Discussion
Landslides are natural disturbances that affect forest ecosystems in high rainfall tropical and subtropical regions and can incur substantial socioeconomic loss and ecological damage. We investigated potential methods of enhancing forest regeneration after landslide events via human mediated restoration efforts, focusing on ways to increase native seed germination rate. Our results show that, first, seed coating technique improves seed germination rate of native tree species. Second, selecting suitable successional groups and seed size can lead to higher germination rates and restoration success.
We found that biochar dominant seed coats had a significantly positive effect on seed germination success across our study sites, and there is evidence that seed coating can further encourage growth for young seedlings (Jones, Schwinning, & Esque, 2014; Lee & Park, 2006; Scott, 1975). Biochar coating also had differential effectiveness in enhancing seed germination in landslide scars across the species we investigated, agreeing with results from other studies showing that seed coating improves seed emergence under different soil conditions (Liu, 2010; Madsen, Davies, Boyd, Kerby, & Svejcar, 2016). Landslide scars are characterised by soil compaction and lack of organic matter, reducing soil nutrients availability and limiting vegetation growth and restoration potential (Błońska, Lasota, Zwydak, Klamerus-Iwan, & Gołąb, 2016; Fetcher et al., 1996). Treating seeds with additional coating helps seeds to overcome this initial barrier to germination due to poor site properties by serving as a carrier of nutrients and a physical shield to avoid physical damage to the seeds. Moreover, the seed coat increases the likelihood of seeds reaching germination favoured microsites, reducing the chance of the seeds being blown or washed away and minimises seed loss from predation (Overdyck et al., 2013; Taylor et al., 2020; A. G. Taylor, Eckenrode, & Straub, 2001). Moreover, adding active ingredients in coating formulation boosts the seed germination rate. In modern seed coating technology, seed coats act as a carrier of different ingredients including nutrients, plant protectant or even plant beneficial microbes for several purposes (Afzal et al., 2020; Rocha, 2019). For instance, in our experiment, biochar dominant seed coating has a 30% higher than clay dominant seed coat and raw seeds which agree with previous findings that biochar coating serves as soil conditioner and enhances water holding ability of soil (Brown et al., 2018; Madsen, Davies, Mummey, & Svejcar, 2014).
In the soil surface treatment experiment, we found no evidence of sowing biochar powder on top of soil surface affecting native seed germination on the landslide scars. Previous research has found that biochar powder applied to topsoil can enhance tree and crop growth through an increase in soil pH, fertility, and water holding ability, resulting in plants with higher biomass (Mohamed, 2015; Thomas, 2015). It seems that biochar soil surface treatment is only able to enhance growth after seed germination but does not offer direct advantages in enhancing seed germination rate.
In our study, the overall germination rates of raw seeds were 17% - 67% (Table 3) while having biochar dominant seed coat gave an additional 9.33% improvement of germination rates. Our finding was much higher than a meta-analysis of direct seeding indicating that the seed germination rate was 23.9% on average (Ceccon, 2016 ). It is also much higher than many restoration programmes, in which seedling establishment rates are often observed at less than 10% (Merritt & Dixon, 2011). The observed much higher seed germination rate in our study can likely be because of the use of native species seed or the test-site in a favourable condition of humid tropical environment, and a more in-depth understanding of the underlying difference and mechanism remain needed in the future attempts.
Our results showed that species selection was an important factor affecting seed germination rates. First, different successional groups of plant species had different seed germination rates, as they require different environmental conditions from the colonizing pioneer species to establish on landslides (Walker, Velázquez, & Shiels, 2009). This also agreed with previous studies showing that species in late-successional group had an advantage in germination rates in other landscapes (D. C. de Souza & Engel, 2018; R. P. de Souza & Válio, 2001). Late successional species have a germination strategy that conserves nutrients, allowing them to be more shade tolerant when compared to pioneer species (Kleijn, 2003). This gives them higher competitive ability but low colonizing ability (Kleijn, 2003; H. Zhang, Qi, & Liu, 2018), which is a potential reason for our observed result. Moreover, pioneer species require special environmental conditions for dormancy break to germinate, so food reserves in seeds may be depleted before these conditions are created, resulting in lower overall germination rates (Baskin & Baskin, 1998). Our results showed that non-pioneer or late-successional species could successfully establish on degraded hillsides, and that late successional species establishment is limited by seed dispersal rather than tolerance to harsh site conditions. Thus, late successional species may be used in direct seeding restoration in landslide scars.
Second, we found a positive relationship between seed size and germination rate, implying that a larger seed size improves germination rate. This also agreed with previous studies that compared the sizes of different tree species (St-Denis et al., 2013; Tunjai & Elliott, 2012). In the natural environment, large and small seeded species have different establishment strategies. Large seeded species appears to be more tolerant to stresses such as drought and shade than small seeded species as they have a larger initial energy reserve (Leishman & Westoby, 1994). Their large cotyledons support embryo development when seedlings compete for limited resources, which may be an advantage in habitats where gaps in the canopy are regularly created. Moreover, large seeded species are able to germinate equally in light and darkness and is more resistant to temperature fluctuation (T. R. H. Pearson, Burslem, Mullins, & Dalling, 2002), while small seeded species produce seedlings that are more susceptible to extreme environmental conditions and do not resist long periods of unfavourable growing conditions (Camargo, 2002). Although small seeded species have an advantage in quantity during seed production (Moles & Westoby, 2004), in human mediate restoration efforts, large seeds mask the effect of production by having rich energy source in cotyledons to enhance the chances of restoration success.
In addition to seed coat treatment and species selection, there are other methods of enhancing restoration success which were not assessed in this study. Pre-sowing seed treatments and site treatments are most commonly used in practice. Pre-sowing seed treatments such as scarification (Kaye & Kuykendall, 2001) and sowing agglomerated seeds (Madsen, Davies, Williams, & Svejcar, 2012), were shown to give higher germination rates than sowing singular raw seeds. Field treatments, such as sowing seeds at different microhabitats and biotopes (Chantal, Kuuluvainen, Lindberg, & Vanha-Majamaa, 2005; Doust, 2006), and creating grass canopy on soil surface before direct seeding (Guarino & Scariot, 2014), were also shown to increase seed germination and establishment rates. In addition to seed and site treatments, the method of seed sowing can also affect seed germination rates. In this study, we focused on direct seeding with seed burial, as seed burial can minimise animal predation on the sowed seeds (Alem, 2020; Garcia‐Orth & Martínez‐Ramos, 2008). The difference between seed coating and bare seeds on germination rate is potentially greater when broadcasting seed than seed burial as the seed balls themselves can act as physical protection to the seeds (Leverkus, Rojo, & Castro, 2015). Broadcasting native seeds on landslide scars is still under investigation and a potential avenue for further research. Furthermore, in our current results, large seeded species only occurred in the late successional stage. It would be interesting to investigate how small seeded late successional species and large seeded species in other successional stages behave on landslide scars.
Landslides are natural disturbances that can cause enormous social-economic loss and severe ecological consequences. Forest restoration on landslide scars can contribute substantially to soil surface protection, reducing future disturbance risk and help to protect human lives and properties whilst restoring destroyed habitats. Effective regeneration of these degraded areas requires successful germination of seeds, which can often be a slow natural process. Human mediated restoration can help to speed up this process. We were able to show the effectiveness of biochar seed coating and planting non-pioneer species with medium to large seeds in improving seed germination rates on landslide scars. We used Hong Kong in this study, but the findings could be applied to post disturbance vegetative restoration and recovery on landslides of other hilly terrain in the seasonal tropics.