Genetic diversity is essential for sustaining biodiversity over time given that a greater genetic diversity increases species likelihood to respond to environmental changes and disturbances (Banks et al., 2013; Hoffmann and Parsons, 1991; Rice and Jain, 2013). Land transformation and fragmentation caused by livestock breeding is affecting ecosystems worldwide (FAO, 2019), although for some areas are reported positive effects (García-Fernández et al., 2019; Shapira et al., 2019), while for others negative effects were observed (Fischer and Lindenmayer, 2007; Hobbs et al., 2008; Yirdaw et al., 2017). Particularly, it is well documented that livestock grazing in arid ecosystems leads to severe reduction in vegetation cover, abundance and richness, affecting their functioning (Li et al., 2017; Paruelo and Aguiar, 2003; Tadey, 2006). Moreover, long-term grazing and its negative impact on ecosystem dynamics can have unpredictable evolutionary consequences on plant communities (Hanke et al., 2014; Herman and Sultan, 2011; Tadey and Farji-Brener, 2007). However, the extent of these effects on plant genetic diversity has received limited attention, and even less the effect of grazing on genetic diversity through generations (Fu et al., 2005; Wu et al., 2010).

Large herbivores can directly or indirectly modify plant fecundity and gene flow during pre-dispersal and post-dispersal stages, which might lead to changes in genetic diversity and structure of consumed plant populations (Grant, 2010; Mulder, 1999). Even if herbivores can help plant species dispersal by zoochory (Bessega et al., 2017; Campos et al., 2011; García-Fernández et al., 2019), reductions in effective population size by consumption of individuals (either by killing or damage) and/or decreasing reproduction (by consumption of flowers and/or fruits) might result in less breeding individuals contributing to the next generation. This process may lead to the reduction of the species’ gene pool and the restriction of gene flow, eroding over time, genetic diversity and increasing divergence among populations (Couvet, 2002; Lanfear et al., 2014; Leimu et al., 2006; Young et al., 1996). Small populations are subjected to inbreeding effects as a consequence of the increased likelihood of mating with closely related individuals (Heinken and Weber, 2013). However, inbreeding may also be increased through genetic drift, changes in allele frequencies caused by cumulative effects and bottleneck, which is a kind of genetic drift where the gene pool of a population is drastically reduced by a stochastic event (Luikart and Cornuet, 1998; Luikart et al., 1998; Piry et al., 1999). Usually, after these events inbreeding may be deepened by limited reproduction, decreasing offspring viability and fitness (i.e. inbreeding depression) (Hedrick and Kalinowski, 2000). Consequently, reductions in genetic variation could severely decrease the adaptive capacity of plant populations, affecting their ability to cope with future natural and/or anthropogenic impacts in a changing world (Cornuet and Luikart, 1996; Frankham et al., 2015; Luikart and Cornuet, 1998; Luikart et al., 1998; Williams, 2001).

From an ecological perspective, a limited number of available mates as a consequence of smaller population size may also affect future generations through reduced pollination and reproduction (Ågren, 1996). The number of mates (i.e. flowering plants) may affect pollination attraction and foraging behavior (Lehtilä and Strauss, 1997; Tadey, 2015, 2008), leading to lower pollen flow within population (Lehtilä and Strauss, 1997). When plants are less attractive, pollinators tend to reduce visitation frequency or to leave to other rich-resources sites (Strauss et al., 1996). Alternatively, when attractive plants are distant from each other, pollinators tend to stay longer within a single plant (Tadey, 2015). These behaviors restrict pollen flow and increase self-pollination, reducing pollen quality (i.e. endogamic pollen) (Aguilar et al., 2008; de Vere et al., 2009; Herrera, 2000; Labouche et al., 2017; Mothershead and Marquis, 2000; Oostermeijer et al., 1994; Tadey, 2008). These last processes may result in inbreeding depression expressed as lower reproduction success by reduced fruit and seed set, germination rates and/or greater seed abortion and seedling mortality (Abrahamsson et al., 2013; de Vere et al., 2009; González-Varo et al., 2012; Oostermeijer et al., 2003; Reed and Frankham, 2003; Waser and Price, 1989). Additionally, plant damage may reduce nutritional resources for new reproductive structures, leading to lower flower production and, consequently, lower plant fitness (Angeloni et al., 2011; Charlesworth and Charlesworth, 1987; Keller and Waller, 2002; Lehtilä and Strauss, 1997; Mothershead and Marquis, 2000; Ramsey and Vaughton, 2016). These negative repercussions of large exotic herbivores on plant fitness may reduce seed banks and seed dispersal, compromising natural revegetation and aggravating genetic diversity losses along generations (Pol et al., 2014; Sternberg et al., 2003; Tadey and Souto, 2016; Tadey, 2007; Trlica and Rittenhouse, 1993).

Particularly in dry environments, large herbivores can be important seed dispersal agents through defecation and trampling, and promoters of seedling establishment for some species by modifying micro-environmental conditions (Aschero and García, 2012; Aschero et al., 2016; Bessega et al., 2017; Campos et al., 2011). In some regions it was observed that livestock may have positive effects on species and ecosystems, such as enhancing seed dispersal and soil quality (Abdalla et al., 2018; Manley et al., 1995;Root-Bernstein et al., 2017; Shapira et al., 2019; Tarin et al., 2016). However, previous works in arid ecosystems showed that livestock may also indirectly affect native dispersers (such as rodents and birds) affecting their interaction with plant species (Marone et al., 2008; Milesi et al., 2002; Tadey, 2019). Additionally, pollen dispersal may also be indirectly affected by livestock grazing. Positive effects were reported for pollination community at intermediate levels of grazing (Tadey, 2015); nevertheless, the increment in pollinators richness and abundance were not reflected in an increase in plant reproduction (Tadey and Souto, 2016; Tadey, 2007). Moreover, a greenhouse study reported increased seedling vigor in some species with increasing livestock density; however, their response is still unknown under natural arid conditions (Tadey and Souto, 2016). Contrarily, other studies found that grazing by large herbivores reduced plant cover, richness, reproduction, germination (Cerda et al., 2012a; Tadey and Farji-Brener, 2007; Tadey, 2006) and even genetic diversity of dominant shrubs (Larrea divaricata and Larrea cuneifolia) (Souto and Tadey, 2018). Therefore, although it seems that livestock may have positive effects in arid regions, finally those effects are counteracted by other negative effects. This underlines the importance of evaluating direct and indirect causal relationships between variables to detect the strongest effect.

Vegetation in the world has evolved in association with herbivory developing adaptive traits. However, in arid regions, such as Patagonian Monte Desert, current vegetation traits suggesting coadaptation are notably weak since browsing reduces reproduction (Cerda et al., 2012b; Tadey and Farji-Brener, 2007; Tadey and Souto, 2016; Tadey, 2020, 2006). This may be partially explained by a weak co-evolution history with herbivores since the large extinct megafauna cited for Monte Desert, or nearby, were mostly described as grazers, found in moister areas associated water and grasses (e.g. records from saline lagoons and endorheic basins), suggesting low association with xeric shrub vegetation (Fernández et al., 2016; Mehl et al., 2019; Tadey, 2020; Villavicencio et al., 2016). In addition, persistent herbivores, such as guanaco (Lama guanicoe, Müller), with a generalist diet including shrubs, are currently scarce in Monte Desert regions and probably were in the past as well (Burgi et al., 2012; Flores et al., 2012). The low abundance of herbivores may also explain a weak coevolution in this xeric vegetation. Desert climatic hostility also modeled vegetation evolution, influencing plant growth which needed to survive under water scarcity and high evapotranspiration rates. Therefore, desert vegetation is characterized by xerophytic lifeforms with slow growth rates and low vegetative reproduction, mainly relying on sexual reproduction to persist over time (Fisher and Turner, 1978). Moreover, as other drylands of the world, Patagonian Monte Desert is an arid ecosystem that is increasingly affected by livestock grazing, which accentuates severe desertification problems, with important socioeconomic impacts on local communities (Maestre et al., 2009; Reynolds et al., 2007). Although it is known that rotation and transhumance practices may mitigate grazing effect (Baied, 1989; Barnes et al., 2008; Root-Bernstein et al., 2017), these grazing management tools are infrequent in Monte Desert (Auffret et al., 2015; García-Fernández et al., 2019; Olea and Mateo-Tomás, 2009). As a consequence of this, serious environmental degradation is observed in most of Patagonian drylands (Ares et al., 2003; INTA, 2011). Estimating the long-term effects of grazing is still a challenge and needs deeper investigation. We studied grazing effects on genetic diversity and fitness in parental and offspring cohorts of Prosopis alpataco Phill., a xerophytic shrub highly consumed by exotic large herbivores in Patagonian Monte Desert. This shrub can be dispersed by livestock, had good germination rates and is abundant in the studied region which makes it suitable to estimate genetic diversity on offspring. To fulfil our aim, we tested the following hypotheses: (1) Plant damage by herbivore consumption of photosynthetic tissue, flowers and fruits reduces the availability of plant resources precluding the allocation to new reproductive structures decreasing plant fitness by reducing the formation of fruits and seeds in terms of quality and quantity. So, as large herbivores increase in abundance, we expect a decrease in plant size (height and diameter), seed weight, seed set and seedling emergence rate; (2) Herbivore consumption of breeding individuals causes lower population reproduction success, eventually decreasing genetic diversity and increasing genetic structure of standing populations, deepening this effect through cohorts. As large herbivores increase in abundance, we expect a decrease in mean number of alleles, effective number of alleles, observed and expected heterozygosity, and percentage of polymorphic loci, and a worsening of these effects in the next cohort. Additionally, we expect that grazing will intensify inbreeding within plant parental populations and that these effects would be deepened in the offspring populations. We assess relative genetic diversity losses between parental and offspring generation (ΔGV) calculated for each genetic parameter separately; expecting greater losses of genetic diversity between generations as grazing intensity increases. Additionally, we tested a hypothetical causal model using D-separation methodology to estimate grazing direct and indirect effects on parental and offspring fitness and genetic diversity. We predicted that individuals with higher genetic diversity will produce heavier seeds, higher seed set and seedling emergence rate, along with lower seedling mortality.

Long lasting disturbances on natural populations may cause genetic diversity erosion through generations. We tested whether exotic large herbivores produce transgenerational negative effects on phenotypic and genetic traits of desert vegetation. As hypothesized, we found that grazing mostly affected parental cohort fitness and its genetic diversity and structure, consequently affecting the offspring. Genetic diversity was lost from parental to offspring cohort along the grazing gradient, although the effect was relatively lower at intermediate levels of grazing. As expected, we observed inbreeding effects on all offspring populations, while parental populations showed inbreeding on moderate and highly grazed rangelands. In addition, there was a negative impact of grazing on offspring survival. Finally, we summarized and discussed the causal relationships among grazing intensity, fitness and genetic diversity parameters between cohorts.

We observed that large herbivores fed on P.. alpataco plants and this occurred in all the studied rangelands, with a mean browsing index for all parental populations of 67 ± 30 % (Fig. 3b). This contradicts previous results in the same region that found that browsing index to other species increased linearly with stocking rate (Tadey, 2006). However, this reflects the species high palatability that even in slightly grazed rangelands, with low encounter probability; there were several highly consumed individuals (>40% browsed in average, Table A.3, Appendix A). We found a significant decrease in plant performance in terms of seed set and seedling emergence along the grazing gradient. These results could be explained by the plant damage effects on reproduction, such as the reduction in resources available to invest in reproduction or by direct consumption of flowers, which in turn could affect pollination services, and/or fruits. Inbreeding may also cause an increase in aborted seeds decreasing seed set (Castilla et al., 2019). Aguirrebengoa et al. (2018) found that ungulates significantly reduced seedling emergence rate, attributing it to lower viability by a reduction in seeds carbon content, which could even explain the higher seedlings mortality. On the other hand, large herbivores did not affect plant height and crown diameter; this apparent discrepancy with the high damage observed could be explained by several facts. First, this spiny shrub sprouts radially generating large crowns up to 8 m in diameter, which generally deters grazing. However, in highly grazed rangelands resources for large herbivores become scarce and this plant species is highly palatable. In this context, herbivores can manage to destroy a crown and open a trail inside of it, foraging the plant center and giving it a ring-shape, without varying its total diameter (Fig. 3a–b). Large herbivores usually consume P. alpataco’s small young leaves, flowers and fruits, and less frequently, stems and spiny branches. We suggest the addition of a measurement of green biomass to estimate grazing damage in future studies. Moreover, we observed no variation in seed weight along grazing gradient, suggesting that seed weight is maintained, probably at the expense of reduced seed quantity, through maternal effects mechanisms (Agrawal, 2001). Nevertheless, our results suggest that grazing is affecting, at least partially, reproduction success through plant damage of a xerophytic shrub that depends mainly on sexual reproduction to persist over time.

Fig. 3.

Prosopis alpataco crowns. a: Oval-shaped crown slightly grazed. b. Ring-shaped crown foraged from plant center.

(0.33MB).

Consequences of long lasting disturbances on plant abundance and reproduction can affect ecosystem dynamics in different ways that, through time, end up impacting on their genetic diversity and structure (Davies et al., 2016). As mentioned above, genetic diversity is essential for species persistence over time. We observed different negative effects of large herbivores on genetic diversity between cohorts. In parental cohort, we found that grazing reduced genetic diversity, which suggests that large herbivores are eliminating standing individuals (i.e. genotypes) through consumption or trampling. P. alpataco is a long-lived, self-incompatible species (Harris, 2003), and probably sampled individuals existed before the introduction of exotic livestock in the region. Therefore, it is also probably that livestock had been selecting (or rejecting) some genotypes, leading to within population gene pool homogenization. Accordingly, we found increasing inbreeding within and divergence among parental populations with grazing intensity. Moreover, genetic drift may lead to further reductions in genetic diversity (Frankham, 1995). Consequently, a reduced population size and/or cross-reproduction could restrict gene flow, depaupereting the gene pool across cohorts. In this way, few individuals and/or fewer breeding individuals contribute to seed production to future generations and, therefore, part of the population’s genetic diversity is lost (Souto and Tadey, 2018). These detrimental effects of grazing on genetic variation may jeopardize population’s ability to persist in stressful or changing environments across generations (Frankham et al., 2002).

Grazing effects on the offspring cohort were more complex. Genetic diversity of parental to offspring populations is expected to vary due to sexual reproduction; greater outcrossing in parental population would lead to greater genetic diversity in the offspring. However, we observed that as large herbivores increased in abundance, offspring genetic diversity showed a negative quadratic response. Even when genetic diversity was lower in offspring than in parental generation in all studied rangelands (i. e. negative ΔHo values), this loss was lower at intermediate grazing intensities (Fig. 2). These results may allude to the “intermediate-disturbance hypothesis” (Connell, 1978; Wilkinson, 1999), suggesting that in offspring cohort, an intermediate level of disturbance would affect genetic diversity loss less than extreme grazing intensities. This might be explained by pre-zygotic mechanisms, mainly during pollination. A recent study reported that intermediate levels of grazing did not impair bee communities (Shapira et al., 2019). Similarly, in the same rangelands included in this study, pollinator richness and abundance were higher at intermediate than at extreme grazing intensities (Tadey, 2015). This might possibly be the result of a reduction in the abundance of dominant species and the increase of rare ones by reduced competition; this in turn would improve pollination services, favoring cross-pollination at intermediate grazing intensities (Castilla et al., 2019; Lázaro et al., 2016; Lehtilä and Strauss, 1997; Tadey, 2015; Winfree et al., 2007). In accordance with this, we observed lower levels of inbreeding by kinship in the offspring at intermediate grazing intensities. So higher pollen dispersal might be preventing inbreeding depression (Gandon and Michalakis, 1999; Gandon, 1999; Hamilton and May, 1977; Taylor, 1988). To some extent, intermediate grazing could also benefit P. alpataco establishment through a genetic structure determinant, such as seed dispersal (Bessega et al., 2017), by seeded droppings and subsequent trampling and burial. Additionally, since P. alpataco has high germination rates once it escapes herbivory it could easily establish in bare soil where there is less competition (Campos et al., 2011; Paez and Marco, 2000). Nevertheless, there is little information on the role of livestock in the dispersion process of P. alpataco (Vega Riveros et al., 2011).

At the extremes of grazing intensity reproductive success may be limited by the level of kinship among intraspecific neighbors, as generally nearby individuals may exhibit higher relatedness levels (Castilla et al., 2019; Loveless and Hamrick, 1984), and by pollinator behavior that could be affecting plant-pollinator interactions, such as distances of pollen dispersal at which the genetic neighborhoods develops (Breed et al., 2015; Kolb, 2008; Lloyd et al., 2018; Tadey, 2015). In the case of low grazing intensities, where there are patches with more plants density and all species bloom almost simultaneously, there is a greater competition between pollinator species and between plants for pollinators (Tadey, 2006). This could be reducing pollen quantity (pollen limitation) and quality through inbred pollen dispersal due to territorial pollinator foraging behaviors (i.e. foraging only on a reduced group of individuals/patches) (Castilla et al., 2019; Tadey, 2008, 2015). This is supported by the higher inbreeding between cohorts observed in low grazing intensity rangelands. Meanwhile, highly grazed rangelands have lower plant density, cover and species richness (Tadey, 2006), leading to more isolated flowering plants and patches increasing pollinator foraging costs (Kolb, 2008; Tadey, 2020, 2007). Consequently, grazing would be indirectly enhancing self-pollination by limiting pollinator movement among patches because pollinators tend to visit more flowers per plant to reduce foraging costs (Harder, 2014; Jennemten, 1988; Kolb, 2008; Sih and Baltus, 1987; Tadey, 2015). In turn, self-pollinated endogamic pollen may result in non-viable seeds, particularly in those plant species with a self-incompatible reproduction system (Harris, 2003), as observed in our results. Therefore, large herbivores could be affecting pollen dispersal and hence gene flow within plant populations differentially along the grazing gradient, whereas at intermediate levels of grazing better pollination services might decrease inbreeding through generations.

In general, genetic diversity losses are associated with a reduction in fitness due to inbreeding depression, both at the individual and population level, which further reduces the population size (Bouzat, 2010). We observed negative grazing effects on parental cohort, both on phenotypic and genotypic variation, with consequent transgenerational effects on their offspring. We found grazing effects on seedling mortality exclusively through indirect mechanisms. Aguirrebengoa et al. (2018) found some similar results, showing direct negative effects of ungulates on offspring cohort, increasing their mortality rate, ascribing it to a reduction in seed quality by parental plant damage. According to our model, grazing negatively affected parental genetic diversity and fitness, which in turn affected offspring genetic diversity, increasing seedling mortality. Among the three proposed indirect pathways that increase mortality rate (Fig. 1b), the pathway including parental fitness (i.e. PCA scores summarizing three variables, viable seeds, seed weight and seedling emergence) through offspring genetic diversity was the strongest; and as above described, we observed a strong decrease in plant performance along the grazing gradient. Grazing may restrict pollination with consequences on plant species demographic structure and/or spatial and temporal distribution of sexually compatible plants, limiting gene flow by decreasing the likelihood of pollen deposition from unrelated plants (Tadey, 2008) and affecting offspring mortality by inbreeding depression, as we observed here. Usually, higher genetic diversity is associated with an increase in plant fitness (Bouzat, 2010). We observed, in our model, that parental genetic marginally increased their fitness (PCA scores), but greater offspring genetic diversity had negative correlations with seedling mortality. Additionally, we observed a marginal and negative association between parental and offspring genetic diversity. This may result from the high variability observed on genetic diversity (PCA scores) along the grazing gradient (Fig. A.4, Appendix A), different type of responses from each generation to grazing intensity (linear vs. quadratic) or similarities in genetic diversity between both cohorts. Suggesting that one generation might not have enough time to see significant changes in genetic diversity or that a larger sample size is needed. However, our results also showed significant losses of genetic diversity between cohorts with long-term effects. The path analysis evidenced that the main mechanism by which grazing affected offspring genetic diversity and survival was through parental fitness. The negative and quadratic associations among seed set and seedling emergence of parental cohort on the effective number of alleles and expected heterozygosity of offspring cohort showed that populations with intermediate production of viable seeds and emergence rates produce offspring with higher genetic diversity (Ae and He). This could be limiting offspring gene pool, whereas at intermediate grazing intensities, with intermediate viable seeds production and seedlings emergence, there is greater genetic diversity in offspring, possibly due to greater richness and abundance of pollinators (i.e. better pollination services). It would be useful to confirm these results with another type of more sensitive and more polymorphic molecular markers than isoenzymes, such as microsatellites. Our results also highlight the need of measuring reproduction success both in quantity and quality, in terms of cross-breeding with non-inbred pollen, seedling vigor and survival as we found that evidenced with the greatest seedlings emergence occurred in populations with less inbreeding (FIS). Since parental fitness was highly associated to parental genetic diversity, we may assume that parental genetic diversity is also important for offspring persistence over time. In this sense, indirect grazing impacts on offspring fitness are expected to be much greater in arid field conditions, compromising natural establishment and exacerbating genetic diversity losses over generations (Pol et al., 2014; Sternberg et al., 2003; Tadey and Souto, 2016; Tadey, 2007; Trlica and Rittenhouse, 1993). In consequence, a decrease in species fitness can decrease population growth rate and potentially decrease their sizes with greater losses of genetic diversity over time, consequently affecting their adaptive potential to environmental and/or anthropogenic changes (Hendry et al., 2018; Lande, 1994; Lynch et al., 1995).

Therefore, the maintenance of genetic diversity needs to be encouraged for the long-term sustainability of unmanaged dryland rangelands (Souto and Tadey, 2018), where extreme environmental conditions make them more susceptible to disturbance than other more humid ecosystems. Some of the measures that could mitigate population reduction are the inclusion of sustainable management practices such as livestock rotation, transhumance practices and the application of ecological restoration. In the case of the latter, increasing biological and genetic diversity should also consider the type of species with different growth rates, such as early or pioneer species, which may be less susceptible to grazing and have a faster recovery rate than late species, have high reproductive success, provide high availability of resources for pollinators, consider their phenology and seed dispersal.

Genetic diversity enables species to persist over time by overcoming disturbance effects. This study shows that grazing by large exotic herbivores can change gene flow in different pathways depending on the intensity, with consequences on genetic diversity and structure. We observed that offspring populations showed significant inbreeding effects. We found that genetic diversity loss between cohorts involved complex, non-linear and indirect mechanisms. Generally, genetic diversity losses were associated with a reduction in parental fitness, which in turn affected offspring genetic diversity and increased seedlings mortality. Hence, studies along gradients of disturbance and understanding relationships among variables in a causal framework are important to detect these processes. Our results highlight the relevance of estimating intra and inter-generational direct and indirect of long-term effects of grazing on plant population dynamics. Finally, the introduction of large exotic herbivores in unmanaged rangelands are seriously affecting dryland vegetation structure, with substantial consequences on ecological processes and ecosystem services. Livestock grazing may be, therefore, jeopardizing desert species evolutionary potential, compromising natural revegetation and aggravating genetic diversity losses along generations that might be deepened under forecasted climate change.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.