Dispersal is a key process in determining the survival of plant species following habitat fragmentation and climate change, as well as driving the introduction and spread of invasive alien species in new regions. Due to its passive nature, seed dispersal is particularly complex, and the rare long-distance events relevant for plant species' responses to environmental change are a barrier to its understanding. Attempts to simplify the seed dispersal process often ignore dispersal by humans, despite the huge influence humans have over ecological systems throughout the world. In this Biodiversity Viewpoint, we describe how the movement patterns of humans and human-mediated dispersal vectors can be useful for understanding potential patterns of dispersal at multiple spatial scales. Humans and their associated dispersal vectors such as livestock and motor vehicles can disperse huge numbers of seeds of many plant species very long distances. Their relationships with the physical environment affect their movement, and therefore the movement of the seeds which they can potentially disperse. Therefore, we believe that a geographical approach can be useful at a time where understanding and managing pathways of dispersal are of direct relevance to the challenges faced by plant species and communities.
Dispersal is a vital process, determining patterns of species distributions at local, regional and international scales. As such, dispersal is directly related to the responses of organisms to anthropogenic environmental change, as species must be able to move between increasingly fragmented habitats or track their climatic ranges to avoid local or regional extinction (Thomas et al., 2004; Helm et al., 2006). On the other hand, the movement of non-native species to new regions has a strong bearing on their potential to establish and eventually become invasive (Wilson et al., 2009).
For flowering plants, whose movement usually occurs through the passive dispersal of seeds away from the parent plant, the response to environmental change depends on seeds dispersing unusually long distances between isolated habitat patches or to new areas or regions (long-distance dispersal; Nathan, 2006). Indeed, long-distance seed dispersal is considered extremely important for biological conservation (Trakhtenbrot et al., 2005), with excessive dispersal being implicated in biological invasions (Wilson et al., 2009) and insufficient dispersal resulting in current and future biodiversity losses in response to habitat fragmentation and climate change (Honnay et al., 2002; Ozinga et al., 2009).
Unfortunately, the nature of seed dispersal makes it very difficult to study and predict. Effective dispersal requires the establishment of a new individual plant, which is the final step in a process involving seed production, transport (including attachment and detachment) and seedling recruitment (Eriksson, 2000). The rarity and relative importance of long-distance dispersal events (Nathan, 2006) makes a broad understanding even more difficult, resulting in the problem that dispersal is quite poorly understood and rarely considered in conservation management (Trakhtenbrot et al., 2005).
Several methods have been suggested to improve the understanding of long-distance seed dispersal and its practical applications by reducing the complexity of the process. For example, using plant life history traits can be useful in predicting mean and maximum dispersal distances (Muller-Landau et al., 2008; Tamme et al., 2013), while assigning arbitrary maximum dispersal distances can estimate connectivity in a certain landscape (Minor et al., 2009). However, dispersal is not only a case of distance, but also direction, which is especially important for successful dispersal between isolated habitat patches or further afield. For this, an understanding of the dispersal vectors by which seeds move is needed. Like approximate dispersal distances, dispersal vectors for a species can be broadly predicted by life history traits (Thomson et al., 2010), but to predict patterns of dispersal in space, it is important to understand both how and where seeds can be dispersed by these vectors. Frugivores such as birds and small mammals have been shown to disperse seeds directionally between habitat patches, according to their habitat preferences and influenced by the landscape in which they move (Carlo et al., 2013; Perea et al., 2013). The realization that the interaction of plants, seed dispersal vectors and the physical landscape influences seed dispersal is significant, but the required detail of such investigations means that only very few plant species can be studied.
Human Activity and Human-Mediated Dispersal
An issue which many of the above studies has in common is that dispersal by humans is generally neglected. This is despite the fact that humans are often present in the study areas of interest and that the interest in dispersal is partly driven by the need to understand plant responses to anthropogenic environmental change. Dispersal by humans and associated dispersal vectors over which they exert control (human-mediated dispersal; Wichmann et al., 2009) is actually common for many plant species, and seeds can be dispersed long distances (Pickering & Mount, 2010; Auffret, 2011). Human-mediated dispersal is implicated in the transfer of invasive alien plant species to new regions (Wilson et al., 2009), while at the same time, a failure to disperse through fragmented landscapes is causing extinctions in human-dominated environments (Ozinga et al., 2009). Thus, human-mediated dispersal is particularly interesting in the way that it can be both insufficient in response to anthropogenic change and excessive as a driver of change itself. Importantly, the human element means that it can be actively managed. It is therefore of clear interest to study how humans influence seed dispersal, both at local and landscape scales in rural environments for conservation (Auffret, 2011), and at regional and international scales for understanding species invasions and plant responses to climatic change (Caplat et al., 2013).
In this Biodiversity Viewpoint, we focus on the unintentional dispersal of seeds by human-mediated dispersal vectors, and how considering the movement patterns of these vectors in relation to the physical environment can contribute to an understanding of dispersal patterns at multiple spatial scales. This is relevant both for target species where dispersal is lacking and invasive alien species where dispersal is a problem. Although species invasions have also resulted from an intentional dispersal of plants by humans (Wilson et al., 2009), these are theoretically easier to control and prevent compared to unintentional introductions (Hulme, 2009). The subsequent spread of introduced species depends on dispersal ability (Sakai et al., 2001), which can often be human-mediated (Trakhtenbrot et al., 2005).
Despite uncertainties regarding seed attachment, detachment and eventual establishment, we believe that it can still be of value to focus on the movement patterns of humans and human-mediated dispersal vectors when studying seed dispersal. As the movement and behaviour of seed dispersal vectors is known to be important in determining dispersal patterns, and human-mediated vectors are known to disperse a large number of seeds and species long and short distances, it is of interest to combine these aspects to understand how humans can influence and manage patterns of seed dispersal at multiple spatial scales. This is especially relevant in the human-dominated landscapes and regions undergoing anthropogenic environmental change. Human movement and how this movement interacts with the physical environment falls within the subject of geography, and we spend the remainder of this Viewpoint outlining why we think that a geographical approach can contribute to a greater understanding of how and where plant species move. Using literature from both conservation and invasion biology, we describe how the movement patterns of different human-mediated dispersal vectors are important for understanding how and where seeds are dispersed from the landscape to the global scale.
The Geography of Seed Dispersal
A guiding principle of (human) geography is that humans follow regular and predictable patterns of movement which manifest on different temporal and spatial scales, constrained or facilitated by the attributes of their physical surroundings. This has been used to map human encounters, the spread of innovation and the movement of an individual over a lifetime (time geography; Hägerstrand, 1970). We believe such an approach can provide a valuable backdrop for understanding spatial patterns of seed dispersal. By controlling the movement of a range of dispersal vectors, as well as to a large extent determining the physical structure of the landscape, humans play an important role in seed dispersal at multiple spatial scales. It is therefore important to understand not only the wheres, but also the whys of human movement, and that of their associated vectors. Below, we identify four prevalent human-mediated seed dispersal vectors and describe how their movements relate to the physical environment at different spatial scales (Fig. 1). We describe both the movement of dispersal vectors through the landscape and the direct transport of vectors between isolated areas in space. These examples illustrate the role humans can play in seed dispersal both in connecting otherwise isolated habitats and in forming potential pathways of invasion and migration.
Humans have long been recognized as a potential means of dispersal for plant species (Woodruffe-Peacock, 1918). However, it is only relatively recently that unintentional seed dispersal by humans has received serious research attention. Huge numbers of seeds can attach to the clothing and footwear of humans in the landscape, representing a significant fraction of available species (Auffret & Cousins, 2013a). Once attached, these seeds are able to disperse far beyond species' regular dispersal distances (Wichmann et al., 2009). The extreme mobility of humans means that seeds can be dispersed at a range of spatial scales. At landscape scales, humans can move species between spatially isolated habitats such as sites of interest for tourists, agricultural parcels, the study sites of a field ecologist or a combination of any of the above. At these scales, humans generally follow paths or tracks (e.g. Wichmann et al., 2009), and in that respect follow predetermined patterns of movement through the landscape (Fig. 1a). Which patches are visited can depend on ease of access or land ownership. At this scale, humans can be expected to have a common point of departure into the landscape, such as a home, car park or tourist centre. Tourism can also be a driver for regional-scale dispersal, for example between sites of interest along a common theme (Fig. 1b; D'hondt et al., 2012a), as seeds can be transported with humans as they travel from the source community (Auffret & Cousins, 2013a). At the international scale, point-to-point transportation by air and sea can connect different environments through seed dispersal by travellers and their possessions (Fig. 1c; Ware et al., 2011; Chown et al., 2012).
Livestock and wild herbivores
Livestock represent another capable seed disperser over which humans nevertheless exert almost total control. Cattle, sheep and horses are the most commonly studied, and together they have an immense capacity for dispersing seeds (Auffret, 2011). Cattle are thought to be able to disperse over 2.5 million seeds per individual and year in their manure (endozoochory), with horses and sheep able to annually disperse 500,000 and 40,000 seeds, respectively (Mouissie et al., 2005). The retention times of germinable seeds inside the livestock mean that dispersal distances depend on how far the animal moves in 2–3 days (Cosyns et al., 2005), while almost any seed has the potential to become – and remain – attached to livestock fur (epizoochory; Couvreur et al., 2004b).
Within pastures, vector movement and seed dispersal are determined by the habitat preferences of the livestock in question (D'hondt et al., 2012b). At the landscape scale, seeds can be dispersed during the organized movement between pastures either guided along paths or transported longer distances (Fig. 1d; Wessels et al., 2008; Hogan & Phillips, 2011; Auffret et al., 2012). The direction of such movement and transportation will depend on land ownership or membership of local conservation schemes. Recreational horse riding is also known to disperse plant species along tracks at the landscape scale (Wells & Lauenroth, 2007). Rotational grazing can connect isolated pastures at regional spatial scales as part of larger conservation initiatives (Fig. 1e; Couvreur et al., 2004a; Rico et al., 2011). The historical herding practices of transhumance and the export of live animals are able to disperse seeds hundreds of kilometres (Bruun & Fritzbøger, 2002; Manzano & Malo, 2006). Although the practice of transhumance exists mainly as a rare cultural relict in Europe, the practice still occurs elsewhere (Bläß et al., 2010), while the transportation of livestock at regional and international scales by land, sea and air is increasing world-wide (Fig. 1f; Hogan & Phillips, 2011).
Similar to livestock, large wild herbivores are also able to disperse seeds on their coat, fur and hooves (Heinken & Raudnitschka, 2002; Jaroszewicz et al., 2013), moving species across landscapes and regions (Myers et al., 2004; Auffret & Plue, 2014). Although their patterns of movement are not directly controlled by humans, all stages of dispersal by larger wild animals are affected by human activity (Markl et al., 2012).
Private motor vehicles
Although there are some exceptions, movement of private motor vehicles is generally confined to road networks (Fig. 1g–i). Despite this, studies from across the globe have identified cars as capable dispersers of a large range of plant species (Ansong & Pickering, 2013). Seeds can remain attached to the undersides of vehicles for tens to hundreds of kilometres (Taylor et al., 2012), while shorter dispersal distances are possible in a vehicle's airflow (von der Lippe et al., 2013). Due to the confinement of cars to roads, dispersal is limited to habitat patches directly adjacent to the road network and to the roadsides themselves (Fig. 1g). However, suitable habitat patches can often be located close to roads (Auffret & Cousins, 2013b), and in rural landscapes, road verges themselves can be valuable habitats (Cousins, 2006). The properties of road verges and dispersal by vehicles can also combine to determine the larger-scale distributions of certain plant species (Fig. 1h–i; Scott & Davison, 1985).
Agricultural machinery and commercial vehicles
Similar to private vehicles, agricultural machinery is another capable human-mediated dispersal vector. However, by having a larger capacity to transport seed-containing mud and plant material, they are able to disperse a higher volume of seeds than cars (Strykstra et al., 1997; Auffret & Cousins, 2013b). Combined, private vehicles and agricultural machinery have been found to disperse a quarter of all known species within a landscape over the course of a growing season (Auffret & Cousins, 2013b). In addition to the larger dispersal capacity, machinery is less confined to the road network, meaning that seeds can be transferred between fields which were otherwise isolated in space within a landscape (Fig. 1j; Strykstra et al., 1997). Seeds of agricultural crops and weeds are nonetheless dispersed along roadsides where they can then establish (Fig. 1j; Bailleul et al., 2012). At regional scales, logging vehicles have been found to disperse seeds both along roads and then locally into log landings (Fig. 1k; Veldman & Putz, 2010). The movement of these vehicles is related to land ownership and locations of the delivery of the raw goods (Veldman & Putz, 2010; Bailleul et al., 2012) and should therefore be quite predictable. At international scales, we expect vector movement related to the delivery of the processed materials to be confined to the road network along regular trade routes, similar to dispersal by private vehicles (Fig. 1l).
What Can Geography Offer Dispersal Ecology?
The above examples illustrate the extent of human influence over seed dispersal at multiple spatial scales. This is evident from both the range of human-mediated dispersal vectors and the quantities of seeds and plant species they can disperse. However, more valuable than identifying the potential for dispersal is trying to understand the manifestation of this potential in space. This depends on the movement patterns of these dispersal vectors, and this is where we think a geographical approach may be useful. Geography can provide a broad understanding of both the physical environment and how humans (and their associated vectors) use that environment. If the seeds are being moved, then it is geography which ultimately determines where they might end up.
Management of human-mediated dispersal
A geographical approach thus provides the opportunity to identify how seeds can be dispersed, and the human element allows us to actively manage the flow of seeds across scales. This is valuable considering the importance of seed dispersal in determining plant species responses to global environmental change. Fig. 1 shows that humans influence seed dispersal through multiple dispersal vectors, at multiple spatial scales, both in systems where biodiversity is threatened by insufficient dispersal by desirable species and where the dispersal is driving the spread of invasive plant species. Therefore, humans can manage dispersal by (1) directly or indirectly influencing the movement of dispersal vectors and (2) by understanding the movement of dispersal vectors and influencing the dispersal. For example with livestock, rotational grazing schemes can be initiated, adapted or halted to immediately change the patterns of dispersal by including or excluding certain habitat patches within a landscape or region. Directing humans or vehicles to take certain routes can influence their movement in space, while linear elements acting as valuable habitat for natives or pathways for invasive species can be identified and managed. On the other hand, if movement is a fact, then identifying hubs for the import of unwanted species to new regions or the movement of vectors influencing their subsequent spread might be useful to intercept vectors and eradicate harmful dispersal. The identification and management of dispersal patterns can even be important before any changes in biodiversity are detected, as time-lags exist both in the invasive spread of introduced alien species and local extinctions following landscape change and a failure to disperse between fragmented habitat patches (Jackson & Sax, 2010).
Although we have concentrated on the movement patterns of dispersal vectors as whole, it is also important to note that the densities of the vectors also play a role in how effective dispersal can be. Fortunately, the human (geographical) element means that such data may in many cases be available. Statistics regarding human populations and livestock are common at quite fine spatial scales, while visitor numbers are also commonly collected data in popular tourist areas (Wichmann et al., 2009; Chown et al., 2012).
Human activity poses many serious threats to biodiversity world-wide. The process of dispersal can act as both a driver of, and response to these threats. Humans also have a huge influence over seed dispersal through a number of effective and largely generalist dispersal vectors. This influence means that dispersal can be identified and managed for the purposes of biodiversity conservation. Recent research has often hinted towards the importance of vector movement in understanding patterns of diversity and long-distance dispersal, and we have used this Biodiversity Viewpoint to try to assemble these under the broad heading of geography. Just as invasion biology has recognized the importance of dispersal pathways for invasions of non-native species (Wilson et al., 2009), we believe that the capabilities of humans to influence seed dispersal means that such pathways can also be of value for managing dispersal for whole communities at the landscape scale.
Although we concentrated on movement rather than density, in many cases, human-mediated dispersal vectors are abundant. From herds of livestock to herds of tourists, seeds are being moved at all times under the direct influence of humans. Our patterns of movement and their relationships with the physical environment are an established scientific discipline, and we think that geography has a place in the ecology of global environmental change.
This work was funded by the Swedish Research Council for Environment, Agricultural Sciences and Spatial planning (FORMAS) and the cross-disciplinary project EkoKlim at Stockholm University. Two anonymous referees provided insightful comments on a previous version of the manuscript.
Alistair Auffret is a postdoctoral researcher interested in the role of humans in determining biodiversity and connectivity in rural landscapes. Through understanding how past and present management affect present-day plant communities, he aims to use his research to make useful recommendations for landscape-scale conservation.
Johan Berg is a senior lecturer in human geography, interested in the history of the Swedish rural landscape and the people who inhabited it.
Sara Cousins is a professor of physical geography. Her research interests are based around the landscape-scale analysis of biodiversity patterns and their relationship with past and present land use.
Author contributions: A.G.A, J.B and S.A.O.C developed the idea, A.G.A wrote the paper, with input from J.B and S.A.O.C.