Introduction

Over time, woodlands have had an important role as an animal feed supply, particularly in the tropics (Franzel et al. 2014), the Mediterranean region (Ibidhi et al. 2018), and in other dry or cold areas across the world (Nagler et al. 2015; Turunen et al. 2020). Grazing oak woodlands (GOW) is a type of silvopasture (Mosquera-Losada et al. 2009) that can be found in Spain (San Miguel 2004; Baumont et al. 2000), Portugal (Castro 2009), Greece (Pantera et al. 2018;), Italy (Pardini and Nori 2011), and in other Mediterranean regions (González-Pech et al. 2015). According to Nerlich et al. (2013), GOW is probably the oldest agroforestry practice in Mediterranean Europe as it goes back almost to the Neolithic period.

Through the ages, oak woodlands (OW) held a prominent place within the economy of these regions in Mediterranean Europe by providing firewood, charcoal, cork, fruit, extracts for medicine and by-products such as tannin, and also by offering an important grazing area for livestock (Debussche et al. 2001). However, the socio-economic changes that occurred after World War II opened the gap between agriculture and forest systems. This was accelerated by the parallel process of rapid intensification and specialization of agriculture in Europe (Kubacka et al. 2016). As a result, silvopastoral systems like GOW have been disappearing over the last decades.

Nowadays, GOW are regaining interest as they offer the chance to respond to numerous environmental concerns. They have the potential to enhance soil fertility, reduce erosion, improve water quality, enhance biodiversity, increase aesthetics and sequester carbon (José 2009; Moreno et al. 2016; López-Santiago et al. 2019). Also, GOW are useful for wildfire prevention (Ruiz-Mirazo et al. 2011) due to the livestock potential for effectively controlling shrub growth.

GOW can also help to mitigate greenhouse gas (GHG) emissions from ruminants through different pathways. Several authors (Sasu-Boakye et al. 2014; Sanz-Cobena et al. 2017) have reported the positive effects on GHG emissions by integrating livestock, crops and/or forest systems; the agroecological transition is indeed a pathway to GHG mitigation at farm level (Martin and Willaume 2016). In addition, GHG depends on the production system, for example, whether it is organic or not, as well as on the mixture of agricultural systems, i.e., the mixture of agricultural crops and livestock (Fazeni and Steinmüller 2011). Moreover, use of local feed production instead of importing feed from other regions reduces long-distance transport, thus also contributing to GHG reduction (Peyrauda et al. 2014). Finally, wildfires are recognized as a significant source of atmospheric emissions at a global scale (UNEP 1999). Despite the relatively small importance of European forest fire emissions, in a global context, the majority of these emissions come from the five EU Mediterranean countries (Barbosa et al. 2009), since approximately 85% of the total burnt area occurs in the EU Mediterranean region (San-Miguel-Ayanz and Camia 2010). Therefore, if livestock grazing is considered an efficient tool for wildfire prevention, GOW can be seen as a means to reduce GHG emissions.

Additionally, various reasons motivate the growing interest for trees and shrubs as fodder for ruminants in the Mediterranean region. They are useful sources of cheap feed for ruminant animals, especially during dry or cool seasons, when conventional forages are scarce and of low quality (Olafadehan and Okunade 2018). Browse species supply energy, protein and other nutrients (González-Pech et al. 2015). Durmic et al (2010) pointed out that the inclusion of woody plants in diets can have multiple benefits in livestock nutrition, including increased feed intake, digestibility and rumen fermentation, and reduced methane emission from ruminants (Pantera et al. 2018). In addition, such plants also provide shelter (Castro 2009) and may reduce internal parasite infestation as some of them have anthelmintic properties (Xhomfulana et al. 2009).

Oak woodlands have been studied in several ways, from productive to conservation viewpoints, taking into account the soil, carbon sequestration, physiology, firewood or timber production, silvopastoral management, etc. However, few studies have been performed about the feeding value of browse species in OW. In this context, there is a need to conduct studies focusing on to the nutritive value and digestibility of the foliage of trees and shrubs, which could guarantee the recommendation of their use in the feeding of ruminants.

This paper aims to evaluate the seasonal variation of the chemical composition and the in vitro organic matter digestibility (IVOMD) of key browse species of OW widespread in the north of Portugal. Additionally, the feeding value was compared with the requirements for goats’ maintenance and late pregnancy with regard to protein and energy, thus providing information to goat producers so that they find the best strategies to take advantage of forest resources for satisfying the nutritional requirements of their flocks.

Material and methods

Sampling

The study was carried out in the Mediterranean region of Trás-os-Montes, Northeast of Portugal, at a latitude of 41° 25′ N to 41° 48′ N, and a longitude of 1° 56′ W to 2° 12′ W, at an altitude ranging from 380 to 820 m a.s.l.. The climate of the area is Mediterranean. The mean annual temperature varies between 11.9 and 14.2 °C and the mean annual precipitation ranges from 520 to 1075 mm. January is the coldest month (mean minimum temperatures ranging from 0.7 to 2 °C) while July is the hottest month (mean maximum temperatures ranging from 27.9 to 31.6 °C). The soils are leptosols and the structure of vegetation is open in the case of Quercus suber L. and Quercus faginea Lam woodlots and quite closed in the Quercus rotundifolia L. and Juniperus oxycedrus L. The prevalent shrubs that comprise the understory vary according to the type of woodlots (mainly altitude and temperature) more details can be seen in Castro and Fernández Nuñez et al. 2014).

For this study, eight OW were selected according to the dominant tree species and their altitude distribution. The species studied were the trees Q. faginea (deciduous), and perennifolios tree Q. suber, Q. rotundifolia, and J. oxycedrus; and the deciduous shrubs Cytisus scoparius L., Cytisus multiflorus (L'Hér.) Sweet, and the perennifolius shrubs Genista falcata Brot., Lavandula stoechas L., and Cistus ladanifer L.. Current year foliage samples were composed of leaves in the case of trees and terminal twigs with their leaves in the case of shrubs.

Samples of the different species were taken throughout the seasons, April in spring, July in summer, November in autumn and February in winter. Q. faginea was sampled only during the leaf production periods. Three composed samples (from five randomly selected plants) were collected from each species per local and season. Samples were air-dried to constant weight in a fan-assisted oven at 60 °C for 48 h, and they were ground in a mill through a 1-mm sieve.

Nutritional analysis

Ash and crude protein (CP) contents were assessed following the methods of AOAC (2000). Neutral detergent fibre (NDF), acid detergent fibre (ADF), and acid detergent lignin (ADL) were determined according to detergent procedures (Robertson and Van Soest 1981; Van Soest et al. 1991). Organic matter (OM) was calculated by difference of the mineral residue. The in vitro organic matter digestibility (IVOMD) was assessed using the two-stage technique of Tilley and Terry (1963) modified by Marten and Barnes (1980).

The dry matter digestibility (DMD) values were used to estimate digestible energy (DE) using the regression equation reported by Fonnesbeck et al. (1984): DE (Mcal kg−1) = 0.27 + 0.0428 × DMD%. Then, DE values were converted to metabolizable energy (ME) using the formula reported by Khalil et al. (1986): ME (Mcal kg−1) = 0.821 × DE.

The ME requirements for the maintenance of goats was obtained according to 424.2 kJ/kg BW0.75 (NRC 1981); in late pregnancy, NRC (1981) included a fixed increase in ME of 5.48 MJ/day above maintenance requirements. The protein requirements were calculated as 7% and 11% of dry matter intake (DMI) for maintenance and late pregnancy, respectively (NRC 1987). The estimate was done for an average animal, considering the mean weight of an adult goat to be approximately 45 kg (the maximum capacity of intake is 1,1Kg DM and 1,3Kg DM for maintenance and late pregnancy, respectively (NCR 1987)).

Statistical analysis

Variables were analysed using two way ANOVA (PROC GLM procedure) following the model Yij = m + Ai + Bj + eij, where Yij is the studied variable; m the variable mean; Ai: season; Bj: species, eij is the error. The LSD test was carried out for subsequent pairwise comparisons (p < 0.05; α = 0.05). The statistical software package SAS (2001) was used for all these analyses.

Results

Chemical composition

The chemical composition (OM, CP, NDF, ADF, and ADL) and IVOMD of leaves and twigs from trees and shrub species at each sampling season are shown in Table 1. Significant differences (p < 0.05) were found among species and seasons.

Table 1 Chemical composition (gkg−1 DM) and IVOMD (%) of shrubs and trees species at different seasons
Full size table

The CP content of the studied species varied significantly, being particularly high in C. scoparius in spring (195.9 g kg−1 DM) and low in J. oxycedrus in summer (41.6 g kg−1 DM). The CP content of legume species (C. scoparius, C. multiflorus, and G. falcata) was generally higher than in other species. The great part of the species showed a significant increase of CP content in spring, but a few have shown a different pattern, for instance C. ladanifer presented a higher CP content in autumn and winter, and L. stoechas in spring and autumn.

The NDF content ranged from 316.7 g kg−1 DM to 632.5 g kg−1 DM, and ADF from 246.0 g kg−1 DM to 468.0 g kg−1 DM; with both parameters being at their lowest in L. stoechas spring samples and at their highest in C. multiflorus summer. Fibre contents (NDF and ADF) tend to be higher in summer for the great majority of species however there are exceptions, for example (e.g.), C. ladanifer. The ADL content varied significantly with the season, being particularly high in winter (210.5 g kg−1 DM) in the case of C. ladanifer.

The highest IVOMD values were observed in spring for L. stoechas (71.34%) and C. scoparius (70.87%) while the lowest were recorded in summer for C. ladanifer (33.02%).

In general, shrubs species showed high contents of CP (111.4 g kg−1DM) and IVOMD (49.8–51.38%) and lower fibre contents than the tree species.

Figures 1 and 2 compare the seasonal variation of CP and ME, respectively for the shrubs and tree species studied, as well as, the requirements for the maintenance and late pregnancy of adult goats with 45 kg of body weight. In this respect, large differences were found among the species under analysis, however, the highest and the lowest values of CP and ME, were reported in spring and summer, respectively. The estimated ME showed significant differences for the different seasons and it also varied significantly between life forms.

Fig. 1
figure 1

Temporal variations of the crude protein (CP) content in the different shrubs and trees studied. Different capital letters indicate a significant difference between species in the same season and different lower-case letters indicate a significant differences between season in the same species (p < 0.05)

Full size image
Fig. 2
figure 2

Temporal variations in the metabolizable energy (ME) content in the different shrubs and trees studied. Different capital letters indicate a significant difference between species in the same season and different lower-case letters indicate a significant differences between season in the same species (p < 0.05)

Full size image

Discussion

Future climate change projections for Portugal show that the mean seasonal precipitation is expected to decrease substantially in all seasons (Soares et al. 2015), while an increase is expected in the duration and intensity of drought. In the Mediterranean region, an important consequence of these projections is that grasslands could decrease their capacity to provide good forage in summer (Sebastià 2007). In this context, farmers with small landholding and greater dependency on natural resources face an important challenge. It is known that browse species may be complementary to the annual herbaceous vegetation and that fodder from these woody species is available at different times of the year considering their differences in phenology and growth activity (Boufennara et al. 2012). The annual diet of goats where this study was conducted is composed of more than 50% of browse species in the summer, autumn and winter, a pattern that only changes in spring (Castro and Fernández-Núñez 2012). The trees and shrub species analysed in this study are consumed in significant quantities, by small ruminants grazing in OW, particularly goats, which shows the importance of woody plants in extensive feeding systems (Mancilla-Leytón et al. 2012; Ravetto et al. 2020).

However, the seasonal chemical composition of these browse species (trees and shrubs) vary within a wide range (e.g., CP content 41.6–195.9 g kg DM−1 and IVOMD 33.02–71.34%), resulting in a highly variable fodder interest of the species in a particular season. These results were consistent with those previously reported by other authors (Castro et al. 2014; Castro and Fernández-Núñez 2018). Also, our results are in line with those found in similar environmental conditions by Martínez (2009), who ranked Q. ilex as low quality feed, due its high lignin content and low digestibility; J. oxycedrus as medium quality given that its cellular contents and digestibility are relatively high for a woody plant and Q. faginea as a high quality resource. According to the author, this last species has a high protein content and high digestibility. Manousidis et al (2018) found results consistent with ours for Juniper's leaves, namely low contents of CP ranging from 3.8 to 6.6%. They also reported that the CP content and the dry matter digestibility decreased over time whereas NDF, ADF and ADL contents were not stable across seasons. Concerning aromatic shrubs, C. ladanifer could be classified as low quality feed, due to their very low CP content (64.7–79.7 g kg−1 DM) and low digestibility (33.02–41.81%), while L. stoechas might be a medium to high quality resource depending on the time of year (e.g., CP content 63.7–111.1 g kg DM−1 and IVOMD 41.80–71.34%). The leguminous shrubs are the most interesting set of fodder species. They have high CP (102.2–195.9 g kg−1 DM), medium IVOMD (50.77–70.87%) and estimated ME (8.02–11.35 MJ kg), which qualifies them as a good quality feed source.

The CP contents of most of the species studied were higher than the minimum level of 70–80 g kg−1 DM required for optimum rumen function and feed intake in goats (Van Soest 1994). Only Q. rotundifolia, J. oxycedrus and C. ladanifer showed different trends. Concerning the digestibility of OM, which is one of the main factors determining the nutritive value of fodder, trees species and C. ladanifer showed values below 50% of IVOMD, suggesting that they may not be able to meet the needs for animal maintenance (Arzani et al. 2006). However, an underestimation of the digestibility of material from shrubs and trees by the in vitro method has been noted since the late 80′s (Nastis and Malechek 1988) and may be related to the presence of substances like tannins, which cause a negative bias in the in vitro estimates (Sidahmed et al. 1981). For instance, Nastis (1982) found a difference of 20% between the in vitro and in vivo digestibility of kermes oak (Q. coccifera) (50% and 70%, respectively).

Other studies provided different results regarding the proximate composition and digestibility for the same species and comparable dates. For example, Ammar et al. (2005) found lower CP, similar NDF and ADF and much higher lignin contents in Q. suber in Tunisia. This is probably related to the faster lignification process in tropical and arid environments than in temperate environments (Kenneth and Hans-Joachim 2001). Also, in the case of C. ladanifer, Guerreiro et al. (2016) found higher CP contents, comparable cell wall content and lower digestibility values than those found in the present study. Mancilla Leyton et al. (2012) reported higher ME values for C. ladanifer (7.9 MJ kg−1 DM) and lower ones for L. stoechas (5.8 MJ kg−1 DM), unlike what was found in our study.

Temporal fluctuations in the chemical composition and IVOMD can be commonly observed in many browse species with similar trends, as well as in other than Mediterranean environments (Ravetto et al. 2020). With a few exceptions such as C. ladanifer, J. oxycedrus, the CP and IVOMD were significantly higher in spring than in all other seasons for both life forms, while the fibre contents followed an opposite pattern. Overall, the temporal decrease of nutritional values related to phenology advancement is more pronounced in herbaceous than in shrubs or tree species, since they have a longer growing season and are able to maintain their nutritive values for longer (Arzani et al. 2006). The trees species studied showed a lower forage quality than the shrubs, as they have a higher fibrous content (NDF, ADF, ADL) and lower CP and IVOMD. This agrees with findings reported by several authors (Castro and Fernández-Núñez 2018). However, it is not a generalizable pattern, since the nutritional value is mainly related to the species or family and not so much to the life form. As reported by Cook and Stubblendieck (1986; cited by Arzani et al. 2006), the chemical content of plant species may differ because of an inherent ability to withdraw certain nutrients from the soil and to concentrate them in tissues. For instance, Ravetto et al. (2020) reported CP contents of temperate fodder tree species comparable to the one of herbage growing in similar alpine environments across the vegetative season. Conversely, in dry environments, other authors reported higher CP contents, in vitro DM and estimated ME in the foliage of deciduous fodder species than in local grass or hay (Gardiner et al. 2013). Also, in a Zagros Mountain region in Iran, Arzani et al. (2006) found higher CP contents and ME in shrub species than in grasses and forbes. As is known, the chemical composition of plants varies with species, phenology advancement (maturity stage) and type of soil (Ammar et al. 2005). Among the tree species, J. oxycedrus showed the lowest CP and highest IVOMD percentage; Q. suber and Q.rotundifolia, showed the lowest IVOMD percentage and Q. faginea, the highest CP and a medium IVOMD percentage.

With regard to the capacity that the woody species studied have to meet the energy needs of livestock (goats with 45 kg of body weight), almost all the species cover the energy needs for maintenance (7.4 MJ day−1, NRC 1987). Only C. ladanifer, Q. suber and Q. rotundifolia reported inadequate ME levels. Depending on the season, only a few shrub species, such as C. scoparius in spring can cover the ME needs in late pregnancy (12,85 MJ day−1, NRC 1987). Considering the protein requirements as 7% and 11% of DMI for maintenance and late pregnancy, respectively (NRC 1987), only leguminous shrubs (C. scoparius, C. multiflorus, G. falcata) cover the needs of both physiological states. Inadequate CP levels were observed in the C. ladanifer, Q. rotundifolia and J. oxycedrus, even for maintenance, while the other woody species only cover the needs of the less demanding physiological states.

On the other hand, diet composition varies over time according to the availability of plant species in the habitat and to their nutritional value. Also, regarding food selection, goats seem to establish a balance between swiftness and best consumption quality (Duncan and Gordon 1999 quoted by Mancilla-Leytón et al. 2012), depending on their nutritional requirements at the time. Therefore, it is expected that livestock choose their forage resources to improve the nutritive value of their diets. None of the species studied are consumed as sole feed given that OW are patchy environments and a flock may well come across extremely diverse vegetation within a single day (Baumont et al. 2000). Hence, OW seems adequate to meet the maintenance requirements of goats, as these animals ingest a diversity of plants throughout their grazing circuits and seek to balance their choices according to their needs.

Conclusion

The chemical composition and digestibility of plant species occurring in OW vary amongst species and seasons, resulting in a highly variable fodder interest of the species in a particular season. In general, the woody species studied cover the energy and protein requirements of goats’ maintenance. However, only leguminous species are able to meet the protein needs in late pregnancy. The energy requirements for this physiological state can only be met by C. scoparius, C. multiflorus and L. stoechas, and only in spring or autumn. The tree species studied showed lower forage quality than the shrubs, as they have a higher fibrous content (NDF, ADF, ADL) and lower CP and IVOMD. In light of this, the conclusion drawn is that OW could/should be integrated in feeding systems, thus reducing the food scarcity periods, providing important information to farmers on how to find the best strategies to achieve the feed requirements of their flocks and improving farmers’ management production systems in a climate change scenario.