decreased by 35 and 17%, respectively. On average, the diet of the goats was largely composed of shrubs (64-90%) and trees (2-35%). However, the contribution of herbaceous did not exceed 8%. The contribution of trees to the diet during spring dropped from 30.3 to 3.7% and from 29.0 to 2.2% in 2016 and 2017, respectively. The diet proportion of R. ulmifolius varied from 0.01 to 3.4% [11].
The nutritive value of the hand-plucked samples, corresponding to the most tender part of the plant, appears to reflect the quality of the diet consumed by grazing goats [6]. For this study, the browsed parts of the plant by goats were analyzed, which represent a mixture of leaves, stems, and twigs.
The chemical traits of browsed species were extremely wide, which are in accordance with previous studies conducted in northern and eastern Mediterranean forest rangelands [6,[12][13][14], deciduous tropical forest [35], South African rangeland [36]. In northwestern Italy, Ravetto Enri et al. [37] reported the relevant effect of the vegetative season on chemical composition and in vitro true digestibility of four tree species selected by goats. These wide variations on the nutritional proprieties of plant species could be explained by soil fertility [15,38], environmental conditions, and stage of growth or age [14,39].
The observed mean CP level found in this study varied from 60 to 240 g/kg DM. Most of the analyzed plant species present high levels of CP than the minimum level of 70-80 g/kg DM required by microorganisms for optimum rumen functioning and feed intake in ruminant livestock [40]; a lower CP content affects negatively feed intake and digestibility [41]. In the present study, the low CP content was particularly recorded in A. unedo (from 52.7 to 70.7 g/Kg DM). In the northwest of Tunisia, a value of 55 g/kg DM in A. unedo, collected in March 1998 from the uplands of Taaref, was reported [15], which is in the range of the current results. The high proportions of mature leaves and twigs in the samples could explain the low CP level in some plant species, such as A. unedo. Overall, the average CP level was higher during spring 2016 in all species because plants contain the maximum CP content during the vegetative stage [42]. The decrease of this parameter in the summer agrees with the literature [19,43] because CP drops with the physiological maturity stage of the plant [44], which explains the negative correlation of CP with ADL and EE that increase with plant maturity as found by Ammar et al. [12] in some Spanish shrub species. As expected, the CP content was higher in C. villosa as it is a leguminous plant. According to Kokten et al. [14], leaves of C. villosa could be used as protein supplements for livestock since their CP contents are high compared to the other Mediterranean shrubs. The high protein level in C. villosa could be attributed to the ability of this plant to fix atmospheric nitrogen thanks to rhizobia associated with their nodules [12,16]. Overall, the older leaves contained less CP and more fiber than the young and tender part of the selected plant species. This statement is in consistent with other studies [6,14]. In terms of CP content, many of the woody species cover the daily maintenance requirements of grazing goat but not for milk or meat production needs, which is above 130 g/kg DM. In another environment (hills of Nepal), Khanal and Subba [45] reported a good nutritional value of leaves from most of the tree fodder species, with a minimum CP of 110 g/kg DM.
Generally, trees had a higher EE content during summer, which coincides with the maturity stage of these groups of plant species. Indeed, plant species had a higher fat content (EE) in the late physiological stage that increases with maturity [12].
According to the species and sampling season, the NDF, ADF, and ADL contents in ligneous species varied from 242 to 629 g/kg, 186 to 482 g/kg, and 70 to 322 g/kg, respectively. However, herbaceous recorded the lower lignin content (64.3-89 g/kg). Overall, these contents significantly increased by advancing maturity. The results are in line with the findings of several authors [14,16,46], who indicated that cell wall content (NDF, ADF, and ADL) augmented with maturity (cell wall lignification). All analyzed samples recorded higher ADL levels during the spring of the dry year compared to the wet year. This higher concentration during the dry year could reflect the response of plant species to water stress (rarity rains), which is associated with the increased level of tannins.
Khanal and Subba [45] reported a high ADL content in most of the fodder trees in the hills of Nepal, with values more than 100 g/kg DM. High fiber content, and lignin especially, means low free-nitrogen extract and soluble carbohydrates contents, which explain the observed negative correlations of fibers with IVOMD and ME. Ammar et al. [47] reported an increase of fiber content in parallel with a decrease of in vitro digestibility, with the maturity of mountain grasses, which confirm the negative correlation between fibers and digestibility.
The observed CT concentrations varied from 2.3 to 184 g/kg DM, showing significantly different and slightly higher values than those obtained by several authors [12,48] with shrub leaves from Northern Spain. These variations could be due in part to the difference in analysis methods. Moreover, it could be owing to the stage of growth and the sampled parts of the plants (leaves, stems, and twigs), to the season and to the nature of the sampling site [49,50]. In addition, the current study concerns the analysis of different parts of the plant selected by goats, not only their leaves.
A CT concentration of 20-45 g/kg DM has a negative effect on protein digestibility and proteolytic bacteria [51], and a concentration above 55 g/kg DM reduces the voluntary feed intake of grazing ruminants [52,53]. Thus, except for herbaceous (2.20-4.17 g/kg DM), C. villosa, L. stoechas, P. media, and O. europaea, all pastoral species had a CT content higher than this maximum level. However, even with a high CT content, the shrubs were highly consumed in spring and autumn [11]. These findings are consistent with Fomum et al. [36] and Mkhize et al. [54], who reported no correlation between CT and feed intake in goats. It could be explained by the ability of goat to balance their diet and dilute secondary compounds by consuming a mixture of plant species [55,56]. Moreover, grazing animals exposed to high CT feed could excrete more saliva richer in proline-rich proteins that has the ability to bind with CT to neutralize it [52,54]. Nevertheless, goats have a specificity compared to other ruminants that their ruminal microbiota is able to valorize feed with low nutritional values due to their cellulolytic bacteria and the tanninase activity [57]. Min et al. [51], reported that high CT concentrations reduced digestibility, which could explain the negative correlation between CT and IVOMD.
Bartolomé et al. [58], who studied the quality of forest resources in the undergrowth of the pine forests of Mallorca (Spain), reported that species, such as P. lentiscus, C. monspeliensis, and A. unedo, could be interesting as feed for goats, as they show protein levels above the minimum of maintenance, and at the same time, high digestibility. Nevertheless, they present a high content of secondary compounds, such as tannins.
In the present study, the higher values of digestibility, mainly observed in spring, are attributed to its negative correlation with ADF and ADL. Ammar et al. [12], also reported a negative correlation between fibers and digestibility of browsed leaves. Most of the pastoral species had low digestibility and, consequently, a low energy content (IVOMD < 550 g/kg; ME < 8 MJ/kg DM), except for herbaceous and some shrubs (A. unedo and Cistus spp.), as their values varied from medium (IVOMD: 550-700 g/kg; ME: 8-10 MJ/kg DM) to higher nutritional values (IVOMD > 700 g/kg; ME > 10 MJ/kg DM), especially during spring [59]. As reported by Paton [60], ME depends mainly on IVOMD, which could explain their high positive correlation. Overall, goats select species with high CP and digestibility and low fiber content basis, in accordance with the literature [34].

Conclusions
This study provides a valuable and useful database on the temporal variations in chemical composition, in vitro digestibility, and ME of the main plant species browsed by goats in the Southern Mediterranean forest rangeland. Most of these plant species showed considerable variation among grazing season. In general, the nutritive value of plant species was highest in spring, and then steadily decreased through the summer. In autumn, the nutritive value decreased, remained the same, or increased compared to summer. Most of the selected plant species presented high levels of CP than the minimum required levels for maintenance needs. All analyzed samples recorded higher lignin levels during the spring of the dry year compared to the wet year. Except for herbaceous, C. villosa, L. stoechas, P. media, and O. europaea, all analyzed species had a CT content higher than this maximum level. The high values of digestibility in spring are attributed to its negative correlation with ADF and lignin. Owing to the morphological and physiological differences between the consumed plant species, changes in chemical compositions and in vitro digestibility could be expected. Consequently, goats are forced to adapt their browsing behavior to the low-quality vegetation, typical of the Mediterranean forest. The results could be used as indicators to assess the nutritional value of the goat diets in forest rangelands throughout the year to help the grazing management strategy, and/or to eventually supplement the goats adequately, to prevent the low farmer incomes due to the animal performance decrease.
Future work would allow knowing if the practiced grazing systems in the Southern Mediterranean region could guarantee the seasonal dietary requirements of grazing goats, in terms of energy and protein, taking into account the physical and physiological conditions of grazing goats. Institutional Review Board Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available from the corresponding author on request.