Social, economic, and environmental challenges have resulted in agricultural land abandonment. Agricultural land abandonment (hereafter old fields) can have both positive and negative consequences (Subedi et al. 2022). For example, old fields improve carbon sequestration and native plant revegetation, but they are also associated with invasive alien plants and wildfires that negatively affect ecosystem services (Subedi et al. 2022). To alleviate the negative consequences of old fields on ecosystem services and human well-being, calls for old field ecosystem restoration have intensified (Subedi et al. 2022).
Old field restoration can take place either passively (unassisted natural revegetation) or through active human intervention (assisted). Unassisted natural revegetation is regarded as slow in achieving native plant recovery compared to assisted active restoration because it is dependent on the natural factors such as availability of soil seed banks and seed dispersal (Parkhurst et al. 2022). Therefore, assessing soil seed bank availability and viability in old fields targeted for unassisted natural revegetation is important. Globally, examinations of soil seed bank in old fields have been mixed, demonstrating both seed bank presence and absence (Heelemann et al. 2013). Few studies have been conducted on soil seed banks in South Africa, highlighting the need for more studies to inform potential old field recovery trajectories (Mndela et al. 2020). To gain a better understanding of the South African context, this study assessed the role of soil seed banks on the revegetation of species diversity and composition in old fields targeted for natural recovery. This study is motivated by the fact that old fields natural revegetation is partially dependent on whether soil seed bank is available; if the soil seed bank is depleted, then assisted active restoration is needed.
This study was conducted in old fields located at Nanaga farm (33°35′41.3″ S, 25°55′44.9″ E) in the Eastern Cape province of South Africa (Figure 1A). Soil cores for the soil seed bank study were collected on study sites where Matimu and Ruwanza (2023) conducted above-ground vegetation surveys. Topsoil was collected from three old field sites adjacent to intact natural vegetated reference areas. The reference areas acted as the control and are on same farm but were never cultivated (Figure 1B). The three old fields were approximately 500 m apart, whereas the distance between the paired intact natural vegetated reference areas and the old fields was approximately 10–15 m (separated by farm roads).
Map of study area showing (A) the study location in Eastern Cape province of South Africa and (B) the old fields at Nanaga farm.
Due to changes in farm ownership, information regarding past cultivation of these old fields is anecdotal, but it is assumed that they were abandoned more than 20 years ago. The old fields are mostly used for wildlife grazing by Cape Mountain Zebra (Equus zebra zebra) and Impala (Aepyceros melampus). The old fields have ridges that are approximately 1-m high and 1-m wide and furrows that are approximately 15-m wide. The ridges and furrows were artificially created to reduce erosion, and cultivation was mostly done on the furrows. The intact natural vegetated reference areas are dominated by Albany Thicket Biome vegetation (Mucina and Rutherford 2006). Soils in the study area are predominantly sandy, temperatures are mild (mean annual temperature is approximately 17.8°C), and rainfall is non-seasonal (mean annual rainfall is approximately 750 mm) (Mucina and Rutherford 2006).
At each of the sites, we collected soil seed bank cores from the topsoil. Within the study site, we established a 100-m transect positioned on the ridge, furrow, and reference areas. The transacts were parallel to the old field ridges and furrows, starting 10-m from the site edge and stretching east to west. At each transect, five replicated plots measuring 5 m2 (1-m wide × 5-m long) were set up. Individual plots were separated by a 10-m buffer. In total, 45 plots were set up (5 plots per transect × 3 treatments [ridge, furrow, and reference] × 3 sites).
Within each plot, soil cores measuring 10 cm in diameter at a depth of 10 cm were collected in late summer (February 2022). The above soil depth is based on previous studies by Ruwanza (2022) which suggested that topsoil depth in Nanaga farm is approximately 8–10 cm. Soil sample collection was done in late summer to capture both the long buried and recently dispersed seed banks. Our soil sampling time is associated with seed rain of most plants in austral summer. Within each plot, five soil cores were collected from the four corners of the plot and at the center, and these were bulked to constitute a plot sample. Sample bulking was done to reduce seed variability clustering and irregular distribution. Collected soil cores were sieved using a 10-mm mesh to remove litter, stones, and plant roots.
The soils were placed in trays measuring 15-cm wide × 20-cm long × 10-cm deep and placed in a passively ventilated greenhouse at Rhodes University, where the air temperature is close to the outdoors. In the greenhouse, trays were arranged based on sites into five columns and nine rows. The trays were rotated monthly to account for greenhouse light penetration and air temperature variations. The trays were watered by hand (whenever required) for nine months (February to October). Emerging seedlings were counted and identified as close to species level as possible. Counted and positively identified seedlings were removed from the trays to avoid duplicate counting. Seedlings that could not be identified using a plant species list generated by Matimu and Ruwanza (2023) were identified at Selmar Schonland Herbarium at Rhodes University in Makhanda. Seedlings were classified based on phytomorphology as trees, shrubs, forbs, and graminoids. The number of seedlings per tray were converted to density (the number of seedlings per m2).
Using seedling counts, we calculated species richness (S), Shannon-Wiener (H’), evenness, and Simpson’s index of diversity per tray. To test for significant differences in the above-mentioned measurements across ridges, furrows, and reference areas, we used a one-way ANOVA in TIBCO STATISTICA version 14.0, since data were normally distributed. To evaluate species similarity between the soil seed bank and above-ground vegetation data collected by Matimu and Ruwanza (2023), we computed the Sørensen’s similarity community coefficient (S) index as follows: S = 2C/(A + B), where A (soil seed bank) and B (the aboveground vegetation) are the number of species in each plot, and C is the number of species common in both. Principal component analysis (PCA) was used to investigate changes in species composition across old field areas using CANOCO for Windows 5 on seedling presence/absence data for the seedlings of 25 frequently occurring species.
Species abundance and density from soil seed banks were higher on ridges compared to furrows and reference areas, but with no statistical (p > 0.05) differences across the three areas (Table 1). Seed bank species richness, Shannon-Wiener, and evenness showed significant (p < 0.05) differences across the three areas, being higher on furrows compared to ridges (Table 1). Simpson’s index showed no significant (p > 0.05) differences among the three areas (Table 1). Seed bank species richness for trees and shrubs was significantly (p < 0.001) higher on reference areas compared to ridges and furrows (Table 1). In contrast, seed bank species richness for forbs and graminoids was significantly (p < 0.001) higher on furrows compared to ridges and reference areas (Table 1). The Sørensen’s index showed lower correspondence of species between the seed bank and above-ground vegetation, an indication that the paired communities shared different species composition (Table 1). However, across the three areas, the reference areas had the highest species similarities compared to the ridges and furrows.
Comparisons of seedling diversity indices among different old field sites. Data are means ± se and one-way ANOVA results are shown. Means with different letter superscripts are significantly different at p < 0.05.
A total of 25 different species (six trees and shrubs, 12 forbs, and seven graminoids) germinated across all areas, with the bulk being native species (19) compared to six alien species (three forbs and three graminoids) (Figure 2). We enumerated 2,256 individual plants across all areas, of which 838 occurred on ridges, 770 on furrows, and 648 on reference areas. Across all areas, four native species were not observed on ridges (Cussonia spp., Oxalis smithiana, Andropogon eucomus, and Ehrharta spp.), whereas five native species (Asparagus spp., Selago cinerea, Secamone alpine, Andropogon eucomus, and Ehrharta spp.) and one alien species (Lactuca spp.) were not observed on reference areas. Principal component analysis biplots showed clear species separations, with species such as Cymbopogon marginatus, Drosanthemum hispidum, Hermannia althaeoides, Diospyros spp., Asparagus spp., and Lactuca spp. dominating both ridges and furrows (Figure 2). The first two PCA axes for all species showed low eigenvalues and accounted for 25% of the variance.
Principal Component Analysis (PCA) biplots of plant species (▲) from ridge (●), furrow (■), and reference (♦) old field sites for seedlings of 25 frequently occurring species. Axes 1 and 2 explain 15.68% and 24.76% of the total species variability, respectively.
We investigated the presence and importance of soil seed banks in the revegetation of old fields under different conditions (ridges, furrows, and reference areas) and demonstrated that soil seed banks do exist in the Nanaga farm fields. Although old fields typically are regarded as heavily degraded ecosystems due to past cultivation, the observed soil seed bank similarities across the three areas present evidence that, even 20 years after cultivation abandonment, soil seed banks can potentially initiate unassisted natural revegetation of native species. Matimu and Ruwanza (2023) reported that vegetation in these old fields was recovering spontaneously; this above-ground vegetation is likely the largest contributor to the soil seed bank. However, Sørensen’s similarity community coefficient index between the soil seed bank and the above-ground vegetation showed lower similarity, an indication that not all seeds are from the existing vegetation. It is possible that some seeds are being dispersed from the nearby intact reference areas. Seed dispersal from intact reference areas to old fields has been reported in the past and is prevalent where the areas are in proximity (Matimu and Ruwanza 2023). Our study areas were separated from native vegetation by farm roads, so seed dispersal from the reference areas to the old fields is highly possible. In this study, 19 seed bank species were present on old fields and reference areas, suggesting that dispersal between the two areas is taking place.
Observations in the old fields alone showed that soil seed bank species diversity was higher in the furrows compared to the ridges; however, a close analysis of the data showed that this was more prevalent for the forbs and graminoids that dominated the furrows. The dominance of forb and graminoid species in furrows compared to ridges could be due to several factors, such as legacy effects from past cultivation that seem to favor recruitment of fast-growing forbs and graminoids in previously cultivated furrows (Ruwanza 2022). This is also likely to explain the presence of some alien forbs and graminoids. Since the study area is susceptible to wildlife grazing, continuous grazing of furrows rather than ridges could also explain why forbs and graminoids dominate furrows through ungulate-mediated plant dispersal.
Our results indicate that the soil seed banks of some shrubs such as Diospyros species dominated both ridges and furrows, whereas others like Searsia lucida dominated ridges only. It is not clear why some species occur in particular old field areas; however, this disparity could be due to seed dispersal mechanisms and the availability of appropriate seed germination conditions between the two areas. Overall, the presence of tree and shrub species seeds in both ridges and furrows is welcome because it points to the positive successional trajectory towards the original Albany Thicket vegetation type. We conclude that both the old field ridges and furrows have diverse soil seed banks that could trigger unassisted natural revegetation of Albany Thicket Biome vegetation. Vegetation recovery of the unique Albany Thicket Biome is important to sustain a higher degree of plant diversity which in turn supports human livelihoods. Under the current Nanaga old field conditions, assisted vegetation recovery (active restoration) is not needed since spontaneous soil seed bank driven unassisted natural revegetation is taking place.
Acknowledgements
Funding was provided by South African National Research Foundation (NRF Research Grant UID 137789). The authors thank the Department of Environmental Science and Botany Department at Rhodes University for providing equipment and tions can provide substantial support for local, regional, and national economies by increasing the populations of greenhouse space, respectively.
