The use of seed banks in ecological restoration: effects of anthropogenic disturbances as ecological filters

The degree and speed of loss of highly biodiverse ecosystems overcomes our capacity to implement actions to protect or restore them, and is leading to ¿a global biome crisis¿ (Temperton et al. 2019). In response to this situation, the United Nations declared 2021¿2030 as the UN Decade on Ecosystem Restoration (Ferrier 2020; Lohbeck et al. 2020). It unites the world around a common goal, to restore degraded and destroyed ecosystems at a massive scale and, in this way, to fight against climate change and improve food security, water supply and biodiversity (Ferrier 2020; Lohbeck et al. 2020). All ecosystems need to be taken into account for restoration (Temperton et al. 2019), at the same time as efforts should be well distributed to achieve the goals (Ferrier, 2020). Several countries have made ambitious restoration plans (Suding et al. 2015; Chazdon et al. 2016; Meli et al. 2016). However, one of the main problems that restorers have to face is the lack of sources of biological material for reestablishment of the lost community, which have led to the implementation of restoration programs based only on widespread and common native species (Moreira da Silva et al. 2016) and, by using commercial seed mixtures that are not generally of plants of local provenance (Kiehl et al. 2010). Moreover, around the aim of UN Decade on Ecosystem Restoration, natural regeneration has been recognized as a key process for restoration (Lohbeck et al. 2020). In both contexts, soil seed banks result in an attractive way to reintroduce indigenous plant populations after a disturbance (Davy 2002). In fact, the recruitment of native species from seed banks has been considered as a necessary condition to guarantee the successful restoration of plant communities (Middleton 1999).At global scale, soil seed banks are biodiversity reservoirs (Vandvik et al. 2016) that buffer the response of plant communities to environmental changes (Plue et al. 2020). Moreover, after disturbances they re-establish the lost vegetation (Kalamees et al. 2012) and provide resilience to the plant community (Vandvik et al. 2016). Furthermore, seed banks are important in efforts to recover the lost genetic diversity where adult populations have become extinct (Mahy et al. 1999; Honnay et al. 2008). In degraded areas, the seed bank accelerates succession because it has a key role in the natural regeneration processes (Ma et al. 2019). For this reason, its use in spontaneous restoration has been recognized (Wang et al. 2020). In addition, the knowledge about seed bank characteristics in degraded ecosystems is used as an indicator of whether the ecosystem is changing towards a more degraded state, and if this change is irreversible (Bhattachan et al. 2014). Moreover, its similarity with the aboveground flora is an indicator of the level of effort that will be required to restore an ecosystem (Ma et al. 2019). This, is the reason why the seed bank is considered a key aspect when choosing an ecological restoration strategy to be applied (Galloway et al. 2017). In particular, the presence or absence of seed banks, their quality and composition might be used as a tool to take these decisions (Cohen et al. 2004). Therefore, this knowledge has an important role in the planning of restoration projects. Specifically, the evaluation of the state of seed banks can be used in the strategy of reintroduction of species in the area to be restored. As a source of local plant material, the seed bank has been recommended to bring back lost species in degraded ecosystems (Holmes & Richardson 1999) and, because it is better adapted to the environmental conditions of the area to be restored when compared to commercial seeds (Schaal & Leverich 2005). Finally, in some cases when the seed bank is depleted, its recovery can be used as criteria to monitor the progress of restoration processes (Hutchings & Booth 1996) and an indicator of the success of the restoration projects (Ghorbani et al. 2007; Moressi et al. 2014). As a result of its importance, several studies have sought to identify the potential use of the soil seed bank for the ecological restoration. The first one was carried out by Bakker et al. (1996), who published a synthesis about the methods used to estimate seed longevity and density and, the role of each kind of seed bank to recovery the above-ground vegetation. The authors identified several gaps in our knowledge such as effects of soil management on seed viability and longevity and the physiological basis of longevity of seeds in the soil. Later, Bossuyt & Hermy (2003) evaluated the potential of persistent seed banks to recover grassland and heathland communities from the results of 16 studies and the database of Thompson et al. (1997) for North West Europe. They concluded that restoration from the recruitment of seeds in the soil could be difficult because no target species increased through succession whereas those seeds of target species declined due to senescence. Moreover, they pointed out that there are limitations to the study of the potential of seed banks to restore ecosystems, for instance, constraints imposed by the sampling effort and the lack of knowledge about specific germination requirements. These limitations were also recognised five years later by Bossuyt & Honnay (2008), who compiled and analysed 102 seed bank and standing vegetation studies conducted in European forests, marshes, grassland and heathlands. After examining seed bank characteristics, they identified that the dominance of few species and the low number of seeds of rare species make restoration of target communities, in most cases, not viable. However, the authors clarified that depending on disturbance the seed bank may be useful to restore specific ecosystems. In 2010, Bedoya-Patiño et al. (2010) published the first synthesis about the role of the seed bank in the recovery of tropical forest. The authors identified that the kind of the disturbance, the land use intensity and temporal changes in soil characteristics affected this role. In the same year, Kiehl et al. (2010) reviewed several techniques employed to reintroduce species in different semi-natural grassland restoration projects in Central and North Western Europe. Among these techniques, the seed bank became important when soil and turf transfer are applied. The next year, Török et al. (2011) carried out a comparable review, but this time, with the commonest techniques to restore grasslands on croplands in Europe. The authors identified that the use of seed banks takes place in spontaneous succession, hay and topsoil transfer. Nevertheless, they highlighted that the role of seed banks depends on site conditions, donor site availability and the time in which the material is collected. Recently, Kiss et al (2018) evaluated the potential of seed banks to buffer the responses of grassland vegetation to climate variation from a review of 42 seed bank studies. The authors suggested that persistent seed banks can be used for passive restoration in habitats characterized by frequent disturbances. Finally, Wang et al. (2020) analysed several seed bank characteristics and its role in the restoration of forest, forest-steppe and steppe located in China from 38 studies. The authors concluded that the seed bank is a useful resource of plant material, during the early successional stages, in degraded grazing-excluded grasslands and abandoned slope farmlands. Despite the contribution of these studies, there are several uncertainties of the use of the seed bank for the ecological restoration. For instance, most of the research has been focused on temperate ecosystems, especially European grasslands. This leads to a bias in our understanding of the potential contribution of seed banks in ecological restoration and limits the use of soil seed bank to recovery tropical ecosystems. Furthermore, most reviews have been focused in only one ecosystem or in ecosystems located in a particular geographic region; in less proportion, some studies have addressed the review to a particular kind of disturbance, instead of an interaction between several factors.The literature shows that use of seed banks as a source of plant material is restricted to the persistence of seeds in the soil. Davy (2002) stresses that transient seed banks have little utility for ecological restoration and that the more persistent is the bank, the greater its ability to be used in restoration. This can be explained because persistent seed banks have the capacity to survive episodes of disturbance (Thompson 2000), they, ¿as a memory of the original plant community¿, are the only one type that can contribute to the recovery of the lost or degraded plant species (Bakker et al. 1996). Therefore, they may be used as a source of plant material in ecological restoration. However, there is not enough knowledge about species that form persistent and transient seed banks, in particular, for tropical ecosystems. According to van der Valk & Pederson (1989) it is necessary to know the seed bank state before initiating any restoration process. But, for the limited time and financial resources of some projects, it is not always possible to do. Furthermore, techniques to assess jointly in situ seed banks and the general state of the ecosystem are not available to land managers to take effective and efficient decisions about the applicability of seed banks in restoration programmes, which reduces their use (Jones & Esler 2004). Therefore, a theoretical approach about the state and future usefulness of seed banks can contribute in these cases. This approach might be based on whether the seed persistence in some ecosystems is higher in comparison with others (Kiss et al 2018). This evidence may contribute to the planning process of restoration projects. The persistence of seeds may be regarded as a species trait (Fenner & Thompson 2005) that depends on genetic and microenvironmen