Research addressing community wide and species-specific seed rain patterns of forests in Sierra Leone is lacking. Most approaches to understanding tropical forest dynamics and plant species availability have focused on patterns of vegetation structure, composition, diversity and distribution (Sesay, 2019; Fayiah et al., 2020; Mattia and Sesay, 2020). Kasewe forest reserve has been through series of human exploitations since it was first gazetted in 1919 (Munro and Hiemstra-van der Horst, 2012a; Munro et al., 2017) for resources like gum copal, timber, poles, and other non-timber forest products (NTFPs) and bush meat. During the Ebola crisis of 2014 - 2016 the population within the forest increased as people fled their communities and found the forest as a source of livelihood. The resulting exploitations have expanded and increased over the years to levels which are no longer sustainable. Exploitation rates increased through artificial deforestation the use of power saws offered by affluent merchants in return for charcoal and related products to their benefactors in cities where these products are in high demand.

 

The most preferred species for charcoal production by the exploiters is G. copalliofera, a species of tree that was heavily exploited for gum copal during the colonial period but no longer being done at present. The high quality of charcoal produced from this species and the increased exploitation it currently faces in this unprotected reserve could pre-dispose this limited geographical range species to extirpation.

 

This study therefore aimed to characterize and evaluate the dynamics of seed dispersal by G. copallifera; specifically, to quantify the seed rain along the different land uses in the forest reserve, as well as analyze for differences in total abundance, species richness, diversity and composition in the seed rain.

Study site

 

Kasewe Forest Reserve (8°18′53″N, 12°15′43″W?? / 8.31472; -12.26194) occupies a small range of steep sided hills of volcanic origin rising up to 500 m above the surrounding plain (Bowden, 1997; Lytwyna et al., 2006). The reserve covers 2,331 hectares, and the dominant vegetation is tropical forest containing a mosaic of moist semi-deciduous forest, evergreen forests and savanna that give way to medium altitude forest on the upper slopes (UNEP, 2008). The reserve is located in Moyamba District in the south-central part of Sierra Leone about 170 km east of the capital city of Freetown. The Bo-Freetown highway is one of the busiest provincial roads in the country, it forms one discernible boundary and allows easy access for exploitation and transport of products such as charcoal and timber from the reserve.

 

Seedrain sampling

 

Construction of seed traps and placement

 

In order to assess the nature of the seed rain and dispersal capability of G. copallifera in the reserve, seed traps were set up along pre-cut transects in degraded areas adjacent to the forest edge and in the interior of forest patches. Transects were up to 300 m long with equally spaced pairs of sampling units (traps). One trap consisted of a 1.5 m² plastic mesh (1mm mesh) with the corners lifted 1.2 m above ground level on poles, while the other consisted of swept ground demarcated by four wooden stakes driven into the ground but with ends sticking out at about two meters from the mesh trap.  Forty-five transects were set out in the forest and five at right angles to the forest edge into the degraded forest. Traps were monitored on a weekly interval for 12 weeks. In the degraded areas, focal trees of G. copallifera were selected from which transects were located with the aim of observing the dispersal dynamics of this species. Transects in the degraded area had traps installed at 10 m intervals out to 300 m from the forest edge.

 

Seed collection and identification

 

Traps were emptied on a weekly basis with standard data sheet for each type of trap that was installed. Contents of each trap within the trap line was emptied into a separate labelled bag or container and leaves and twigs removed. Physical states of the fruits and seeds were noted: rotten, germinated, partly eaten, insect infested and animal feeding. Seeds were grouped in two classes namely G. copallifera (Kobo) and others.  Trap contents were taken into the laboratory and sorted by hand. All potentially viable G. copallifera seeds (>1.5 mm diameter) were extracted and identified (Cottrell, 2004). Other species were identified by comparison to the seed collections held at the National Herbarium of Sierra Leone in the Department of Biological Sciences, Njala University.

 

Statistical analyses

 

Majority of the analyses were done in R version 3.6.2 (R Core 2020). ANOVA single-factor analysis, paired t-tests and linear regression were used to determine the effectiveness of the different types of traps, dispersal in the different environments, changes over time. Variability with seed collected in the different traps was visualized using box plots.  

Types of traps and layout of the experiment

 

Trap types were installed at 45 transect locations to determine if there was a significant difference in capture efficiency. A total of 1,164,880 seeds were collected. Among these 44.94% (523,467) of seeds collected were from stations with both ground and mesh traps; while 22.88% (266,518) from ground traps and 20.97% (244,244) from mesh traps. The effectiveness of the traps types was determined by a t-test. This showed there was no statistical difference between the two types of traps (p>0.05, df=118, t=0.089).

 

Floristic composition, species richness and diversity of seed rain

 

A total of 1,164,880 seeds were collected resulting in a final density of 2,107 seeds m² (300 seeds m² week^-1). Eighty-five species from forty-one genera were collected; 65 species were identified to species level, 9 to genera, 6 to family and 5 remained unidentified. The large majority of species were dispersed by wind (50 species), followed by animals (20) including birds and bats (15).

 

Guibourtia copallifera was the most abundant species, followed by Nesogordonia papaveriefera (A.Chev.), Hymenocardia lyrata Tul. and Memecylon normandii Jacq. Leguminosae-Caesalpiniaceae had the highest number of species, followed by Euphorbiaceae, Apocynaceae. Rubiaceae and Sterculiaceae. Species belonging to Rhizophoraceae, Chrysobalanaceae, Ixonanthaceae and Ulmaceae were present in very low numbers. Seeds of G. copallifera made up 58.43% of all seeds and were slightly more likely to be captured by the mesh traps (61% of seeds) compared to the ground traps (54%). Species richness ranged from 3 to 12 species per trap during the collection period.  

 

Status of seeds in traps

 

Seed conditions were classified as intact (80 to 85%), germinated (8 to 12%), insect infested (1 to 2%), decayed (rotten) (2 to 3%) or partly eaten (2 to 4%). Intact seeds germinated in 3 to 7 days but the majority of intact seeds failed to germinate.

 

Seed quality and production are main considerations overseeing the regeneration, structure, also progression of trees in natural forests (Galipeau et al., 1997; Viglas et al., 2013).   Harsh and Joshi (1993) have announced 70% harm to Albizia seeds because of insects and diseases out of which 40% was because of the insects, Bruchus bilineatopygus and B. sparsemaculatus. Diminished viability and germination disappointment of these damaged seeds have been accounted for (Ponnuswami et al., 1990). Singh and Bhandari (1986) have detailed scope obliteration of chilgoza pine seeds in India at Kalpa, Himachal Pradesh, by Dioryctria abeitella Schiff. causing up to 50% seed damage. Essentially, seeds of one more significant species, teak (Tectona grandis) are damaged by Dichocrosis (Conogethes) punctiferalis (Guenee) (Pyralidae) causing up to 70% seed destruction in storage.  Locally, decomposition on forest floor can be constrained by the chemical quality of litter (Krishna and Mohan, 2017; Prescott et al., 2017), microenvironmental conditions (Rodríguez-Paredes et al 2012; Tymen et al., 2012) and soil properties (Li et al., 2018; Kravkaz-Kuscu et al., 2018).

 

Weekly seed collection

 

Weekly seed collection was virtually the same for both trap types; median average 1217.4, while that of mesh is 1227.4 (Figure 1).

 

No significant linear trend was observed for the first 8 weeks the transect traps were set out, but there was then a rapid decline close to zero by week 12. An ANOVA single-factor analysis was performed to confirm that the results were highly significant (p<0.001). A “broken stick” regression was then conducted to show the effective termination of dispersal at week 8 (Figure 2).

 

 

 

Seed dispersal in the forest

 

The seed rain of the forest was highly variable but with an average of 297 seeds m^-2 week⁻¹, but with considerable spatial and temporal patchiness. More than 60 species were represented in the seed rain. These were dominated by a small number of species producing a large seed crop, with most species represented by few seeds during the period. Total seed rain from wind-dispersed versus animal-dispersed tree species did not match their proportions in the forest. Wind dispersed seeds were relatively common compared to others. Additionally, no previous studies have assessed the proportion of seed rain of this species and this forest patch in general.  

Types of traps and layout of the experiment

 

The study shows no statistically significant difference between ground traps and net traps in terms of total seeds collected. Although the total number of seeds is highest in traps with large area, seed density does not vary significantly across traps with different areas which is consistent with Morris et al. (2011). In forest seed rain studies, large seed traps are usually used to capture large seeds of fleshy-fruited and tree species (Vergara-Tabares et al., 2021).   However, our values seem consistent with records from other locations such as Carrillo-Arreola et al. (2020)  who  collected  8,500  seeds m^-2 forest matrix.  Kitamura et al. (2005) observed 22 species of birds and six species of mammals foraging in 15 different species of fruit-bearing plants in Khao Yai National Park in Thailand.

 

Floristic composition, species richness and diversity

 

Results on the species richness and diversity indicate that despite the disturbed nature of the forest there is still a high amount of diversity. G. copallifera seeds were the most abundant, followed by Nesogordonia papaveriefera (A.Chev.), Hymenocardia lyrata Tul. and Memecylon normandii Jac.-Fél. The abundance and seed density values in this study (85 species from 41 plant families) are higher than those in several other studies of primary and secondary tropical forests, 16 species (Chalermsri et al., 2020), 26 species (Wang et al., 2019), 15 species (Ce´sar et al., 2017), 40 species (Sritongchuay et al., 2014). Similar levels of diversity are the 65 species recorded in the Serra da Bodoquena National Park (Brachtvogel et al., 2020), 56 species (Dhillon et al., 2003) and 46 species by Villicana- Hernandez et al., (2020)  in Yucatan, Mexico.  

 

Seed density was found to be higher in this study relative to two similar studies in semi-deciduous tropical forest for Cameroon (≅297 seeds.m-². year^-1; Hardesty and Parker, 2002 Nuñez  et al., 2021;), and the southeast of Brazil (≅442 seeds.m-². Year^-1; Grombone-Guaratini and Rodrigues, 2002). The higher number of seeds found in this study reflect that the large number of mature trees producing a high number of seeds produced by G. copllifera (Peili et al., 2019; Wickert et al., 2017). Peak abundance of tree seeds coincides with the start of the rainy season, while seeds of other herbaceous species peak at other times.

 

Human disturbances in this forest also strongly affect the plant community structure; more specifically, G. copallifera species are gaining dominance in non-degraded areas and losing dominance in degraded areas. The seed dispersal distance for this species was strongly correlated with the plant than with seed traits (Augspurger et al., 2017; Thomson et al., 2011). Indeed, the dispersal of G. copallifera seeds in our study was largely due to plant character more than the seed size. The parent plant crown and the height positively influenced the seed dispersal of this species.

 

Status of seeds in traps

 

Conditions of collected seeds in traps varied depending on the several prevailing factors like rainfall, wind, humidity, sunlight, germination rate and volume of leaf litters. Damage was less in the mesh traps as a result of the fact that ground trap is being affected by run off, ground animals and insect prey which reduces the evidence on the ground compared to mesh trap which drains water and is more difficult to access by terrestrial seed predators.

 

Germinated seeds were recorded even in the absence of soil especially for G. coppallifera.  Regarding seed decay, the principal factors thought to control paces of litter decay include barometrical relative humidity (Gregorich et al., 2017; Beyaert and Voroney, 2011), the herbaceous plant layer (Ossola et al., 2016; Zirbel et al., 2017), and soil temperature and dampness (Cortez 1998; Mao et al. 2018; Sun and Zhao, 2016).  Song et al. (2021) observed that decomposition rate of leaves increased under mats of exotic weed Tradescantia fluminensis Vell.. This is attributed to a positive microclimate and living space for the decomposers.  Studies on seed insect pests, conducted at the Pakistan Forest Institute, Peshawar, revealed that seeds of 30 out of 70 tree species were infested with 24 insect species. Among them 86% were Coleoptera, and 5.6% were Lepidoptera, while 8.4 % was a hymenopteran parasite (Rehman, 1993).  

 

Weekly seed collection

 

The number of seeds captured weekly on the interior and edges of the forest seem to coincide with the frequency of monthly wind, sun and rain data for that area.  More seeds are captured in weeks with more frequent rainy days and less in less windy and raining days in the weeks. When the rate and duration of precipitation increased in September, the data showed a significant decrease in the number of seeds captured particularly for G. copallifera, since this was at the end of its fruiting season. The temporal variability of seed rain related to rainfall has been reported by other studies (Gummadi et al., 2017; Perini et al., 2019).

 

Seed rain was observed to be greater in the interior of the forest than at the edges. Studies on the interior and edges observed that seed rain on both sites was higher along the forest edge, also observed by other studies (Diogo et al., 2015; Dunham et al., 2018; de Melo et al., 2006; Magrach et al., 2014; Razafindratsima et al., 2021). The low abundance and diversity of seed captured in the degraded area shows a limitation on seed dispersal either due to felling of trees, and or increasing human activities. Similar results by other studies for seed dispersal from forests adjacent to pasture found that seeds were mostly anemochory and limited to 10m distance (Aide and Cavelier, 1994), 5 m (Holl,1999), and 4 m (Zimmerman et al., 2000; Cubiña and Aide, 2001) from the forest edge.

 

The fact that the seed rain dynamics in both forest edge and interior was affected by the surrounding tree heights, number of trees, species and intensity of human activities, shows that matured and taller trees produce more seeds and have a higher chance to disperse more seeds (Thomson et al., 2011).

 

Seed rain numbers

 

Over sixty species were shedding seed at the same time as G. copallifera, however, it is possible that some species have been missed due to human error (misclassification or identification), low density or being physically restricted to small areas in the forest; additional species disperse at other times of the year. G. copallifera seeds dominated seed rain traps.  Numbers of seed dispersed by this species is plentiful on the forest floor. 

 

Relationship between distance and dispersal frequency

 

The most general and perhaps obvious dispersal pattern found in the study was the fairly uniform seed rain out to about quarter of a kilometer from the seed source (Figure 3). The results illustrate how an analysis of a species seed dynamics can contribute to a better understanding of the general ecological phenomena, including species dispersal and spread. This is an indication to the fact that this species has the ability to establish itself in distant locations in the absence of the parent plant.  The proportion of seed that is randomly dispersed over relatively long distances from the source plants of G. copallifera ranged from 5 to 50 % of the total seed set. Despite this long-distance dispersal advantage of G. copallifera seeds, its’ establishment in the new area is opposed by predation, soil conditions, depth of seed in soil / leaflitters and human activities. 

 

G. copallifera has an extremely restricted geographical distribution in Sierra Leone with the majority of the population in a single small and degraded forest reserve. Disturbance to the forest through charcoal burning allows the pioneer species to be dominant and in some grass patches which form fire-climax vegetation communities are establishing. Our results showed that G. copallifera produces a lot of wind dispersed seeds and densities of up to 70 to 100 per sq meters were observed 250 meters from the forest edge. Only 0.8% of the seeds were found to be empty. Dispersal of G. copallifera seeds would seem to be sufficient to allow it to maintain its population in the forest but not to disperse to other similar forest patches which are many kilometers away.  

 

The results obtained demonstrate that Kasewe Forest Reserve still has a high natural regeneration potential for G. copallifera; nevertheless, there are several causes for concern which will be critical for continued natural regeneration; the on-going disturbances have reduced tree cover, increased the number of clearings and increased the availability of light, thereby facilitating the dominance of pioneer species. Thus, the national and district authorities must be careful for the implementation of the conservation measures and enhance effective co-management and protection program involving the local people to ensure fruitful conservation of this species in its most abundant site. Protection and aided natural regeneration may be another alternative option for effective natural regeneration and conservation of this species especially in degraded habitats within the reserve.

The authors have not declared any conflict of interests.

We thank the staff of the Sierra Leone National Herbarium in the Department of Biological Sciences for their technical support. We are grateful to the Ministry of Forestry and Food Security for granting us permission to conduct this research in Kasewe Forest Reserve.  We would like to thank Andrew Kamanda, Josaphus Junan, Joseph Longha, Janet Saidu, Emmanuel kamanda Kinnie Jarawallay for their assistance with fieldwork and data collection. Special thanks to Maxwell Williams and Emmanuel Saidu for data compilation and input.