Abstract
The increase in severity of droughts associated with greater mortality and reduced vegetation growth is one of the main threats to tropical forests. Drought resilience of tropical forests is affected by multiple biotic and abiotic factors varying at different scales. Identifying those factors can help understanding the resilience to ongoing and future climate change. Altitude leads to high climate variation and to different forest formations, principally moist or dry tropical forests with contrasted vegetation structure. Each tropical forest can show distinct responses to droughts. Locally, topography is also a key factor controlling biotic and abiotic factors related to drought resilience in each forest type. Both dry tropical forests and ridges (steeper and drier habitats) are more sensitive to droughts than moist tropical forest and valleys (flatter and wetter habitats). The most important biotic factors are leaf economic and hydraulic plant traits, and vegetation structure. The most important abiotic factors are soil nutrients, water availability and microclimate. Here we show that topography has key roles controlling biotic and abiotic factors in each forest type. Our synthesis highlights that gradients of altitude and topography are essential to understand tropical forest’s resilience to future drought events. We described important factors related to drought resilience, however many important knowledge gaps remain. Filling those gaps will help improve future practices and studies about mitigation capacity, conservation, and restoration of tropical ecosystems.
Keywords: Climate change, El Niño, mortality, growth, recovery, resistance
Introduction
Drought is a natural phenomenon characterized by abnormally low precipitations and high temperatures, which strongly affect plants at different scales, from individuals, to populations and communities, and thus ecosystems worldwide (Dai 2011, Fauset et al. 2012). Drought events are predicted to modify species composition and abundance, as well as ecosystem functioning and dynamics (Boeck et al. 2017), especially in tropical forests (Reichstein et al. 2013) due to high exposure to El Niño events (IPCC, 2018). Nevertheless, resilience to droughts in tropical forests remains poorly understood (Xu et al. 2013, Meir et al. 2015). Identifying biotic and abiotic factors related to mortality and low growth induced by droughts would allow us to better understand and predict forest resilience and dynamics.
The interest in studying drought effects in tropical forests has increased more than four times during the last 20 years (Fig. 1). This is likely because tropical forests are among the most vulnerable ecosystems to climate change (Bellard et al. 2014). Indeed, tropical forests were strongly affected by the two most severe droughts ever recorded, in 2015/2016 (Otto et al. 2015, Kogan and Guo 2017) and 1998/1999 (Slik 2004). The former and most recent one led to a high net carbon loss, even higher than the 2010 drought in the Amazonian forest (Liu et al. 2017). Tropical forests provide key ecosystem services – such as water supply, carbon sequestration, pollination and climate control – to urban and rural areas (Joly et al. 2014, Silva et al. 2021). Although our understanding of drought effects worldwide is growing, the impacts of droughts in tropical forests are still poorly understood, which limits our ability to model forest responses to future climate scenarios.
Repeated occurrences of strong droughts can select for drought-tolerant species due to higher mortality of vulnerable species (Aguirre-Gutiérrez et al. 2020). Thus, it has been predicted that trees of tropical forests will become smaller with denser wood (Phillips et al. 2010), due to high mortality rates of canopy trees and shade-tolerant species (Fauset et al. 2012). There is a field of evidence that tropical forests can change to a drier ecosystem due to the interactions between recurrent fires and climate change. This process drives communities through a savannization (Sansevero et al. 2020) and/or a secondarization process (Barlow and Peres 2008), thus decreasing carbon sequestration and enhancing climate change effects, because dry ecosystems sequester less CO2 than moist forests (Taylor et al. 2017). Therefore, repeated drought events associated with high deforestation and recovery can drive tropical forests into a different type of forest following the future climate scenarios. This process can lead to changes in forest composition, vegetation structure, traits and ecosystem services (Aguirre-Gutiérrez et al. 2020). Predictions of which plant species are most vulnerable to drought effects, and where they occur, are needed to model consequences of climate change and to determine the best mitigation strategies.
Climate changes impact forests at multiple spatial scales, and this is the interaction of processes across scales that may determine forest resilience (Reyer et al. 2015). For instance, drought effects are influenced by gradients varying at global, regional and local scales (O’Brien et al. 2017). Thus, looking at multiple spatial scales may offer the best way to predict forest resilience because plant mortality and reduced growth can occur locally with potentially negative effects, but they may smooth out at larger spatial scales (Reyer et al. 2015). The aim of this study was to synthesize biotic and abiotic factors, mostly controlled by topography, that can be used to predict drought-induced mortality in tropical forests. We expect that this review will help identify knowledge gaps in drought resilience research and provide a framework for future studies.
What is drought resilience and how do we quantify it?
Resilience is the capacity of a biome, an ecosystem or a species to both resist and recover from a disturbance (Oliver et al. 2015). Resistance is the ability of a system to maintain its properties and functions during a disturbance (Mariotte et al. 2013) whereas recovery is the ability of a system to return to initial conditions (Verbesselt et al. 2016). Droughts are associated with vegetation mortality and reduced growth (Adams et al. 2009, Verbesselt et al. 2016, Greenwood et al. 2017, Meir et al. 2018). Thus, quantifying mortality and growth to estimate drought resilience is an important approach to measure vulnerability and drought resilience of plant species (Redmond et al. 2018). Evidences suggest a growth-survival trade-off, showing for example, that survival rates are negatively correlated to growth rates (Wright et al. 2010, O’Brien et al. 2017). Another pattern can be found when recovery rates are measured after a normal condition, for example, after a La Niña event. In this case, a resistance-recovery trade-off is expected, with sites or species showing greater resistance rate during drought but lower recovery rate after a rainy period (Gazol et al. 2017). Furthermore, during the recovery period, new individuals can be recruited, mainly because of the available niches resulting from large tree mortality (Redmond et al. 2018). Tree resistance and recovery to droughts will likely determine the long-term trajectory of tropical forests. Therefore, resilience is an important component to determine community and ecosystem sensitivity to disturbances and to predict shifts in forest carbon stocks caused by climate changes (Sánchez-Salguero et al. 2018).
Drought timing and severity, as well as forest type, are important factors to take into account when estimating the resistance and necessary time for the recovery. For instance, in 2017, savannas and grasslands had already recovered from the 2015/2016 drought but tropical forests did not because tropical trees have slower growth and recruitment rates (Wigneron et al. 2020). Drought legacy, i.e.,the time of recovery after a drought, can take longer according to the drought magnitude (Huang et al. 2018, Kannenberg et al. 2020). For instance, tropical forests in Costa Rica recovered from the 1997 drought within two years (Silva et al. 2013) and the Puerto Rican forests recovered from the 2015 drought within 1 year (Schwartz et al. 2019, 2020). Quantifying the local magnitude of droughts can be difficult due to spatial variation in precipitation and temperature. An El Niño event does not mean the same dry condition because some forests can be more exposed to such conditions depending on the latitude and the topography.
Drought resilience in different types of forests
Tropical forests include dry and moist forests (Olson et al. 2001). Dry tropical forests (also called seasonally dry tropical forest) range from tall forest on moister regions to shrublands on the driest regions and the vegetation is mostly deciduous during the dry season (Pennington et al. 2009). Moist tropical forests are characterized by the presence of evergreen plant species in high-altitude moister areas of the tropics but also by the presence of deciduous species in areas with seasonal climatic regime (WCMC 1992). Therefore, this classification (dryvs. moist tropical forest) is mainly due to different climatic conditions, which are driven by altitude, latitude and longitude (Fig. 2).
Distinct patterns of drought resilience likely exist among different forest types because each type of forest will impact differently soil nutrients and water availability, vegetation structure and functional traits (Fig. 2). For instance, African tropical forests are more resistant to climatic extremes than Amazonian and Asian forests due to differences in their vegetation structure (Bennett et al. 2021). Furthermore, El Niño impacts vary latitudinally in the Atlantic Forest, with tropical areas being more impacted by El Niño events than subtropical areas (Rodrigues et al. 2011). Eastern and southern regions of the Amazon Forest experienced stronger drought impacts than the north-western region (Van Schaik et al. 2018).
Tropical dry forests are less resistant to droughts (Allen et al. 2017). Furthermore, tropical forest communities in West Africa that normally experience higher seasonal water deficit, and that became drier through time, tended to become more homogeneous in functional, taxonomic and phylogenetic diversity (Aguirre-Gutiérrez et al. 2020). Some areas of tropical forest are secondary forest patches recovering from land use, timber extraction or natural disturbances (Rüger et al. 2020). Pioneer species occurring in those secondary forests present high mortality during drought periods (Rocha et al. 2020), but high recovery rates after a drought event (Poorter et al. 2016, Gazol et al. 2017). Those findings highlight that each forest type of tropical forest can show different patterns of growth and mortality and this regulates how much carbon each forest type will store under the future climate.
Drought resilience across topographic gradients
Locally, topography is also a source of environmental heterogeneity within each forest type (Jucker, Bongalov, et al. 2018, Nettesheim et al. 2018). Abiotic factors strongly vary across topographic gradients, such as water availability, soil nutrients and microclimate (Fig. 2, Fig. 3, Fyllas et al., 2017). Topographic variation of tropical forests can be distinguished into two main habitats: valleys (flatter and wetter habitats) and ridges (steeper and drier habitats). Biotic factors, such as species traits and vegetation structure, are also driven by topography (Fig. 2, Fig. 3, Jucker et al., 2018a; Rodrigues et al., 2019). Topography is thus an important factor to take into account to understand drought impacts in plant communities due to the role of those factors in survival and growth of plant species (O’Brien et al. 2017, Hollunder et al. 2021).
Species occurring in different topographical habitats can show distinct patterns of mortality and growth rate. Studies in tropical forests have shown that woody plant species occurring in valleys can be more resistant than the ones occurring in ridges (Table 1). Valleys can act as refuges, providing nutrients and water during droughts (Costa et al. 2020, Hollunder et al. 2021). Furthermore, shade availability in valleys can reduce drought impacts and improve the performance of plant species during drought periods (Holmgren et al. 2012). Species occurring in dry habitats are living under their microclimate limits and under water stress, and a small change in precipitation and temperature can be physiologically stressful (Allen et al. 2017, Aguirre-Gutiérrez et al. 2020, Cartwright et al. 2020). In turn, species from dry habitats can show high recovery rates during a post-drought period (Schwartz et al. 2019), following the resistance-recovery trade-off. However, it has been suggested that dry habitats can also show strong drought legacies (Anderegg et al. 2015, Allen et al. 2017).
Other studies have shown that valleys can be less resistant than the ones occurring in ridges (Table 1). Species from valleys do not have traits related to drought tolerance, while species occurring in ridges have traits related to survival during periods of water-stress (Allen et al. 2017, Gessler et al. 2017). Such distinct evidences indicate that there is no general pattern of drought impacts for all tropical forests, though most of the studies found stronger resistance in wetter habitats. Biotic and abiotic factors controlled by topographic gradients are also strong drivers of tree growth and mortality, as well as of forest resilience, and can thus contribute to help elucidating such patterns (Fig. 2).
Abiotic factors