Skip to main content
NCBI home page
Plant Diversity logoLink to Plant Diversity
. 2026 May 15;48(5):884–892. doi: 10.1016/j.pld.2026.05.003

Latitudinal diversity gradients of soil seed bank and aboveground vegetation in China

Yantong Zhao 1, Yangqing Luo 1, Shilu Zheng 1, Tiantian Zhang 1, Jiajia Liu 1,⁎
PMCID: PMC13615778  PMID: 42801213

Abstract

The latitudinal diversity gradient is one of the most general biogeographical patterns. However, current latitudinal diversity gradients have been mostly focused on aboveground vegetation, while such a pattern exists in soil seed bank, which is an irreplaceable natural resource that provides multifaceted benefits to terrestrial biodiversity, is not well understood. Here, we compiled plant diversity data of soil seed bank and corresponding aboveground vegetation from 496 sample sites in China, to detect whether the latitudinal diversity gradient of soil seed bank differs from that of aboveground vegetation. We found that soil seed bank exhibited a significant but weaker latitudinal diversity gradient compared to aboveground vegetation, and the latitudinal diversity gradient of soil seed bank was not observed for ecosystems dominated by herbaceous plants. In contrast to increasing species diversity of aboveground vegetation towards the tropics, species diversity in soil seed bank remained stable across the latitudinal gradient for grasslands and wetlands. Moreover, environmental factors associated with species diversity of soil seed bank and aboveground vegetation were different, with mean temperature and temperature range as the major driver of species diversity in soil seed bank while elevation and temperature seasonality as the major drivers of aboveground vegetation species diversity. Our results provide empirical evidence that biogeographic pattern of plant diversity generated from aboveground biota do not necessarily reflect that of belowground diversity, and aboveground-focused conservation may fail to protect plant diversity stored in soil seed bank.

Keywords: Soil seed bank, Aboveground vegetation, Latitudinal diversity gradient, Soil biodiversity, Species composition, Belowground community

Highlights

  • •

    Soil seed bank show a weaker diversity change with latitude than aboveground plants.

  • •

    Herbaceous-dominated seed bank diversity stays constant across latitudes.

  • •

    Aboveground conservation efforts may not protect unique diversity in soil seed bank.

1. Introduction

The latitudinal diversity gradient is one of the most general biogeographical patterns, with increasing species richness from the poles towards the tropics (Hillebrand, 2004; Mittelbach et al., 2007). Such pattern has been widely reported across aboveground taxa, including terrestrial plants (Kreft and Jetz, 2007; Kerkhoff et al., 2014), birds (Jetz et al., 2012), mammals (Davies et al., 2008), amphibians (Pyron and Wiens, 2013) and insects (Economo et al., 2018). Hypotheses such as increasing species-carrying capacity (Kerkhoff et al., 2014; Usinowicz et al., 2017), diversification rates (Jablonski et al., 2006; Mittelbach et al., 2007), as well as long-term climate stability towards the tropics (Francis and Currie, 2003; Kreft and Jetz, 2007), are proposed to explain this spatial variation in diversity. In contrast, a growing number of studies highlight that latitudinal diversity gradient does not necessarily exist for belowground taxa, such as soil bacteria (Fierer and Jackson, 2006), fungi (Tedersoo et al., 2014), and earthworms (Phillips et al., 2019). Therefore, belowground biodiversity may not be parallel to the aboveground components. However, whether such a mismatch exists in plant diversity remains unknown, mainly due to the lack of paired data on belowground and aboveground diversity at a large scale.

Soil seed bank, comprised of viable seeds stored in the soil, is an important component of belowground biota and plays an important role in shaping aboveground plant diversity (Thompson and Grime, 1979; Simpson et al., 1989). For example, soil seed bank can promote species coexistence (Chesson, 2000), prevent local extinction (Gioria and Pyšek, 2016), and enhance ecosystem resistance and resilience (Kiss et al., 2018; DeMalach et al., 2021; Plue et al., 2021; Auffret et al., 2023). Soil seed bank can be beneficial to the survival and expansion of plant populations (Gioria et al., 2020), and maintain regional biodiversity by facilitating regeneration rapidly in degraded ecosystems (Poorter et al., 2021). Therefore, understanding the distribution patterns and drivers of soil seed bank will provide novel insights into the mechanisms governing large-scale plant diversity.

Soil seed bank represents the results of a balance between seed rain, persistence and germination, and may lead to a different latitudinal diversity gradient from that of aboveground vegetation. First, soil seed bank is often dominated by herbaceous species with high seed production and small seed size, which could survive in cold soils and have prolonged storage (Warr et al., 1993). Second, soil seed bank diversity is strongly controlled by microhabitat variables (Gioria et al., 2020; Zou et al., 2021; Gasperini et al., 2021), while climatic variables play more important roles in aboveground vegetation (Kerkhoff et al., 2014). For example, soil nutrients (nitrogen and carbon), which are higher in temperate regions compared to tropical ones (Post et al., 1985), can significantly affect seed persistence and germination rate and hence modulate the latitudinal diversity gradient of soil seed bank (Pichon et al., 2023). Therefore, consistent with the latitudinal gradients in herbaceous species diversity (Taylor et al., 2023), species diversity in soil seed bank may peak in mid-latitude regions and exhibit a distinct pattern compared to aboveground vegetation.

Recent studies of soil seed bank diversity patterns have suggested a weak latitudinal diversity gradient in soil seed bank in temperate wetlands in the United States (Lei and Middleton, 2021) and temperate forests in north-western Europe (Plue and Cousins, 2013). Indeed, a global meta-analysis of 130 grassland soil seed bank confirmed a weak latitudinal diversity gradient (Jabot and Pottier, 2017). In contrast, based on a global dataset including different ecosystem types, Yang et al. (2021) found a rather strong latitudinal diversity gradient in soil seed bank. These contradicting results may be due to the lack of references for aboveground diversity, as local plant community and regional source pools can significantly influence the diversity of local seed bank. However, a comprehensive study examining whether there are differences in biogeographical patterns of soil seed bank and aboveground plant diversity is lacking mainly due to the lack of sufficiently paired data. This literature gap is critical to fill because resources for plant biodiversity conservation will always be limited and could be poorly targeted without knowledge of whether soil seed bank is represented by aboveground vegetation.

Here, we present a synthesis of plant diversity of belowground soil seed bank and corresponding aboveground vegetation for 496 sample sites with a total of 2484 species in China, spanning a latitudinal range from 21.6°N to 53.5°N. We aim to investigate differences in the latitudinal diversity gradient between soil seed bank and aboveground vegetation, distinguishing between different ecosystem types, and to assess the relationships between latitudinal diversity gradient and environmental drivers for both belowground seed bank and aboveground plants. We couple investigations on soil seed bank and aboveground vegetation in the same community and ask: (1) Does soil seed bank exhibit a similar latitudinal diversity gradient compared to aboveground vegetation? (2) Which factors influence the differences in geographic patterns of plant diversity belowground and aboveground? By simultaneously considering the plant diversity of both aboveground and belowground components, our study provides a critical assessment of the biogeographical patterns of belowground biodiversity and will be useful for decision makers aimed for effective biodiversity conservation.

2. Method

2.1. Data compilation

We complied plant diversity data of soil seed bank and corresponding aboveground vegetation from field studies in China, which were conducted following a standard germination protocol (Yu and Jiang, 2003). We obtained data of soil seed bank by searching studies published during 1990 and October 2022 via the Web of Science (https://webofscience.clarivate.cn). The keywords used were (‘soil seed bank’ OR ‘soil seedbank’) AND (‘diversity’ OR ‘richness’ OR ‘species composition’). To include relevant studies published in Chinese, we performed an additional literature search using same keywords in the largest searching engine of China, the China National Knowledge Infrastructure (CNKI; https://www.cnki.net). Conference proceedings and review publications were excluded. At this stage, we integrated 542 local studies forming a Chinese soil seed bank database including 2885 sample sites and 3126 species.

To compare differences in species richness between soil seed bank and aboveground vegetation of natural ecosystems, we manually checked each article and only included studies: (1) that were conducted in undisturbed ecosystems or with duration of secondary succession > 50 years without human interference; (2) which contained species richness and abundance data or relative abundance within the sampling unit; (3) which collected at least three soil cores per site to avoid contingency in species richness data of soil seed bank; (4) of which data were obtained without experimental manipulation (e.g., light, temperature, moisture, and nutrient); (5) with location information; (6) that were conducted in terrestrial ecosystem. Finally, we removed duplicate data from our dataset. In total, we collected 829 species diversity data of soil seed bank and aboveground vegetation from 182 studies (106 studies published in Chinese and 76 in English), with study sites spanning a latitudinal range from 21.6°N to 53.5°N and a longitudinal range from 101.5°E to 123.5°E (Fig. 1a). Specifically, forests (n = 194), grasslands (n = 229) and wetlands (n = 406) were the three most studied ecosystem types in the comprised dataset. In addition, there were a total of 2484 plant species across all sites, including 1126 woody plant species (trees, shrubs, and woody lianas) and 1358 herbaceous plant species (herbs, and herbaceous lianas).

Fig. 1.

Fig. 1

Open in a new tab

Species diversity patterns of soil seed bank in China. (a) Distribution pattern of species richness in soil seed bank in China. Points on the map show the geographical locations of surveyed communities (n = 494 sites). (b) Box plots show species richness in soil seed bank and aboveground vegetation (n = 494 in soil seed bank and n = 335 in aboveground vegetation). .

We implemented standardized procedures to ensure the comparability of all data collected with different soil depths. Species diversity in soil seed bank generally decreases with soil depth (Thompson and Grime, 1979; Fenner and Thompson, 2005). Therefore, 0–10 cm was the most reported and was chosen as the soil depth in the calculation of species diversity of soil seed bank (Data S1). Details about study design and site characteristics such as location, sample size, soil depth and duplication, and method of detecting seeds were included in the database (Data S1).

Our dataset represents a significant expansion in both depth and scope compared to previously published database (Yang et al., 2021; Auffret et al., 2024). While previous focused primarily on community metrics such as total seed density and species richness, our database provides species-level resolution, documenting the identity and specific abundance or density of each species in each site. Furthermore, we recorded comprehensive methodological metadata and corresponding diversity in paired aboveground vegetation, enabling a direct assessment of relationship between below- and aboveground plants. To facilitate environmental analyses, we also integrate site-specific environmental variables, including climate, topography, and soil properties. Full references for all primary data sources are provided in the Supplementary Database (Data S1).

2.2. Species richness data

We used species richness, which intuitively represents the number of species in an assemblage, as an indicator of species diversity. To account for potential biases associated with species abundance, we additionally calculated the Shannon index for soil seed bank based on seed density of each plant. To evaluate the relationship between the diversity of soil seed bank and the corresponding aboveground vegetation along latitude, we obtained paired species richness data from both soil seed bank and aboveground vegetation, according to the following criteria: (1) species richness in one community with soil seed bank were reported simultaneously; (2) species composition data including relative abundance or percentage per species per sampled community; (3) aboveground vegetation surveys published in another journal or dissertation were matched with soil seed bank sampling at the similar time (within three years) by the same authors. Studies without detailed species lists of aboveground vegetation were excluded from the analysis of species composition. In total, we obtained 496 soil seed bank and 335 aboveground vegetation sample sites with plant diversity data, within which 326 samples from 137 studies had both soil seed bank and aboveground vegetation diversity data (Data S1).

Species richness and seed density data were extracted from each study through text, tables, figures, and supplementary materials. We used Plot Digitizer (https://plotdigitizer.com/) to extract the data when species richness was only presented in figures. To ensure consistent plant taxonomy across sites, we standardized scientific names based on taxonomic information from Species (2000) (http://www.sp2000.org.cn/), which provides both scientific and Chinese names of species. In total, the database included 2484 species, with 1551 found in soil seed bank and 1779 in aboveground vegetation.

2.3. Environmental variables

To assess relationships between latitudinal diversity gradient and environmental drivers for both belowground seed bank and aboveground plants, we included climate, soil and topographic variables for each site, which have been reported to be associated with species diversity of aboveground vegetation and belowground soil seed bank (Hoyle et al., 2013; Plue and Cousins, 2013; Gasperini et al., 2022). We selected mean annual temperature, mean annual precipitation, temperature seasonality, temperature annual range, and precipitation seasonality as climate variables extracted from the WorldClim database v.2 for 1970–2000 (Fick and Hijmans, 2017). As the most important topographic variable, elevation was extracted from the WorldClim database v.2 (Fick and Hijmans, 2017). The soil texture (clay content, sand content, silt content, coarse fragments volumetric) and fertility (cation exchange capacity, total nitrogen, soil organic carbon density, soil organic carbon stock, soil organic carbon content) variables were retrieved from the SoilGrids250m (Hengl et al., 2017). In addition, we obtained soil moisture data from the National Tibetan Plateau/Third Pole Environment Data Center (Yao et al., 2021), which have been found to be associated with species richness of soil seed bank (Zou et al., 2021; Gasperini et al., 2021). All environmental layers were standardized to a 30 arcsec resolution (∼1 km2 at the equator), and all soil variables used were determined by the top layer of soil at a depth of 0–5 cm.

2.4. Statistical analyses

To test differences in latitudinal diversity gradients between soil seed bank and aboveground vegetation, we applied linear mixed-effects models (LMMs) on species richness with a Gaussian error distribution using the ‘lme4’ package (Bates et al., 2015). Species richness was loge-transformed before analyses because of high skewness of distribution. In each model, we treated species richness as the response variable, the latitude of each sampling site, community group (soil seed bank or aboveground vegetation) and the interaction between these two variables as fixed predictors. Study ID was treated as a random intercept to account for potential non-independence among observations within each study. Residual spatial autocorrelation was tested to improve the accuracy of models involving spatial variables.

The similarity of species composition between soil seed bank and aboveground vegetation can be used to predict vegetation change, and low similarity indicates a higher resistance and resilience in plant communities (Hopfensperger, 2007). To understand the geographical pattern of the similarity between soil seed bank and aboveground vegetation, similarity index from original research was also extracted and unified as Sørensen similarity index according to reported data about the number of common species, total species separately found in soil seed bank and aboveground vegetation. For samples without direct similarity data, but with available information on species composition, the Sørensen index was calculated by using the ‘vegan’ package (Oksanen et al., 2022).

To assess the role of multiple environmental factors in driving the geographic distribution patterns of soil seed bank and aboveground vegetation, we fit the linear mixed-effects models (LMMs) for species richness and environmental variables. To address multicollinearity among soil properties, we performed two separate Principal Component Analyses (PCAs) for soil texture and soil fertility. We extracted the first principal component (PC1) from each analysis to represent these characteristics. The extracted PC1 for soil texture and soil fertility accounted for 52% and 62% of the total variance within their respective variable sets. We considered four groups of predictors, spanning a range of spatial scales in each LMMs containing the following terms: (i) topography: elevation; (ii) climate factors: mean annual air temperature, mean annual precipitation, temperature seasonality, temperature annual range, and precipitation seasonality; (iii) soil factors: mean annual soil moisture, soil texture, and soil fertility; (iv) seed diversity: seed density for soil seed bank. Preliminary analysis revealed high collinearity between temperature seasonality and temperature annual range (Variance Inflation Factor > 10), preventing their simultaneous inclusion. To determine which of environmental variables better explained the latitudinal diversity gradients, we conducted a model comparison approach using AICc (corrected for small sample sizes). In each LMM, species richness was the response variable, with environmental variables as the fixed predictors and study ID as a random intercept. All predictors and response variables were scaled to allow the comparison of effects and responses between groups. All statistical analyses were performed using R program v.4.3.2 (R Core Team, 2025).

3. Results

Species richness of soil seed bank was significantly lower than that of aboveground vegetation, with an average of 21.0 ± 15.5 (mean ± s.d.) in seed bank and 33.5 ± 30.4 in aboveground vegetation (n = 829; P < 0.001; Fig. 1b and Table S1). Moreover, the difference was higher in forest ecosystems (33.8 ± 21.4 in seed bank and 56.5 ± 43.4 in aboveground vegetation; n = 194; P < 0.001), and lower in grassland and wetland ecosystems (15.5 ± 11.0 and 19.1 ± 12.1 in seed bank, 20.7 ± 12.5 and 26.3 ± 16.4 in aboveground vegetation; n = 270 in grassland, P = 0.001; n = 528 in wetland; P < 0.001; Fig. 1b).

Our results found that strikingly contrasted latitudinal diversity gradient between aboveground vegetation and soil seed bank, which further varied across ecosystem types (Fig. 2). While species diversity of aboveground vegetation increased significantly with latitude for different vegetation types, the soil seed bank exhibited weaker latitudinal diversity gradients relative to aboveground vegetation (n = 748; interaction t-value = 3.03, P = 0.003; Fig. 2a and Table S2). Specifically, there was a significant latitudinal diversity gradient for woody species in forests (n = 182, P < 0.001), but the slope of latitudinal diversity gradient in soil seed bank decreased slower compared to aboveground vegetation (interaction t-value = −0.15, P = 0.136; Fig. 2b and Table S2). The latitudinal diversity gradient of soil seed bank was not observed for ecosystems dominated by herbaceous plants (n = 182 in forest herbaceous layer; n = 229 in grassland; n = 406 in wetland; Fig. 2c–e). In addition, we found a significant decrease in the slope of the latitudinal diversity gradients from aboveground vegetation to soil seed bank for herbaceous species in forests (interaction t-value = −3.26, P = 0.001; Fig. 2c), grasslands (interaction t-value = 3.96, P < 0.001; Fig. 2d), and wetland ecosystems (interaction t-value = 2.16, P = 0.031; Fig. 2e).

Fig. 2.

Fig. 2

Open in a new tab

Latitudinal diversity gradient of soil seed bank and aboveground vegetation. The latitudinal diversity gradient for overall pattern (a, n = 748), woody and herbaceous species in forests (b and c, n = 182), grasslands (d, n = 229), and wetlands (e, n = 406) for soil seed bank (brown) and aboveground vegetation (green). Lines represent model predictions with shaded areas representing standard errors. The solid line indicates a significant difference and the dashed line indicates no significant difference between soil seed bank and aboveground vegetation. The asterisks indicate the statistical significance level of the difference in latitudinal diversity gradient between soil seed bank and aboveground vegetation.

Similar patterns were observed when species diversity was quantified using the Shannon index, confirming the consistency of the latitudinal trend across different diversity facets (Fig. S1 and Table S3). We further verified that the observed weak latitudinal diversity gradient was not confounded by variations in sampling effort across studies. Sampling area showed no systematic relationship with latitude (Pearson's r = 0.076, P = 0.099; Fig. S2). The divergence of latitudinal diversity gradient between soil seed bank and aboveground vegetation remained robust accounting for sampling area across different ecosystems (interaction: β = 0.017, t = 3.01, P = 0.003). Moreover, species–area relationship across different latitudinal bands revealed inconsistent correlations (Fig. S2), demonstrating that our findings reflect generally ecological patterns rather than methodological artefacts.

Across 275 paired-site surveys, community composition similarity between soil seed bank and aboveground vegetation was positively correlated with latitude (slope = 0.008, SE = 0.002, P < 0.001; Fig. 3a), especially in forest (slope = 0.013, SE = 0.002, P < 0.001; Fig. 3b), though the pattern was not observed in grassland and wetland ecosystems (P = 0.69 in grassland, P = 0.98 in wetland; Fig. 3c and d). The results remained highly consistent with the Simpson index also increasing significantly with latitude (slope = 0.01, SE = 0.001, P < 0.001; Fig. S3).

Fig. 3.

Fig. 3

Open in a new tab

Similarity in species composition between soil seed bank and aboveground vegetation.

The similarity of community composition between soil seed bank and aboveground vegetation for overall communities (a, n = 275), forests (b, n = 88), grasslands (c, n = 85), and wetlands (d, n = 102). Lines represent model predictions, with shaded areas representing standard errors. The asterisks indicate the statistical significance level of the relationship between similarity and latitude.

Species richness in soil seed bank was negatively associated with temperature annual range (β = −0.42, SE = 0.12, P < 0.001), mean annual temperature (β = −0.39, SE = 0.13, P = 0.002), precipitation seasonality (β = −0.14, SE = 0.07, P = 0.042), and soil moisture (β = −0.08, SE = 0.04, P = 0.031), while exhibiting positive relationships with seed density (β = 0.16, SE = 0.02, P < 0.001), soil texture (β = 0.08, SE = 0.03, P = 0.002), and soil fertility (β = 0.06, SE = 0.03, P = 0.031) (Fig. 4a; Table S5). In contrast, species richness of aboveground vegetation declined with increasing elevation (β = −0.52, SE = 0.14, P < 0.001), temperature seasonality (β = −0.52, SE = 0.14, P < 0.001), mean annual temperature (β = −0.39, SE = 0.16, P = 0.015), but increased with soil moisture (β = 0.11, SE = 0.05, P = 0.042) (Fig. 4b and Table S5).

Fig. 4.

Fig. 4

Open in a new tab

Environmental variables associated with diversity patterns of (a) soil seed bank and (b) aboveground vegetation. Dots represent standardized coefficients and bars represent 95% confidence intervals from linear mixed models. Brown indicates variables significantly influencing soil seed bank diversity, green indicates variables significantly influencing aboveground vegetation diversity, and grey indicates non-significant variables in their respective models.

4. Discussion

While plant diversity of aboveground vegetation increased towards the equator, we found that soil seed bank experienced a weakened latitudinal diversity gradient. Moreover, environmental covariates associated with species diversity of soil seed bank and aboveground vegetation are different, with mean temperature and temperature range as the major driver of species diversity variation in soil seed bank while elevation and temperature seasonality as the major drivers of aboveground vegetation species diversity. Our results indicate that the ecological patterns generated from aboveground biota do not necessarily reflect that of belowground diversity. Therefore, current strategies to conserve aboveground biodiversity may fail to protect belowground diversity.

4.1. Weak latitudinal diversity gradient in soil seed bank

Previous studies have found that plants have a strong latitudinal diversity gradient (Kreft and Jetz, 2007; Kerkhoff et al., 2014), and dozens of hypotheses have been proposed to explain the observed pattern (Usinowicz et al., 2017; Cerezer et al., 2022; Tietje et al., 2022; Delavaux et al., 2024). Our results also found strong signals of latitudinal diversity gradient in aboveground vegetation (Fig. 2). However, the strong latitudinal diversity gradients were weakened and even disappeared in soil seed bank (Fig. 2), implying that the distribution patterns of belowground diversity are different from that of aboveground vegetation (Cameron et al., 2019).

The observed weakening of the latitudinal diversity gradient in soil seed banks may be attributed to differences in the balance between seed accumulation and loss across latitudes. First, plants in the tropics do not produce seeds annually resulting in higher seed input variation between years (Kelly and Sork, 2002), whereas in aboveground vegetation, species diversity could be maintained at relatively high levels through diverse reproduction. For example, Parashorea chinensis, Garcinia cowa, and Pittosporopsis kerrii are dominant woody species of aboveground vegetation in the seasonal rainforest of China, and they do not store their seeds in soil seed bank (Data S1). Therefore, tropical soil seed bank has lower species richness compared to that of aboveground due to limiting seed input to soil seed bank. On the contrary, stable seed production of plants in temperature and boreal causes more seed accumulation in soil seed bank (Jakobsson and Eriksson, 2000; Moles et al., 2005). In boreal communities, dominant species in aboveground vegetation also dominated in soil seed bank, such as Picea schrenkiana of boreal forests in the Tianshan Mountains and Carex parvula of alpine grassland in the Qinghai-Xizang Plateau (Data S1). Therefore, species diversity stored in soil seed bank is comparable to that in aboveground vegetation at high latitudes.

In addition to slower seed accumulation, an acceleration in seed loss at lower latitudes represents another important factor contributing to the weakened latitudinal diversity gradient in soil seed bank. Seeds of tropic species are characterized as large size and recalcitrant seeds with high germination rates (Tweddle et al., 2003; Pritchard et al., 2004), which precludes them from forming persistent soil seed bank. Second, warmer and humid environments in the tropics tend to accelerate loss of seeds stored in the soil, mainly because of rapid decay of seeds and the loss of seed viability (Ooi, 2012). Our results also found higher temperatures in low-latitude areas could reduce species richness in soil seed bank (Fig. 4a), consistent with studies conducted in Ethiopia (Birhanu et al., 2022) and Brazil (Oliveira et al., 2019). Third, soil herbivores and pathogens can reduce seed persistence by damaging seed structure and integrity (Dalling et al., 2011). The mean herbivory rate in tropical regions was 3.3% higher than that of non-tropical regions in the Northern Hemisphere (Zhang et al., 2016). These factors synergistically decreased the probability of seed stored in the soil and contributed to a lower diversity of soil seed bank relative to aboveground vegetation in the low-latitude regions. In contrast, species living in high-latitude areas usually have small seed sizes with a lower probability of seed predation (Moles et al., 2005; Hargreaves et al., 2019), and suitable storage with cold and dry conditions, which can maintain seed viability and extend storage time with likely forming persistent soil seed bank (Wang et al., 2024), some of which can persist for decades and centuries (Telewski and Zeevaart, 2002). For example, orthodox seeds have 40–60 years of longevity at low humidity and temperature storage environments, longer than that under ambient conditions for 5–10 years (Solberg et al., 2020). Thus, species diversity in belowground seed bank is comparable to that in aboveground vegetation at high latitudes through long-term stable seed input and lower seed loss in soil seed bank.

4.2. Vegetation and ecosystem dependencies of weak latitudinal diversity gradient

Although the general pattern in weakened latitudinal diversity gradient of soil seed bank was found, the magnitude depends on vegetation types and life forms. Strong latitudinal diversity gradient of soil seed bank was also detected in woody species of forest ecosystems, despite being weaker than that of aboveground vegetation (Fig. 2b). On the contrary, there is a paucity of compelling evidence to substantiate the latitudinal diversity gradient hypothesis for herbaceous species of forest, grassland and wetland in soil seed bank (Fig. 2c–e), which has also been proved in global grasslands (Jabot and Pottier, 2017). The latitudinal diversity gradient of soil seed bank becomes less pronounced in communities dominated by herbaceous species, which is closely linked to the community assembly of soil seed bank. Many herbaceous species are widespread, instead of near the parent plant as woody species, resulting in community assembly being more regulated by seed dispersal rather than climatic factors (Mabry, 2004). That is also confirmed in our results that widespread species such as Bistorta vivipara, Persicaria hydropiper, and Setaria viridis, are relatively common herbaceous in soil seed bank across different communities from tropical forest to boreal wetland (Data S1). In addition, many herbaceous species produce more seeds than woody species (Jurado et al., 1991) and retain certain seeds by dormancy in soil seed bank to resist adverse circumstances and improve survival rate of the population (Faist et al., 2013; Plue and Cousins, 2013; Vandvik et al., 2016). That bet-hedging strategy facilitates herbaceous species to reserve in soil and dominates in the community composition of soil seed bank (Fenner and Thompson, 2005; Bossuyt and Honnay, 2008). Woody species generally have low seed production but high germination rates, which reduces the probability of persistence in soil seed bank. Smaller seed sizes of herbaceous species relatively woody are more favourable to avoiding predation and deeper distribution (Bakker et al., 1996), also facilitating longer seed persistence in soil seed bank. These factors synergistically increased herbaceous species diversity in soil seed bank, even in the high latitudes where plant diversity is relatively lower in aboveground vegetation.

We acknowledge that sampling areas for soil seed bank are typically smaller than those for aboveground vegetation, which may constrain direct comparisons of absolute species richness between these two biotic layers. While conducting surveys across identical areas at each site would be an ideal approach, it is often logistically prohibitive at large spatial scales. Nevertheless, because soil seed bank and aboveground vegetation follow distinct species–area relationship (SAR) curves, forcibly aligning their sampling scales could introduce artificial nonlinear biases. Therefore, comparing their relative response rates along latitudinal gradients remains a robust and scale-appropriate approach that is independent of baseline richness values.

4.3. Implications

Our study suggests that soil biodiversity cannot be represented by aboveground biodiversity, and protecting aboveground diversity may fail to sufficiently conserve biodiversity preserved in soil seed bank. While conservation efforts on aboveground vegetation such as the establishment of protected areas should be a priority (Le Saout et al., 2013), there is an urgent need to identify and protect hotspots of soil seed bank. Unfortunately, current recognition of hotspots of soil seed bank is rare because large-scale surveys involving intensive fieldwork are expensive and time consuming. In recent years, local studies on soil bank have been accumulating rapidly (DeMalach et al., 2021; Lei and Middleton, 2021; Gasperini et al., 2022). Therefore, we should improve accessibility to the vast information held in these publications for both domestic and international researchers, and it is vital to synthesize data from previous field surveys and build a sharing platform to help identify hotspots of belowground diversity. Meanwhile, it is critical to investigate less-explored areas with high diversity of soil seed bank, such as primary forests in tropical areas and alpine peatlands in the Tibetan Plateau (Data S1). Specifically, the unexpected high species diversity of soil seed bank in alpine peatlands provides a potential buffer against the negative effect of global climate change (Plue et al., 2021; Auffret et al., 2023), and these areas deserve urgent conservation attention. In addition, our study implies that current aboveground-focused conservation and management policies may ignore areas with low aboveground biodiversity but high belowground biodiversity such as logged forests and degraded ecosystems. For example, numerous studies have proved that logged forests maintained high resilience (Malhi et al., 2022), partly because of soil seed bank in these forests provides resources to restore ecosystem functions similar to that of primary forests (Kiss et al., 2018; Ma et al., 2019). As a result, species diversity of many abandoned lands has recovered rapidly (Crouzeilles et al., 2017), and natural regeneration can recoup 78% of their old-growth values within 20 years (Poorter et al., 2021). Seeds and tumps that remained after clearing probably accelerated the process. Therefore, degraded ecosystems with low aboveground biodiversity retain considerable conservation value in terms of belowground biodiversity, and they should not be overlooked (Ewers et al., 2024).

5. Conclusions

By using field surveys of plant diversity in soil seed bank and corresponding aboveground vegetation, we found that soil seed bank exhibits a weakened latitudinal diversity gradient compared with that of aboveground vegetation, confirming a mismatch between belowground biodiversity and aboveground biodiversity. As plant diversity in soil seed bank occurs primarily in habitats that may be different from aboveground vegetation, current aboveground-focused conservation and management policies may fail to protect biodiversity preserved in soil seed bank. Therefore, soil seed bank should be considered and included in future conservation planning, and there is an urgent need to identify and protect hotspots of soil seed bank.

Author contributions

JL conceived the idea. YZ collected the data. YZ, YL, SZ and TZ conducted the data analyses. All authors contributed to the writing.

Data availability statement

The data used in the analysis are publicly available in the Dataset S1.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This study was funded by the National Natural Science Foundation of China (32471730).

Footnotes

Peer review under the responsibility of Editorial Office of Plant Diversity.

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.pld.2026.05.003.

Appendix B. Supplementary data

The following are the Supplementary data to this article.

Multimedia component 1
mmc1.xlsx (452.4KB, xlsx)
Multimedia component 2
mmc2.docx (452.2KB, docx)

figs1.

figs1

Open in a new tab

figs2.

figs2

Open in a new tab

figs3.

figs3

Open in a new tab

References

  1. Auffret A.G., Ladouceur E., Haussmann N.S., et al. A global database of soil seed bank richness, density, and abundance. Ecology. 2024;105 doi: 10.1002/ecy.4438. [DOI] [PubMed] [Google Scholar]
  2. Auffret A.G., Vangansbeke P., De Frenne P., et al. More warm-adapted species in soil seed banks than in herb layer plant communities across Europe. J. Ecol. 2023;111:1009–1020. [Google Scholar]
  3. Bakker J.P., Poschlod P., Strykstra R.J., et al. Seed banks and seed dispersal: important topics in restoration ecology. Acta Bot. Neerl. 1996;45:461–490. [Google Scholar]
  4. Bates D., Mächler M., Bolker B., et al. Fitting linear mixed-effects models using lme 4. J. Stat. Softw. 2015;67:1–48. [Google Scholar]
  5. Birhanu L., Bekele T., Tesfaw B., et al. Soil seed bank composition and aboveground vegetation in dry Afromontane forest patches of Northwestern Ethiopia. Trees For. People. 2022;9 [Google Scholar]
  6. Bossuyt B., Honnay O. Can the seed bank be used for ecological restoration? An overview of seed bank characteristics in European communities. J. Veg. Sci. 2008;19:875–884. [Google Scholar]
  7. Cameron E.K., Martins I.S., Lavelle P., et al. Global mismatches in aboveground and belowground biodiversity. Conserv. Biol. 2019;5:1187–1192. doi: 10.1111/cobi.13311. [DOI] [PubMed] [Google Scholar]
  8. Cerezer F.O., Machac A.T., Rangel F., et al. Exceptions to the rule: relative roles of time, diversification rates and regional energy in shaping the inverse latitudinal diversity gradient. Global Ecol. Biogeogr. 2022;31:1794–1809. [Google Scholar]
  9. Chesson P. Mechanisms of maintenance of species diversity. Annu. Rev. Ecol. Systemat. 2000;31:343–366. [Google Scholar]
  10. Crouzeilles R., Ferreira M.S., Chazdon R.L., et al. Ecological restoration success is higher for natural regeneration than for active restoration in tropical forests. Sci. Adv. 2017;3 doi: 10.1126/sciadv.1701345. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dalling J.W., Davis A.S., Schutte B.J., et al. Seed survival in soil: interacting effects of predation, dormancy and the soil microbial community. J. Ecol. 2011;99:89–95. [Google Scholar]
  12. Davies T.J., Fritz S.A., Grenyer R., et al. Phylogenetic trees and the future of mammalian biodiversity. Proc. Natl. Acad. Sci. U.S.A. 2008;105:11556–11563. doi: 10.1073/pnas.0801917105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Delavaux C.S., Crowther T.W., Bever J.D., et al. Mutualisms weaken the latitudinal diversity gradient among oceanic islands. Nature. 2024;627:335–339. doi: 10.1038/s41586-024-07110-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. DeMalach N., Kigel J., Sternberg M. The soil seed bank can buffer long-term compositional changes in annual plant communities. J. Ecol. 2021;109:1275–1283. [Google Scholar]
  15. Economo E.P., Narula N., Friedman N.R., et al. Macroecology and macroevolution of the latitudinal diversity gradient in ants. Nat. Commun. 2018;9 doi: 10.1038/s41467-018-04218-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Ewers R.M., Orme C.D.L., Pearse W.D., et al. Thresholds for adding degraded tropical forest to the conservation estate. Nature. 2024;631:808–813. doi: 10.1038/s41586-024-07657-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Faist A.M., Ferrenberg S., Collinge S.K. Banking on the past: seed banks as a reservoir for rare and native species in restored vernal pools. AoB Plants. 2013;5 [Google Scholar]
  18. Fenner M., Thompson K. Cambridge University Press; Cambridge: 2005. The Ecology of Seeds. [Google Scholar]
  19. Fick S.E., Hijmans R.J. WorldClim 2: new 1-km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 2017;37:4302–4315. [Google Scholar]
  20. Fierer N., Jackson R.B. The diversity and biogeography of soil bacterial communities. Proc. Natl. Acad. Sci. U.S.A. 2006;103:626–631. doi: 10.1073/pnas.0507535103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Francis A.P., Currie D.J. A globally consistent richness-climate relationship for angiosperms. Am. Nat. 2003;161:523–536. doi: 10.1086/368223. [DOI] [PubMed] [Google Scholar]
  22. Gasperini C., Bollmann K., Brunet J., et al. Soil seed bank responses to edge effects in temperate European forests. Global Ecol. Biogeogr. 2022;31:1877–1893. doi: 10.1111/geb.13568. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Gasperini C., Carrari E., Govaert S., et al. Edge effects on the realised soil seed bank along microclimatic gradients in temperate European forests. Sci. Total Environ. 2021;798 doi: 10.1016/j.scitotenv.2021.149373. [DOI] [PubMed] [Google Scholar]
  24. Gioria M., Pyšek P. The legacy of plant invasions: changes in the soil seed Bank of invaded plant communities. Bioscience. 2016;66:40–53. [Google Scholar]
  25. Gioria M., Pyšek P., Baskin C.C., et al. Phylogenetic relatedness mediates persistence and density of soil seed banks. J. Ecol. 2020;108:2121–2131. [Google Scholar]
  26. Hargreaves A.L., Suárez E., Mehltreter K., et al. Seed predation increases from the Arctic to the Equator and from high to low elevations. Sci. Adv. 2019;5 doi: 10.1126/sciadv.aau4403. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hengl T., de Jesus J.M., Heuvelink G.B.M., et al. SoilGrids250m: global gridded soil information based on machine learning. PLoS One. 2017;12 doi: 10.1371/journal.pone.0169748. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Hillebrand H. On the generality of the latitudinal diversity gradient. Am. Nat. 2004;163:192–211. doi: 10.1086/381004. [DOI] [PubMed] [Google Scholar]
  29. Hopfensperger K.N. A review of similarity between seed bank and standing vegetation across ecosystems. Oikos. 2007;116:1438–1448. [Google Scholar]
  30. Hoyle G.L., Venn S.E., Steadman K.J., et al. Soil warming increases plant species richness but decreases germination from the alpine soil seed bank. Glob. Change Biol. 2013;19:1549–1561. doi: 10.1111/gcb.12135. [DOI] [PubMed] [Google Scholar]
  31. Jablonski D., Roy K., Valentine J.W. Out of the tropics: evolutionary dynamics of the latitudinal diversity gradient. Science. 2006;314:102–106. doi: 10.1126/science.1130880. [DOI] [PubMed] [Google Scholar]
  32. Jabot F., Pottier J. Macroecology of seed banks: the role of biogeography, environmental stochasticity and sampling. Global Ecol. Biogeogr. 2017;26:1247–1257. [Google Scholar]
  33. Jakobsson A., Eriksson O. A comparative study of seed number, seed size, seedling size and recruitment in grassland plants. Oikos. 2000;88:494–502. [Google Scholar]
  34. Jetz W., Thomas G.H., Joy J.B., et al. The global diversity of birds in space and time. Nature. 2012;491:444–448. doi: 10.1038/nature11631. [DOI] [PubMed] [Google Scholar]
  35. Jurado E., Westoby M., Nelson D. Diaspore weight, dispersal, growth form and perenniality of central Australian plants. J. Ecol. 1991;79:811–828. [Google Scholar]
  36. Kelly D., Sork V.L. Mast seeding in perennial plants: why, how, where? Annu. Rev. Ecol. Evol. Syst. 2002;33:427–447. [Google Scholar]
  37. Kerkhoff A.J., Moriarty P.E., Weiser M.D. The latitudinal species richness gradient in New World woody angiosperms is consistent with the tropical conservatism hypothesis. Proc. Natl. Acad. Sci. U.S.A. 2014;111:8125–8130. doi: 10.1073/pnas.1308932111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Kiss R., Deak B., Torok P., et al. Grassland seed bank and community resilience in a changing climate. Restor. Ecol. 2018;26:S141–S150. [Google Scholar]
  39. Kreft H., Jetz W. Global patterns and determinants of vascular plant diversity. Proc. Natl. Acad. Sci. U.S.A. 2007;104:5925–5930. doi: 10.1073/pnas.0608361104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Le Saout S., Hoffmann M., Shi Y., et al. Protected areas and effective biodiversity conservation. Science. 2013;342:803–805. doi: 10.1126/science.1239268. [DOI] [PubMed] [Google Scholar]
  41. Lei T., Middleton B. Germination potential of baldcypress (Taxodium distichum) swamp soil seed bank along geographical gradients. Sci. Total Environ. 2021;759 doi: 10.1016/j.scitotenv.2020.143484. [DOI] [PubMed] [Google Scholar]
  42. Ma M., Baskin C.C., Li W., et al. Seed banks trigger ecological resilience in subalpine meadows abandoned after arable farming on the Tibetan Plateau. Ecol. Appl. 2019;29 doi: 10.1002/eap.1959. [DOI] [PubMed] [Google Scholar]
  43. Mabry C.M. The number and size of seeds in common versus restricted woodland herbaceous species in central Iowa, USA. Oikos. 2004;107:497–504. [Google Scholar]
  44. Malhi Y., Riutta T., Wearn O.R., et al. Logged tropical forests have amplified and diverse ecosystem energetics. Nature. 2022;612:707–713. doi: 10.1038/s41586-022-05523-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Mittelbach G.G., Schemske D.W., Cornell H.V., et al. Evolution and the latitudinal diversity gradient: speciation, extinction and biogeography. Ecol. Lett. 2007;10:315–331. doi: 10.1111/j.1461-0248.2007.01020.x. [DOI] [PubMed] [Google Scholar]
  46. Moles A.T., Ackerly D.D., Webb C.O., et al. A brief history of seed size. Science. 2005;307:576–580. doi: 10.1126/science.1104863. [DOI] [PubMed] [Google Scholar]
  47. Oksanen J., Simpson G.L., Blanchet F.G., et al. Vegan: community ecology package. Manual. 2022 [Google Scholar]
  48. Oliveira P.A.N., Fraga dos Santos J.M., Lima Araújo E., et al. Natural regeneration of the vegetation of an abandoned agricultural field in a semi-arid region: a focus on seed bank and above-ground vegetation. Braz. J. Bot. 2019;42:43–52. [Google Scholar]
  49. Ooi M.K.J. Seed bank persistence and climate change. Seed Sci. Res. 2012;22:S53–S60. [Google Scholar]
  50. Phillips H.R.P., Guerra C.A., Bartz M.L.C., et al. Global distribution of earthworm diversity. Science. 2019;366:480–485. doi: 10.1126/science.aax4851. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Pichon N.A., Kaarlejärvi E., Eskelinen A. Seed limitation interacts with biotic and abiotic factors to constrain novel species' impact on community biomass and richness. Ecol. Lett. 2023;26:908–918. doi: 10.1111/ele.14219. [DOI] [PubMed] [Google Scholar]
  52. Plue J., Cousins S.A.O. Temporal dispersal in fragmented landscapes. Biol. Conserv. 2013;160:250–262. [Google Scholar]
  53. Plue J., De Frenne P., Acharya K., et al. Buffering effects of soil seed banks on plant community composition in response to land use and climate. Global Ecol. Biogeogr. 2021;30:128–139. [Google Scholar]
  54. Poorter L., Craven D., Jakovac C.C., et al. Multidimensional tropical forest recovery. Science. 2021;374:1370–1376. doi: 10.1126/science.abh3629. [DOI] [PubMed] [Google Scholar]
  55. Post W.M., Pastor J., Zinke P.J., et al. Global patterns of soil nitrogen storage. Nature. 1985;317:613–616. [Google Scholar]
  56. Pritchard H.W., Daws M.I., Fletcher B.J., et al. Ecological correlates of seed desiccation tolerance in tropical African dryland trees. Am. J. Bot. 2004;91:863–870. doi: 10.3732/ajb.91.6.863. [DOI] [PubMed] [Google Scholar]
  57. Pyron R.A., Wiens J.J. Large-scale phylogenetic analyses reveal the causes of high tropical amphibian diversity. P. Roy. Soc. B-Biol. Sci. 2013;280 doi: 10.1098/rspb.2013.1622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. R Core Team . R Foundation for Statistical Computing; Vienna, Austria: 2025. R: a Language and Environment for Statistical Computing. [Google Scholar]
  59. Simpson R.L., Leck M.A., Parker V.T. Pages 3–8 Ecology of Soil Seed Banks. Elsevier; 1989. Seed banks: general concepts and methodological issues. [Google Scholar]
  60. Solberg S.Ø., Yndgaard F., Andreasen C., et al. Long-Term storage and longevity of Orthodox seeds: a systematic review. Front. Plant Sci. 2020;11 doi: 10.3389/fpls.2020.01007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Taylor A., Weigelt P., Denelle P., et al. The contribution of plant life and growth forms to global gradients of vascular plant diversity. New Phytol. 2023;240:1548–1560. doi: 10.1111/nph.19011. [DOI] [PubMed] [Google Scholar]
  62. Tedersoo L., Bahram M., Põlme S., et al. Global diversity and geography of soil fungi. Science. 2014;346 doi: 10.1126/science.1256688. [DOI] [PubMed] [Google Scholar]
  63. Telewski F.W., Zeevaart J.A.D. The 120-yr period for Dr. Beal’s seed viability experiment. Am. J. Bot. 2002;89:1285–1288. doi: 10.3732/ajb.89.8.1285. [DOI] [PubMed] [Google Scholar]
  64. Thompson K., Grime J.P. Seasonal variation in the seed banks of herbaceous species in ten contrasting habitats. J. Ecol. 1979;67:893–921. [Google Scholar]
  65. Tietje M., Antonelli A., Baker W.J., et al. Global variation in diversification rate and species richness are unlinked in plants. Proc. Natl. Acad. Sci. U.S.A. 2022;119 doi: 10.1073/pnas.2120662119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Tweddle J., Dickie J., Baskin C., et al. Ecological aspects of seed desiccation sensitivity. J. Ecol. 2003;91:294–304. [Google Scholar]
  67. Usinowicz J., Chang-Yang C.H., Chen Y.Y., et al. Temporal coexistence mechanisms contribute to the latitudinal gradient in forest diversity. Nature. 2017;550:105–108. doi: 10.1038/nature24038. [DOI] [PubMed] [Google Scholar]
  68. Vandvik V., Klanderud K., Meineri E., et al. Seed banks are biodiversity reservoirs: species–area relationships above versus below ground. Oikos. 2016;125:218–228. [Google Scholar]
  69. Wang X., Ge W., Zhang M., et al. Large and non-spherical seeds are less likely to form a persistent soil seed bank. P. Roy. Soc. B-Biol. Sci. 2024;291 doi: 10.1098/rspb.2023.2764. [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Warr S.J., Thompson K., Kent M. Seed banks as a neglected area of biogeographic research: a review of literature and sampling techniques. Prog. Phys. Geogr. 1993;17:329–347. [Google Scholar]
  71. Yang X., Baskin C.C., Baskin J.M., et al. Global patterns of potential future plant diversity hidden in soil seed banks. Nat. Commun. 2021;12 doi: 10.1038/s41467-021-27379-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Yao P., Lu H., Shi J., et al. A long term global daily soil moisture dataset derived from AMSR-E and AMSR2 (2002–2019) Sci. Data. 2021;8 doi: 10.1038/s41597-021-00925-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Yu S., Jiang G. The research development of soil seed bank and several hot topics. Chin. J. Plant Ecol. 2003;27 [Google Scholar]
  74. Zhang S., Zhang Y., Ma K. Latitudinal variation in herbivory: hemispheric asymmetries and the role of climatic drivers. J. Ecol. 2016;104:1089–1095. [Google Scholar]
  75. Zou C., Martini F., Xia S.W., et al. Elevation and micro environmental conditions directly and indirectly influence forests' soil seed bank communities. Glob. Ecol. Conserv. 2021;26 [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Multimedia component 1
mmc1.xlsx (452.4KB, xlsx)
Multimedia component 2
mmc2.docx (452.2KB, docx)

Data Availability Statement

The data used in the analysis are publicly available in the Dataset S1.

Articles from Plant Diversity are provided here courtesy of Kunming Institute of Botany, Chinese Academy of Sciences

RESOURCES