Potential for dispersal

Dispersal of earthworms can be categorised as passive through anthropogenic or natural processes and active over the soil surface or through the soil. Both cocoons and adults can be dispersed passively by surface run-off, water currents, heavy rainfall, temporary inundation of a certain area, transported by other animals, e.g. birds or through plant materials and adhesion to soil particles. The active dispersal of earthworms can be triggered by various factors, such as increased earthworm density, low quality habitat or adverse conditions, like heavy rain or flooding, surface applications of irritating fluids, contamination with heavy metals or pesticides, in general, and copper compounds in particular, acid or highly alkaline soils or occurrence of roads and cabins. This was confirmed by Mathieu et al. (2010) who showed in a mesocosm study that dispersal can be reduced by: 1) high habitat quality including the presence of litter; 2) low density; and 3) pre-use of the soil by conspecific individuals that are no longer present.

Data on dispersal rates of earthworms through soil are reported in Table3, showing mean horizontal movements ranging between 2.5 and 14 metres per year (m/y) (Eijsackers, 2011;

Emmerling and Strunk, 2012; Dupont et al., 2015). For A. caliginosa, maximum dispersal of 72 m in 8 years was reported. In the case of L. rubellus, a dispersal between 5 and 11 m/y has been measured. Overall, in agricultural sites, limited variation has been reported in the dispersal rate between different species, and earthworm-population development started after an adaptation period in the range of 2–6 years after introduction.

All types of earthworm species show the ability to disperse over the soil surface by crawling at night. For example, L. terrestris has been shown to crawl 19 m in one night and A. longa 23 m. No directionality in crawling has been demonstrated. However, earthworms can detect and avoid adverse conditions as reported above and, thus, colonisation by earthworms may not occur for years in the case of soil contaminated by persistent substances (Eijsackers, 2011).

Caro et al. (2013b) recorded a high variability within each earthworm functional group concerning dispersal behaviours. Habitat quality significantly influences the dispersal rates of both anecic and endogeic species. In a homogeneous environment, anecics dispersed further and in greater proportion than the majority of endogeics. Overall, the authors concluded that anecic species might show more active dispersal than most endogeic ones.

Earthworms dispersal behaviour can be triggered by environmental conditions, such as habitat quality. In this respect, Mathieu et al. (2010) reported that 90% of individuals belonging to the endogeic species Aporrectodea icterica dispersed when inoculated into a low quality soil, while only 20% dispersed when inoculated into a soil which was demonstrated as largely preferred by earthworms (see Figure2).

Figure 2: Aporrectodea icterica dispersal rates in response to soil properties. Suit: suitable soil (high pH, high org. matter); Uns: unsuitable soil (sandy soil, low pH). Reprinted from Mathieu et al. (2010), Copyright (2010) with permission from Elsevier

Table 3: Mean dispersal rate of earthworms species in various habitats (from Eijsackers (2011) and Emmerling and Strunk (2012)

Species Land use/soil/environment Dispersal rate (m/y)

Lumbricus rubellus Grazed grassland 7–8

Arable land 14

Peat soil > 10

Arable polder 14

Aporrectodea caliginosa Grass strips orchards 6

Grazed grassland 9–11

Arable land 7

Grassland 6

Irrigated desert soil 3.5–5

Pasture 10

Grassland/reclaimed peat 2.5–10

Arable polder soil 7

Allolobophora chlorotica Grass strips orchards 4

Aporrectodea longa Grazed grassland 5–8

Grassland 6

Lumbricus terrestris Grazed grassland 4

Grassland 1.5

Octolasion cyaneum Grassland and arable soils 8

The greater part of terrestrial gastropod diversity comprises very small animals living as detritivores in the litter layer (< 1 cm diameter in greatest dimension), some even have maximum diameters of

< 1 mm (see e.g. Barker, 2001; Sturm et al., 2006). They occur in all kinds of agricultural habitats (as grassland, acres, specialised crops, seminatural habitats) and have often rather specific preferences in terms of habitat and environmental conditions (see e.g. Kerney et al., 1983).

Barker (2001) states that the dispersal abilities of terrestrial gastropods are so low that it can be assumed that mating will be predominantly driven by inbreeding at the level of the local population.

Their ability to recolonise disturbed areas is low and affected by various environmental factors, such as the height of the corn (Wolters and Ekschmitt, 1997). For rape fields, the same authors reported that several snails only invaded about 3 m into fields from woodlands and hedges. Hof and Bright (2010) found that the number of terrestrial gastropods significantly decreased with increasing distance from thefield edge of arable fields.

Mesofauna

Dispersal is an important characteristic of mesofauna with implications for impacts and recovery of PPPs. Kattwinkel et al. (2012, 2015) reviewed the literature on recovery and concluded that Collembolan species reacted significantly differently to population perturbation (including the recovery pattern), meaning that a coarse taxonomic assessment might not be sufficient to detect adequately effects of pesticides. For example, when changing the same plots from the treated to untreated management, responses of individual species varied, e.g. numbers of the species Entomobrya nicoleti remained close to zero, whereas the abundance of Isotoma viridis were the highest recorded during the study. They also concluded that unexposed field margins play a key role as source of recolonisation confirming the role of buffer zones for recovery for mobile surface dwelling collembolan.

There is a considerable body of evidence for the importance of dispersal. Rantalainen et al. (2005) reported the ability of various members of the detrital food web to colonise newly established habitat patches under field conditions, showing that the presence of habitat corridors promoted community diversity. However, rates of movement, although varied are generally low. Dispersal rates for the fungivore species of Collembola, Onychiurus armatus, using connected distinct patches of two different soil types covering a distance of 40 m, ranged from 0.020 to 1.42 per day suggesting that on average species moved less than 10 centimetres per day (cm/d). Dispersal depended on population density, soil type and length of fungal mycelium, being inversely proportional to mycelial length, especially in a sandy soil. When a soil patch at 40 cm distance from the release point was enriched with a favoured food item, dispersal rate was increased by more than four times (Bengtsson et al., 1994). The role of hedgerows for external recovery (recolonisation) of springtails has also been investigated and demonstrated, especially in arable fields. Habitat preference and dispersal ability of different collembolan species have also been investigated by Auclerc et al. (2009) at a small scale study conducted in France. The authors showed that 6% of the identified species were land-use generalists (not restricted to a given habitat), 30% were soil generalists and 36% recolonised defaunated soil blocks within a week. The results also demonstrated discrepancies between preference in land-use and soil, indicating that land-use specialists may not always be also soil specialists. However, food availability was suggested as stimulating dispersal considering that the meadow soil was more attractive than the forest, whatever the land use preference of the species. In addition, it was also shown that dispersal ability might not be predicted based on the morphological features (antenna, legs, etc.) of the species.

Other factors that can alter dispersal rates include reproductive strategy and pheromones. There is an indication that parthogenetic species may colonise more quickly (Chahartaghi et al., 2009).

Recovery and recolonisation of Collembola may be also enhanced by the existence of pheromones which induce aggregation. As mating in Collembola may be indirect, involving deposition of spermatophores by males and subsequent taking up by females, aggregation may increase the efficiency of reproduction (Verhoef and Nagelkerke, 1977; Verhoef, 1984).

Since dispersal is important but limited there are implications for the interpretation of field study data. According to Duffield and Aebischer (1994), the recovery of invertebrate population also depended on the size of the treated plot. In addition, two different recovery patterns were identified:

(i) recovery progressing from the edge to the centre of treated areas; (ii) more rapid recovery in the centre of the large treated areas. The first recovery pattern was mostly associated with the predatory groups such as Carabidae, Staphylinidae and Linyphiidae and it can be associated with a recolonisation of the pesticide-treated plots from the untreated surroundings. The second recovery pattern was associated with the prey groups such as Aphididae and Collembola. The recovery appeared to be

faster in areas with less predation pressure. The results suggested, as also reported by Kattwinkel et al. (2012, 2015), that recovery assessment in small in-field areas might be a source of uncertainty since it could underestimate the pesticide effects on large predators, while overestimating them on microarthropods. This was extensively addresses by the EFSA PPR Panel (2015a) when dealing with the risk of intended uses of PPPs to non-target arthropods. Some patterns might be also valid for the larger Collembola species living on the soil surface and able to move significant distance between fields and edge-of-fields. One of the major threats to biodiversity is landscape fragmentation as it can result in the transformation of continuous (hence large) habitat patches into isolated (hence smaller) patches, embedded in a matrix of another habitat type. In turn, this leads to a loss of biodiversity, especially if species have poor dispersal abilities, such as Collembola (Martins da Silva et al., 2012). A recently created habitat might suffer from a reduced biodiversity because of the absence of adapted species that need a certain amount of time to colonise the new patch (e.g. direct metapopulation effect). Thus, landscape dynamics leads to complex habitat, spatiotemporally structured, in which each patch is more or less continuous in space and time. Patches can also display reduced biodiversity because their spatial or temporal structures are correlated with habitat quality (e.g. indirect effects).

Heiniger et al. (2014) demonstrated that habitat temporal structure is a key factor shaping collembolan diversity, while direction and amplitude of its effect depend on land use type and spatial isolation.

Soil microorganisms

For a long time, the ‘Baas-Becking hypothesis’, stating that ‘everything is everywhere, the environment selects’ (Beijerinck, 1913; Baas-Becking, 1934; Fierer, 2008), made a strong imprint on thoughts and views regarding microbial biogeography. It is well known that microbial communities can exhibit spatial variability at scales ranging from millimetres to thousands of kilometres and that microbial community composition in natural environments can be influenced by a large number of biotic and abiotic environmental factors. However, the impact of specific aspects of the environment on the spatial patterns of microorganisms is still poorly understood. Typical features of microbial communities, such as large population sizes and short generation times, may result in biogeographical patterns. However, unlimited microbial dispersal may lead to constant turnover and increase in genes flow (Eisenlord et al., 2012). For example, many phylogeographical and population genetic studies on plant pathogenic fungi, but also on wood decay species, have reported efficient dispersal and gene flow at a regional or even continental scale. According to Finlay (2002), free-living microbial eukaryotes are probably sufficiently abundant to have worldwide distribution. Accordingly, prokaryotes, which are much smaller and several orders of magnitude more abundant, are even less likely than microbial eukaryotes to be restricted by geographical barriers.

In addition, in the case of fungal spores, for example, dispersal patterns can affect gene flow, population structure and community structure. Dispersal mode can vary among different fungi. While epigeous fungi are mainly transported by wind, the majority of hypogeous fungi are biotically dispersed since they have fewer opportunities than epigeous fungi for being passively dispersed. For example, fungivorous mammals and invertebrates may be an important dispersal agent for many ectomycorrhizal (EM) and arbuscular mycorrhizal (AM) fungi that form sporocarps. It has been reported that arbuscular mycorrhizal fungal spores can remain viable after passing through digestive tracts of earthworms, sowbugs, and crickets. Besides ingestion, dispersal by adhesion to external surfaces of in-soil organisms is another mode of dispersal for spores of soil fungi (Lilleskov and Bruns, 2005).