The life cycle of plants is intricately linked to the climate. Climate change, therefore, may threaten life history strategies that have arisen over evolutionary timescales. Understanding how plant communities will respond to anthropogenic climate change has become an integral part of ecological research (Warren et al., 2013). Seeds are particularly useful in this endeavor as they enable us to evaluate changes in vegetation at the community level (Walck et al., 2011). Many species have evolved dormancy mechanisms that regulate when germination takes place – an adaptation that enables plants to inhabit environments with volatile climates (Finch-Savage and Leubner-Metzger, 2006; Willis et al., 2014). However, the warming of Earth’s climate could cause seeds to germinate at unfavorable times for seedling establishment, potentially leading to local extinctions. Unfortunately, our understanding of how a warmer climate will affect the germination ecology of species with dormant seeds is hampered by a lack of long-term studies.

Seed dormancy can be advantageous in the tropics, where synchronizing germination with the rainy season ensures seedlings are well established before the onset of the harsh dry season (Garwood, 1983; Khurana and Singh, 2001). Tropical plants often avoid dry season germination by either producing dormant seeds or dispersing their seeds in the wet season (Sautu et al., 2007; Salazar et al., 2011; Ramos et al., 2017; Escobar et al., 2018). Thus a trade-off exists, whereby species that disperse seeds during the wet season tend to be non-dormant, germinating immediately following dispersal, while species that disperse their seeds in the dry season tend to be dormant, germinating only when conditions become favorable again in the wet season (Salazar et al., 2011; de Souza et al., 2020). Of the latter, various strategies are used to achieve seed dormancy.

Physical dormancy (PY), caused by a seed/fruit coat that is impermeable to water in some genera of only 18 contemporary angiosperm families is found in 25% of all flowering plants (Baskin and Baskin, 2004; Hudson et al., 2015). It is the second most common class of dormancy after physiological dormancy (PD) and is achieved late in seed development when seeds dry below threshold moisture content (Barrett-Lennard and Gladstones, 1964; Geneve, 2009; Jaganathan, 2016). Impermeable seeds persist in the soil until specific environmental conditions stimulate dormancy-break by opening structures on the seed coat known as “water-gaps” through which water enters and initiates germination (Van Assche et al., 2003; Baskin and Baskin, 2004). PY is typically broken by temperature fluctuations prior to the growing season (Cook et al., 2008; Jayasuriya et al., 2009; Rodrigues-Junior et al., 2018), however, it can also be broken by erratic events such as fire (Jaganathan, 2015) or passage through an animal gut (Jaganathan et al., 2016; Milotić and Hoffmann, 2016). PY can benefit plant fitness not only by synchronizing germination with favorable growing conditions (Baskin et al., 2000), but also by establishing a long-term soil seed bank from which only a portion germinate each year, spreading germination risk across years (Baskin and Baskin, 2014), and protecting seeds against pathogens, soil microbes (Dalling et al., 2011), and seed predators (Paulsen et al., 2013).

While global mean annual temperature is predicted certainly to rise, e.g., 4°C by 2100 (Warren et al., 2013), consequences this change will have for species that employ seed dormancy remain poorly understood. Much of our knowledge of how a warmer climate affects seed dynamics comes from studies conducted in alpine ecosystems (Mondoni et al., 2012; Hoyle et al., 2013); presumably, because they are disproportionately vulnerable to changes in temperature. Tropical ecosystems, on the other hand, have received comparatively less attention (Perez et al., 2016; Stroud and Feeley, 2017). This is partly due to disagreement over the extent to which temperature will rise in the tropics, with predictions ranging widely from 0.26 to 5°C (Corlett, 2012). Regardless of extent, consensus is emerging that warming in the tropics could have serious consequences for plant life (Feeley and Silman, 2010). Soil temperature acts as a bottleneck that controls, not only when seeds break dormancy, but also the proportion of seeds germinating each year. Climate warming may therefore alter the soil environment in which the seeds are present following dispersal. This may lead to increased mortality via the excess evaporation of soil water (Ooi, 2012) or the stimulation of dormancy-break (and consequently germination) at unfavorable times, such that seedlings die before establishing (Walck et al., 2011).

Seeds banks are vital to species that inhabit ecosystems with unpredictable climates or short growing seasons (Long et al., 2015; Jaganathan et al., 2019). They can enable species to persist locally, even when seemingly all plant life has been destroyed (Jaganathan et al., 2015). However, while it is becoming increasingly apparent that warming temperatures may significantly affect the regeneration ecology of species with seed dormancy (Orru et al., 2012; Newton et al., 2020), responses appear to be largely species-specific (Seal et al., 2017). Further, despite efforts to understand such effects for species that employ PY (Hudson et al., 2015; Cochrane, 2017), to date, no such study has been done in the tropics. To close this gap, we performed a long-term warming experiment to investigate how a warmer climate will affect the germination ecology of five tropical species with PY. Specifically, we asked:

The moisture content of the seeds of each species ranged from 5.86% for A. lebbeck to 9.06% for A. chundra (Table 1). Impermeable seeds at the time of collection varied from 4% for A. lebbeck to 31% for B. tomentosa (Table 1). Permeable seeds of most of the species swelled and germinated within 10 days after being placed on a moist substrate, although a few seeds of A. lebbeck did not absorb water until after 10 days. Soil temperatures peaked during the summer months (May–July) and were lowest during the winter months (Dec–Feb). Mean soil temperature ranged from 63 to 19°C (Figure 1).

The implications of climate change for the regeneration of tropical species are poorly understood, largely because of disagreement about the degree to which the tropics are expected to warm. However, there has been renewed interest in this area (Clark, 2004; Corlett, 2012; Mau et al., 2018). To understand how climate warming might affect the regeneration of tropical legumes with PY, we documented the germination dynamics of five species under two future climate scenarios using a long-term warming experiment. PY was confirmed in all five species using imbibition tests (Table 1). While PY has been confirmed in four of these species previously (Jayasuriya et al., 2013), our study appears to be the first to document PY in A. chundra.

Many tropical species with PY disperse their seeds at the transition of the wet to dry seasons (Vázquez-Yanes and Orozco-Segovia, 1993; Khurana and Singh, 2001). In the tropics, dormancy is broken by seasonal temperature fluctuations and the extreme soil temperatures (up to 60°C) experienced in summer. In our experiment, a proportion of seeds of all five species incubated under current climatic conditions germinated after 6 months (Figure 2). Data-loggers confirmed that soil temperature exceeded 60°C before the 6-month retrieval, suggesting that current summer temperatures break dormancy in a proportion of seeds (Figure 1). Similar patterns of dormancy break have been observed in Dodonea viscosa (Jaganathan and Liu, 2014), D. hackettiana (Cook et al., 2008), Delonix regia (Jaganathan et al., 2017), and Adenanthera pavonina (Jaganathan et al., 2018), where seeds artificially buried at a depth of 3 cm–5 cm exhibited 20–40% germination after the first summer.

Physical dormant seeds can persist in the soil for extensive periods of time (Baskin and Baskin, 2014). Indeed, most seeds in the five species we examined remained dormant after 2 years under current climatic conditions. However, the predicted warming in the tropics of 2–5°C by 2100 (Betts et al., 2011; Joshi et al., 2011; Corlett, 2012) may reduce the ability of physically dormant seeds to persist in the soil for so long. Results from our mild warming treatment (+2°C) were idiosyncratic among species. Bauhinia tomentosa and A. chundra exhibited little to no difference in germination percentages relative to control across the entire 2-year period. Seeds of A. precatorius also were not affected by 2°C of warming for much of the experiment, although increased germination was observed in month 24. Interestingly, C. fistula seeds exhibited markedly higher germination rates when incubated under both warming treatments, indicating that this species is particularly sensitive to changes in temperature (Figure 2; Table 2). Such idiosyncratic responses to climate warming might be explained by differences in the conditions required to break dormancy, such as seed moisture content (Jaganathan, 2016), seed coat thickness (Russi et al., 1992; Venier et al., 2012), and seed size (Halloran and Collins, 1974; Rodrigues-Junior et al., 2018), all of which could act independently or in tandem. Why C. fistula seeds were particularly sensitive to warming remains unknown and warrants further study.

In contrast to our mild warming treatment, seeds of all species subjected to our extreme warming treatment (+4°C) exhibited considerably higher germination percentages relative to control seeds (Figure 2). Cochrane (2017) observed similar results in the seeds four Acacia species from Australia that were incubated at 20/65°C for 112 days. Our extreme warming treatment generated very high soil temperatures during both summers (65 and 67°C, respectively), mirroring those of Cochrane (2017). Collectively, these results suggest that temperatures exceeding 65°C deplete the persistence of physically dormant seeds in the soil. Soil temperatures of 65°C are relatively common in tropical soils (Cook et al., 2008; Jaganathan and Liu, 2014), however, temperatures may reach 70°C in some open canopy areas (although direct soil temperature measurements are lacking). Thus, a threshold temperature appears to restrict the conditions under which PY can effectively operate. How exactly the warming of tropical soils beyond this threshold will affect the distribution of these species (and others) is not known.

The conditions required to break PY in the field have received some attention (Quinlivan, 1961, 1968; Taylor, 2005). Vázquez-Yanes and Orozco-Segovia (1982) demonstrated that impermeable seeds of Heliocarpus donnell-smithii matured in Mexico become permeable only when soil temperature exceeds 30°C. Further, more seeds become permeable to water in open canopy sites than under closed canopy, mainly because diurnal temperatures fluctuations in open canopy sites are more extreme. McDonald (2000) similarly found that subjecting seeds to varying temperature regimes of 57/23 and 70/23°C breaks PY, and that this effect was especially pronounced at high temperatures. In the future, a warmer climate is likely to bring more extreme diurnal fluctuations at higher temperatures, potentially rendering PY ineffective as a means of long-term persistence for some species in the soil.

Jaganathan (2016) previously classified PY into two levels: shallow and absolute. For seeds with shallow PY, their moisture content approaches the range at which impermeability to water is achieved (e.g., 8–10%). Seeds with a much lower moisture content of 5–8%, on the other hand, comprise the absolute PY category. We observed considerable variation in moisture content both within and among species (Table 1). Idiosyncratic responses to our warming treatments may, at least in part, be explained by this variation in moisture content. Alternatively, they may represent species-specific adaptations to certain temperature thresholds.

Taylor (1981) proposed that PY loss occurs in two-steps: (1) preconditioning, where seeds are conditioned and made sensitive, but seeds remain impermeable to water; and (2) actual dormancy-break, during which the specialized structures present in seeds (e.g., the lens in Fabaceae) opens, through which water enters and hydrates the embryo. Similar mechanisms of dormancy break have been observed in other families, such as Convolvulaceae (Jayasuriya et al., 2009) and Geraniaceae (Gama-Arachchige et al., 2012). These mechanisms are directly controlled by seasonal changes in temperature. Accordingly, we hypothesize that both stages of dormancy loss will be altered under a warmer climate. Consequently, the bet-hedging mechanism (Philippi and Seger, 1989; Ooi et al., 2009) to spread the risk of germination over time is particularly expected to be altered. One possibility is that warmer temperatures and reduced relative humidity in the future may dry seeds out more, rendering physical dormancy harder to break, and thereby compensating for higher summer temperatures (Bolingue et al., 2010). The effects of climate warming may also occur at the stage of seed maturation on the mother plant. Indeed, the effects of maternal environment on germination dynamics of PY species is gaining attention (Liyanage and Ooi, 2015; Jaganathan and Biddick, 2020), and the ultimate effect of climate warming on species with PY may result from an interplay of processes occurring both before and after dispersal. However, exactly how each stage of dormancy-break will be affected in the future requires further study.

In conclusion, it appears that an increase in temperature of 4°C will significantly increase dormancy loss and alter germination timing of species that employ PY in the tropics, potentially affecting plant establishment and community composition. A milder increase in temperature of 2°C is likely to affect fewer species than 4°C. How severely the germination ecology of species will be affected by climate warming is difficult to ascertain as responses are highly idiosyncratic.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

GJ conceived the idea, performed the experiment, and wrote the manuscript. MB analyzed the statistics and wrote the manuscript. All authors contributed to the article and approved the submitted version.

The financial support by National Science Foundation China (NSFC) with grant number 31750110474 is gratefully acknowledged. Funders play no role in the design of the study and publication of the results.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.