Introduction

Temperatures have increased substantially in the Tropics in recent decades, by as much as 0.5 °C per decade in some regions (Jiménez-Muñoz, Sobrino, Mattar, & Malhi, 2013). In many areas, these increases in mean temperatures have been accompanied by increasingly frequent heatwaves and extreme temperature days (Coumou & Robinson, 2013). Although many tropical forest tree taxa have been exposed to warmer temperatures in the past – e.g. during the Pliocene-Eocene thermal maximum, the rapid warming rates currently experienced are unprecedented (Dick, Lewis, Maslin, & Bermingham, 2013). Forests in the southern Amazon experience the highest temperatures in Amazonia, with monthly maximum air temperatures during dry periods frequently reaching >40°C. They are also amongst the planets most rapidly-warming tropical forests (Gloor et al., 2018; Jiménez-Muñoz et al., 2013) making them a natural laboratory for studying the effects of global warming on tropical forests. However, despite their increasing exposure to heat stress, there is currently no empirical data on the sensitivity of southern Amazonian forests to high temperatures.
Photosynthesis, which underpins the substantial productivity and biomass storage of tropical forests, is heavily temperature-dependent (Berry & Björkman, 1980). Most investigations to date have focused on the temperature sensitivity of net CO2 exchange, at the leaf (e.g., Doughty and Goulden (2008), Lloyd and Farquhar (2008) and Slot, Garcia, and Winter (2016)) and canopy (e.g., Tan et al. (2017)) scales, leading some authors to suggest that tropical forests currently exceed their photosynthetic temperature optima (Doughty & Goulden, 2008). However, studies focusing on gas exchange generally only span the typical range of leaf temperatures to which leaves are exposed under ambient climatic conditions and thus focus only on warming impacts that are reversible in nature. Irreversible changes to the photosynthetic apparatus can ensue in leaves exposed to very high temperatures (G. H. Krause, Cheesman, Winter, Krause, & Virgo, 2013) but little is known of the temperature thresholds associated with photosynthetic impairment in Amazonian tree taxa. In the rapidly-warming southern Amazon, there is a concern that maximum leaf temperatures may be approaching critical thresholds, particularly in outer canopy leaves that receive maximum irradiance, but we currently do not know the point at which irreversible thermal changes will occur. Furthermore, regions of Amazon rainforest vary in the length of dry periods as well (Fu et al., 2013) as drought intensity (Marengo, Nobre, Tomasella, Cardoso, & Oyama, 2008). In our study region, the peak of the dry season coincides with the peak of maximum air temperatures. Seasonal acclimation potential of thermal traits in trees, particularly the ones exposed to long dry/warm periods will influence the potential for carbon gain in the forest and may be very important for maintaining plant photosynthetic function under a rapidly warming climate. Variation in thermal traits across seasons has not been measured previously in Amazonia.
The measurement of leaf thermotolerance typically focuses on the response of one of two diagnostics of chlorophyll a fluorescence quenching, measured under dark-adapted conditions (G. H. Krause & Weis, 1991). This is usually analysed via heat shock treatment of leaf discs (G. H. Krause et al., 2013; G. Heinrich Krause et al., 2010) where changes in the ‘minimal’ fluorescence (F 0) the ratio of variable to maximum fluorescence yield (F v/F m), a parameter that is often referred to as the maximum quantum yield (QY) of photosystem II (PSII) are measured (Kitajima & Butler, 1975). The first measurement defines Tcrit as the temperature at whichF0 rises sharply, while the second usesF v/F m to calculateT50 as the temperature associated with a 50% decline in F v/F m. Although the two metrics are indicators of the functional integrity of photosystem II, they are reporters of different underlying physiological mechanisms. A temperature-induced increase in F 0indicates disruption of the light-harvesting antenna, which is attributed to increased thylakoid membrane fluidity (Figueroa, 2003). The decline in F v/F m, on the other hand, signifies a loss of PSII function that has been attributed to disassembly of the light-harvesting complex from the core of PSII (Kouřil et al., 2004; Lípová, Krchňák, Komenda, & Ilík, 2010; Zhang, Liu, & Yang, 2011). Hence, heat-induced changes inF v/F m indicate damage to PSII and are less prone to measurement artefacts (G. Heinrich Krause et al., 2010; Slot, Krause, Krause, Hernández, & Winter, 2018).
In this study, heat-induced changes inF v/F m were used to characterise thermal tolerance for seven dominant tree species in a rapidly-warming southern Amazonian rainforest over two different seasons i.e., the end of the wet season and end of the dry season. The following questions were addressed: a) does the high-temperature tolerance of sun-exposed evergreen trees in the hottest Amazonian site vary across species? b) what is the extent of seasonal plasticity in photosynthetic thermal tolerance over dry and wet seasons? In addition, c) how does thermotolerance of southern Amazonian species compare with published data for trees from other tropical regions where maximum temperatures are typically lower? In addition to T 50, as a proxy for the first heat effects, we assessed two further parameters: 1)T5 , the temperature associated with the onset of the temperature-induced decline inFv /F m and 2)T95, the temperature at whichF v/F m decreased below 95% of the maximum level, taken to be the limit, where PSII functions are effectively lost.