Responses to future disease outbreaks require a comprehensive theoretical framework in which to formulate our predictions. Being able to identify species more likely to become invasive or turn into emerging diseases would greatly facilitate prevention tasks [22]. Understanding the reasons for the success of an invasive species is often possible in retrospect, however problems arise when we try to predict which species will have a higher likelihood of establishment and spread in the future [23,24]. Several traits can discriminate between invasive and non-invasive species [25], and this applies to some emergent pathogens that share common features such as a broad host range [24]. Unfortunately, trying to predict which species will become invasive based on climate matching and geographical range is often unsatisfactory. In addition, the small proportion of cases in which a species becomes invasive poses a fundamental problem when making predictions [23].

The complexity of the process leading to an emergent disease or to an invasion also represents a difficulty when trying to formulate theories. As an example, the process of invasion can be understood in three steps: (1) introduction into a new habitat; (2) initial colonization of and successful establishment in the habitat; (3) followed by subsequent dispersal and secondary spread into additional habitats. Host jumps have been identified as a major driver for new pathogens to cause damage [24] and, in the case of plant pathogens, they are frequently initiated by humans [26,27]. Anthropogenic activities can also facilitate introduction and colonization for example, due to increased niche opportunities after disturbing native communities [28]. While arrival into a new habitat is typically the direct or indirect result of human activities, final establishment may result from more complex relationships, and the characteristics needed in one of the steps above may be different from the characteristics needed for another [29,30].

Once established in a new region, the transmission potential of the pathogen within the host population will determine the size of the outbreak. In theory, there is a non-linear association between transmission and size of epidemics: small changes in the transmission potential may result in large differences in the dynamics of the epidemic [24]. Changes in the transmission potential can be due to a wide variety of reasons such as changes in host ecology and environment, changes in host distribution, changes in host phenotype, changes in host genetics and changes in pathogen genetics [24]. Again, when trying to predict emergence or invasion, all these different theoretical causes may result in a broad range of possible outcomes. An efficient dispersal, such as spores spread via wind or water, seems to be a common feature amongst some invasive species. The fact that the entire population of trees of widespread genera such as Castanea, Ulmus, and Fraxinus quickly became infected over large areas in the last century reveals the importance of the pathogens' spread capacity for invasiveness (e.g., [31,32]). Besides the dispersal capacity of a pathogen, vectors can play a crucial role on the emergence of infectious diseases [33]. For example, the causal agents of beach bark disease, Neonectria spp., have a low capacity to spread and cause damage on their own, but became invasive due to vectoring by the native beech scale insect, Cryptococcus fagisuga [34]. Similarly, in the case of the Dutch elm disease, the level of damage is strictly dependent on the activity of the bark beetle vectors, although in this case the fungus is highly pathogenic. By carrying Ophiostoma novo-ulmi to new elm trees causing disease and tree mortality, the bark beetles also create a suitable substrate for further breeding creating a positive feed-back loop increasing the speed of the epidemic.

Invaders with large impacts may exhibit small but critical differences to native species previously occupying the invaded niches, providing them with a higher competitive ability to exploit local resources [35,36]. The emergence of ‘novel weapons’ is a useful concept in order to explain how a pathogen can become invasive [37]. Mirroring what has been observed for invasive plants, fungal pathogens with ‘novel weapons’ may allocate resources normally used for defense or competition into reproduction therefore increasing their spread [36]. Plant and forest pathology disciplines assume that pathogens co-evolved with their hosts under specific environmental conditions leading to equilibrium between host, pathogen and environment [38]. When a pathogen with ‘novel weapons’ is interacting with a host lacking an appropriate defense system, the resulting level of disease will be more severe than expected from a native pathosystem. For example, Pinus contorta provenances originating from regions outside the natural distribution of sweet fern rust (Cronartium comptoniae) were much more susceptible to the disease than provenances originating from within the region when tested in a common garden experiment [39]. Moreover, pathogens invading a new environment may experience a lower competitive pressure from the microbial community compared with that experienced within their native range and this allows the introduced pathogens to more readily reach their pathogenic potential [37].

When trying to predict the probability of an ecosystem to be invaded, we might encounter a similar sort of complexity as when trying to predict which pathogens will become invasive. General concepts such as lower diversity, host continuity, or unfavorable climatic conditions, normally associated with a higher vulnerability may or may not relate to a higher invasiveness. Theoretical models suggest that a pathogen entering into a new community is less likely when the latter has a higher species diversity [28]. High infection rates are often observed when tree species are planted in dense monocultures whether in their native environment or as exotics [2,37]. In contrast, when invaders are better competitors or use different resources to native species the effect of diversity becomes weaker [35]. Unfortunately, a higher diversity does not always protect against pathogen invasion and could potentially increase the odds of having an invasion promoter i.e., a species that could facilitate the expansion of the invader [35]. An invasion promoter, for example, is essential for macrocyclic rust species such as Cronartium ribicola, the causal agent of white pine blister rust, where the presence of the alternate host is a prerequisite for the survival of the pathogen (see also [40]). In order to address the complexity of managing emerging diseases, general management recommendations need to be backed up by a thorough understanding of the underlying mechanisms in each pathosystem [41].

Understanding past and present drivers for emergence may not necessarily increase our potential to anticipate future disease outbreaks when new factors may arise and be more influential. Future drivers of emerging plant diseases are hypothesized to be introductions of new pathogens, climate change, and intensification of management [27]. At present, introductions have been the result of human activities, while in future, it is hypothesized that climate change may act as a major driver for emerging diseases [12,27]. Moreover, there is a discrepancy between the scale of understanding climate processes (landscape, county) and the scale in which plant pathology operates (forest, field). In this sense, monitoring epidemiologically relevant variables may strengthen our capacity to make predictions [42].

The relatively low frequency of long distance spore/insect spread minimizes the risk of intercontinental infections occurring naturally. This is partly because of the low probability of spore/insect survival and partly due to the low probability of exposure of susceptible plants to such long distance spread. Instead, most new diseases are associated with infected plants or plant material that have been transported across borders for planting or packaging [5]. For example, the most common pathway for plant pathogens introduced into Great Britain between 1970–2004 was vegetative imports such as seedlings, tubers and scions [43]. Seeds can also vector pathogens; the outbreak of pine pitch canker (caused by Fusarium circinatum) in South Africa is thought to be the result of importing contaminated seed from Mexico [44].

Nurseries are not only a frequent source of vegetative imports but could also be an ideal site for the creation of new species of pathogens. New species of pathogens can develop as the result of hybridization between closely related ancestors [45]. It is very likely that nurseries provide a venue for allowing pathogens from different hosts to meet in a common place. Should hybridization occur between species resulting in novel pathogenicity patterns, tree nurseries provide a multitude of new potential host species. Nurseries also offer a venue for pathogens from different hosts to meet in a common place. This can facilitate host jumps leading to the development of new diseases. In addition, by being centers of trade for plant material, nurseries provide dissemination routes for any diseased plants [46,47]. Since symptoms of new diseases are normally not well described and latent or cryptic disease phases may go undetected for prolonged periods, plants may be disseminated without any disease problems being recognized.

Commercial tree species that are planted worldwide, such as Eucalyptus spp. or Pinus radiata, have not only suffered from extensive damage due to introduced species, but have also been responsible for the movement of several exotic pathogens into native hosts [48,49]. Fusarium circinatum, a typical pathogen of Pinus radiata has been observed on native Pinus pinaster plantations in Spain [50]. The canker pathogen Neofusicoccum eucalyptorum was regarded to be specialized in Eucalyptus, until it jumped onto three native tree species after introduction in Uruguay [51]. Host-jumps from native species into commercial species have also been observed [26], for example guava rust caused by Puccinia psidii jumped from native Myrtaceae in South America onto introduced Eucalyptus and now threatens eucalypts in Australia [52]. Another example comes from boreal forests in Sweden, where the introduced Pinus contorta became infected by the local canker fungus Gremmeniella abietina [53]. Such host jumps may also threaten the new hosts at their native origin, should the pathogen unintentionally be carried back there. International legislation has put very little emphasis on minimizing this type of risk.

Wood and wood packaging can also be a common pathway. Heterobasidion irregulare was introduced with wood into the Italian Peninsula during World War II [54] and Ceratocystis platani, causal agent of canker stain of plane (Platanus orientalis), was introduced into Europe on packaging material in the 1940s [55]. The importance of this pathway has been recognized through international regulations governing the treatment of wood packaging material to reduce pest and pathogen movement (as the International Standard for Phytosanitary Measures No.15 [56]). There is an increased recognition of the need for handling problems in a generic manner through pathway analysis where the means of spread with the highest risk are recognized and threats can be handled efficiently without the need for identification on a case by case basis [5]. However, regulations based on pathway analysis may be regarded as being too generic, thus not meeting the requirement of minimal impact for trade within the International Plant Protection Convention (IPPC) of 1951 [57].

Responses in the society to threats posed by pathogens are dependent on a proper awareness of the types of problems that we encounter. Yet there are several obstacles to the perception of emerging diseases. Organisms are generally assumed to be relatively homogenous and genetically stable and the complication of intraspecific variation is not fully recognized [37,58]. Improvements in molecular techniques have increased our knowledge of the complex breeding systems and taxonomy of pathogens. Cryptic species with minimal or no morphological differentiation have been identified in a multitude of fungal taxa [37,59,60]. They can be associated with different host ranges and pathogenicity patterns. To handle these issues in a legislative context represents a major challenge. Without proper naming of the threats, it is also hard to define the countermeasures that need to be taken. One example of how these issues have been treated by taxonomists is provided by the root rot pathogen Heterobasidion spp. Studying mating incompatibility and host range, the Heterobasidion annosum species complex was found to include several taxa with varying host preference and pathogenicity during late 1970s and 1980s. The species present in Australasia was not the pathogenic H. annosum and highlighted the need, from a plant quarantine perspective, to quickly differentiate the saprotrophic Australasian species from its pathogenic relative. Thus, the southern hemisphere species was renamed as H. araucariae and H. annosum was declared a quarantine threat in Australasia [61].

Another problem arises from our lack of appropriate common language to describe intraspecific variation in fungal populations. A fundamental problem lies in the way we build concepts, i.e., we frequently categorize observations based on visual similarity. However, pathogens evolve from common ancestors into populations with pathogenic traits that affect the dominant local tree species. Such traits may not have an obvious external phenotype, and are therefore hard to recognize without knowledge of the evolutionary history or performing pathogenicity tests. In terms of legislation there are few examples where subspecific naming has been used to identify particular severe tree pathogens. One example is provided by Dutch elm disease (DED) that was epidemic in Europe after the First World War [4]. The pathogen and vectors were exported to North America in the thirties and there the causal fungal partner of the disease complex (Ophiostoma ulmi) was altered or even replaced with a closely related fungal species that remained unrecognized. This new species was later reintroduced into Europe and it became obvious that the newly introduced American variant was much more aggressive. It was later described as a new species, Ophiostoma novo-ulmi. Today this new species is spread all over Europe. Had the biological distinction between the two species of Ophiostoma been recognized and named prior to the unfortunate re-entry into Europe, we would potentially have had the appropriate concepts and words available that could have helped to ban the import of infested plant material and saved elms from facing eradication.

As well as recognizing species differences, variation in genotypes may also be important. Introduction of another genotype may significantly increase the risk in an area if the new genotype is another mating type, which then enables sexual reproduction [29,62]. The ability to reproduce sexually allows pathogens to adapt to changing environmental conditions [62]. So, even though the species is already present in a location it may be important to limit further introductions. Introduction of closely related pathogens may also give rise to interspecific hybridization potentially causing new diseases. The alder phytophthora, Phytophthora alni, which was first discovered in U.K. in the early 1990s, is thought to be a hybrid between two other Phytophthora species [63], which are individually much less aggressive on alder [64]. In this case, the species dynamics was not understood until sufficient research resources was allocated to the disease; this only happened after substantial damage had already been caused to alders throughout most of Europe.

Society has taken measures to limit dispersal through the most common pathways like trade to meet the perceived problems with introduced diseases. These measures may include sanitary actions such as heat treatment of any pathogens in imported wood products, or a sender-end inspection of potentially diseased plant materials. However, quarantine measures are not always effective and there are large problems in performing perfect control at borders. As an example, despite the international regulations governing the treatment of wood packaging material as the International Standard for Phytosanitary Measures No. 15 [56], not all countries have adopted these guidelines and this standard is not necessarily effective for all species of pathogens [71].

Within the EU, trade of plant material is allowed between member states and only a sanitary passport is needed as regulated in the Council Directive 2000/29 EC [72]. The member states not only issue sanitary passports to their products, but also to products imported from outside the EU which are then distributed further internally. Two independent organizations recommend to the European Commission which organisms may be listed as harmful: the European and Mediterranean Plant Protection Organization (EPPO) and the European Food Safety Authority (EFSA) [57]. Efficacy of quarantine policies within a region will only be as good as the weakest of the quarantine controls of its borders [15]. In addition, quarantine measures rely on the accuracy of existing knowledge of the global distribution of pathogens. A review of the records on the distribution of the quarantine organism Fusarium circinatum revealed several cases where observations needed confirmation, while other observations were not included [66]. The authors highlighted observations of the pathogen in nursery stocks in Northern Spain [73] which were not included in the official documentation, and of a record in Italy which needed further confirmation. It was not until the disease was officially reported almost 10 years later in both places [50,74] that the EU officially limited plant movement from the infected regions (2007/433/EC) [75].

Current risk assessment and regulation of new pathogens in international plant health protocols is based on lists of identified organisms recognized as a potential threat [5]. The process of adding species to the list is often too slow since once a species is recognized as a threat it is usually already a problem. Ash decline, for example, was allowed to spread from Poland and Lithuania through northern Europe for at least 10 years [32] before the causal pathogen was identified in 2006 [76], and its sexual stage in 2009 [77,78]. Phytophthora ramorum was identified as the causal agent of sudden oak death disease seven years after the first outbreaks were observed on tanoaks (Lithocarpus densiflorus) in California, U.S. [67]. This means that unknown harmful organisms are not on the list and thereby are unregulated. In addition, as mentioned earlier, pathogens frequently do not cause disease on their original host or within their native range, so may not be recognized as threats until they are already introduced to a new host or into a new area. A pest risk analysis based solely on identified species is also associated with other problems related to the complex nature of pathogen speciation mentioned earlier.

Import of regulated plant material into Europe is allowed if examined and found to be free of regulated fungi and other pests and pathogens [72]. There might, however, be problems in identifying a diseased plant. Following infection there usually is a latency period where the plant remains free of symptoms and this latency period can occur for a relatively long time period. Symptoms might also be restricted to belowground parts of plants that are difficult to examine in an efficient manner. Molecular analysis applied on routinely collected samples may enable identification of otherwise undetected pathogens, however this process may be costly and time consuming, and the problem of relying on lists of already identified threats still persists. Other detection methods can include a global network of sentinel plantings which can provide early warnings for the transfer of unknown pathogens between countries. Currently, several initiatives from the EU, New Zealand, CABI and the U.S. are on-going [79]. Monitoring within countries is also important for the early detection of new pests/pathogens. High throughput sequencing tools (e.g., 454 pyrosquencing, Solexa, SOLiD) can provide an efficient tool for screening high numbers of samples [80], which can complement current monitoring schemes based on symptoms (defoliation, chlorosis).

When a new pathogen has been introduced, actions are frequently taken in order to limit further spread and prevent establishment. Preventing establishment is, however, extremely difficult as can be exemplified by the introduction of the pine wood nematode, Bursaphelenchus xylophilus, in Europe [81]. The pine wood nematode originates from North America and has been found to kill pines and larch of susceptible species outside its native range. When the pine wood nematode was first discovered in Portugal in 1999 [82], the EU adopted large scale control measures in order to prevent its establishment and further spread [83]. Trees showing symptoms in the infested area as well as within a buffer zone of 20 km were felled and removed. It has, however, been found subsequently in various areas of Portugal and a 20 km demarcation zone was established along the Spanish border, in order to stop further spread into Europe. Despite this, the pine wood nematode has been discovered and eradicated in Spain in 2008 and 2010 [84].

Even when a pathogen is already established in an area, further countermeasures may still be worth taking. Control strategies need integrated efforts at different spatial and temporal scales: tree, landscape (or forest stand) and at a regional or international scale [85]. A good example of taking a broad, inclusive approach to controlling forest disease can be found in the management of Dutch elm disease on the isolated island of Gotland in Sweden. Since there is little possibility of natural reintroduction of the pathogen from the mainland, the governing body of the county has worked with local interest groups to preserve the elm population. A critical part of the control strategy is to involve the public by disseminating information about the disease and its control [68], as seen in other schemes elsewhere e.g., Phytophthora ramorum in the U.S. [85] or P. cinnamomi in Australia [69]. The removal of infected elms seems to successfully decrease the advance of the disease in Gotland (R. Vasaitis pers. com. 2010). This can be compared to a similar strategy of removing infected hosts made in the city of Malmö on mainland Sweden, where the eradication may have failed because the control program did not incorporate a sufficiently large geographical area. Time is also a crucial factor that needs to be considered since this will determine both the size of the potential damage and the size of the actual damage. In particular cases large scale control operations can be implemented. In Sweden for example, large numbers of trees were removed after the devastating storm in 2005, to reduce potential Ips typographus outbreaks [86]. This was possible as a result of collaborative efforts across all levels of the forest sector including private forest owners, companies and local administrators. At later stages of an epidemic, control can be difficult as well as costly and a strategy towards protecting remaining uninfected areas is normally adopted e.g., [87].