1 Introduction

The diversity and productivity of complex ecosystems, such as tropical forests, are associated with the supply of resources and their efficient use by plant species (Valladares and Niinemets 2008; Ishii et al. 2013; Girardin et al. 2016). Along the tropical forest vertical profile, vegetation structure and composition lead to heterogeneity in resource availability. For example, light intensity is reduced exponentially from the canopy top to the understory (Castro 2000; Kenzo et al. 2015). Only 1–2% of the irradiance available at the top of the canopy is available in the understory layer (Chazdon and Fetcher 1984; Nicotra et al. 1999; Kenzo et al. 2015). Adapted genotypes or acclimated phenotypes can cope with low irradiance in the understory by mechanisms to efficiently intercept, absorb and utilize this resource in the photosynthetic process and thus allow a positive carbon balance (Valladares and Niinemets 2008). On the other hand, the damage caused by energy excitation excess in the photosynthetic apparatus due to the very high irradiance at the top canopy layer, which can exceed 2600 µmol m−2 s−1, is attenuated by tolerance mechanisms (Wright and Colley 1994; Demmig-Adams and Adams 2000; Kenzo et al. 2006).

There is evidence that the microclimate along a tropical forest vertical profile imposes a substantial variation in photosynthesis-related leaf traits (Carswell et al. 2000; Domingues et al. 2005; Kenzo et al. 2006; Kosugi et al. 2012; Weerasinghe et al. 2014; Kenzo et al. 2015). Reduced light availability in the understory favors leaf traits that maximize interception and absorption of radiant energy, targeting a positive carbon budget. These traits include: low leaf mass to area ratio (Evans and Poorter 2001), altered concentration of photosynthetic pigments (Lichtenthaler and Babani 2004), nitrogen partitioning in the thylakoid pigment–protein complexes (Evans 1989; Kenzo et al. 2006), thylakoid stacking in the chloroplast ultrastructure (Boardman 1977) and a low-light compensation point (Craine and Reich 2005; Valladares and Niinemets 2008). High irradiance in the upper canopy favors a different suite of leaf traits that increase light-use efficiency. These include high rates of photosynthesis and leaf dark respiration (Weerasinghe et al. 2014; Kenzo et al. 2015), high palisade parenchyma thickness (Kenzo et al. 2006), and high leaf nitrogen and phosphorus contents (Kenzo et al. 2006; Kosugi et al. 2012; Weerasinghe et al. 2014; Kenzo et al. 2015).

Considering that light intensity is the main factor affecting plant establishment in tropical moist forest (Poorter and Arets 2003), an investigation of the integrity and functionality of the photosynthetic apparatus associated with electron transport efficiency of upper canopy and of understory tree species in tropical forest is essential to an understanding of their possibly disparate mechanisms of light harvesting and use. The traits related to electron transport efficiency can be accessed by the OJIP transient of chlorophyll a fluorescence induction parameters (Stirbet and Govindjee 2011) and thus improve our understanding of light harvesting and use along the tropical forest profile.

Light energy absorbed by photosynthetic pigments can be used by photochemistry, or it can be lost as heat or as re-emission by fluorescence. In dark-adapted leaves, a polyphasic transient in chlorophyll a fluorescence is observed after sudden saturating light exposure (Stirbet and Govindjee 2011; Stirbet et al. 2018). This polyphasic-induced fluorescence curve is the basis for the JIP-test, which extracts parameters associated with successive electron acceptor reductions along the electron transport chain in the thylakoids (Strasser et al. 1995, 2010; Stirbet and Govindjee 2011). Thus, both the light energy assimilated and the energy lost to fluorescence can be obtained, providing an indicator of photosynthetic performance (efficiency). The most widely used JIP-test parameters related to light-use efficiency are: FV/FM, representing the maximum quantum yield of primary PSII photochemistry; PIABS, a performance index for energy conservation from PSII to intersystem electron acceptors (further than quinone A) and PItotal, a performance index for energy conservation from PSII to the reduction in PSI end acceptors (Strasser et al. 2010; Stirbet et al. 2018).

The maximum quantum yield of primary PSII photochemistry (FV/FM) has been widely utilized in studies investigating response of light and CO2-saturated photosynthesis to abiotic stresses (Kalaji et al. 2016; Stirbet et al. 2018). However, under drought or ozone stress, parameters that integrate more electron transport steps (e.g., PIABS and PItotal) are more sensitive than FV/FM (Bussotti et al. 2011; Campos et al. 2014; Kalaji et al. 2016). A body of evidence demonstrates a close association between potential (saturated) photosynthesis and the last steps of electron transport chain parameters accessed by the JIP-test under different stressors (e.g., light and phosphorus availability, Lin et al. 2009; Cunha et al. 2016; ozone stress, Cascio et al. 2010; Bussotti et al. 2011; ozone and water stress, Desotgiu et al. 2012).

Therefore, to extend the use of chlorophyll a fluorescence measurements to examine photosynthetic performance in plants, we ask if JIP-test parameters are more closely associated with light-saturated photosynthesis along the vertical profile of a tropical forest than the more commonly used traits. We hypothesize that the changes in light-saturated photosynthesis caused by the gradient of light intensity from the understory to the canopy can be well represented by fast kinetics fluorescence parameters.

2 Materials and methods

Site and sampling description

– The study site is a Central Amazon terra firme forest on a low plateau with well-drained clay-loam Oxisol. All studied trees were within 200 m of the “K34” micrometeorological tower (2°36.5′S, 60°12.6′W), maintained by the LBA program (Large-Scale Biosphere–Atmosphere Experiment), 90 km north of Manaus, Brazil. The climate is characterized by average annual precipitation of 2400 mm with two dry months (< 100 mm mo−1) in August and September and very little variation in average monthly air temperature (24.6–26.9 °C) seasonality (Araújo et al. 2002).

We compared the leaf traits of a group of 19 upper canopy trees and 26 understory trees (Online Resource 1). The upper canopy group included 16 species, while the understory group had 24. Forty-two percent of the understory trees were from taxa confined to understory during their entire life cycle. The rest were juveniles of mid- and upper-canopy taxa, included because they constitute a large portion of all understory stems. In each of the two environments, we examined leaf traits during four seasons in the year 2015: wet (April); wet/dry transition (July); dry (September) and dry/wet transition (November). All leaf traits were measured during the four seasons, except anatomical traits measured only in the wet season. For all measurements, we used leaves of similar conditions. All were fully expanded, mature and healthy. Upper and middle canopy leaves were accessed using the “K-34” micromet tower and two canopy walkways, each 30 m long and suspended 20–25 m above the ground. We accessed each understory tree from the ground or with a ladder.

Canopy openness and PPFD

– Canopy openness was measured with hemispherical photographs from a Digital Plant Canopy Imager (CI-110, CID Bio-Science, Camas, WA, USA). These were taken above each measured leaf in the four seasons under a cloudless sky at dusk or dawn. PPFD was logged at 30-s intervals from sunrise to sunset for 5–7 days during each of the four seasonal field campaigns. Two PAR sensors (MQS-B/ULM-500 logger, Heinz Walz, Germany) were installed on the LBA tower, one at 52 m height (about 15 m above the surrounding trees) and one in the understory at 1.3 m height.

Leaf morpho-anatomical traits

– We measured the thickness of the entire leaf (LT), of palisade parenchyma (PP) and of spongy parenchyma (SP). These traits were obtained only in the wet season field campaign, using three leaves from each tree. A section of the leaf middle was fixed in FAA (formalin, acetic acid and alcohol) then transferred to 70% ethanol, sectioned with a microtome (15–25 µm thickness) and photographed with a light microscope (Zeiss Axioskop 2 and Zeiss Axio Cam MRC; Zeiss, Jena, German). The 200 × micrographs were analyzed with ANAT QUANTI 2.0 software (Aguiar et al. 2007). Leaf mass per area (LMA), as dry leaf mass in grams at 65 °C per fresh leaf area in m2, was obtained from ten leaf disks from each of three leaves per tree per season.

Chloroplast pigments

– Chloroplast pigment concentrations were obtained for three leaves per tree in each of the four seasons. We macerated 0.1 g of fresh leaf in 10 ml of 80% acetone with MgCO3 then immediately added 10 ml of 100% acetone. The extract was filtered and absorbance measured (Ultrospec 2100 Pro UV/visible, Amersham Biosciences, Cambridge, UK) at the wavelengths of 663, 645 and 480 nm (Lichtenthaler and Wellburn 1983). Following Hendry and Price (1993), we calculated chlorophyll a and b contents per unit of leaf area (µmol cm−2) and their ratio (Chl a:b).

Leaf nitrogen and phosphorus

– Nitrogen and phosphorus concentrations were obtained for three leaves per tree per season. Leaves were oven-dried at 65 °C to constant mass. For leaf N, 0.1 g of dry leaf sample was digested for 10 h in H2O2 + H2SO4 at gradually increasing temperature to 350 °C (Miyazawa et al. 1999). An aliquot of 25 ml of extract was used for total N by the Kjeldahl method of distillation and titration (Bremner 1996). Concerning leaf P concentration, 0.5 g of dry leaf sample was subjected to nitro-perchloric digestion and concentration was determined at 750 nm by spectrophotometry following Murphy and Riley (1962).

Chlorophyll a fluorescence

– Fluorescence and photosynthesis measurements were made using attached leaves and branches. We used a portable fluorimeter (PEA, MK2-9600, Hansatech, Norfolk, UK) to measure chlorophyll a fluorescence, always between 08:00 and 10:00 h. Three leaves per tree per season were selected and kept in the dark for 30 min, sufficient for complete oxidation of the photosynthetic electron transport chain. This dark-adapted leaf was then exposed to saturating light of 3000 µmol m−2 s−1 at a wavelength of 650 nm during 1 s. The chlorophyll a polyphasic transient steps were recorded and the parameters calculated according to the JIP-test (Online Resource 2; Strasser et al. 1995, 1999; Tsimilli-Michael and Strasser 2008; Strasser et al. 2010): After light exposure, the photon flux is absorbed by the PSII antenna (ABS) and a part of the energy is trapped by PSII reaction centers (TR) causing the quinone A (QA) reduction; the Q −A electron is transported to intersystem electron acceptors (ET) and finally to PSI electron acceptors (RE). Using the JIP-test, we calculated the reaction center density (γRC/(1 − γRC) or RC/ABS), maximum quantum yield of PSII (φP0 or FV/FM), efficiency with which an electron moves further than quinone Q −A (ΨE0 or ET0/TR0) and efficiency with which an electron from the intersystem electron carriers is transferred to reducing end electron acceptors at the PSI acceptor side (δR0 or RE0/ET0). From these parameters, we derived the performance index (PIABS) and total performance index (PItotal):

$$ {\text{PI}}_{\text{ABS}} { = }\left( {\frac{{\upgamma{\text{RC}}}}{{ 1 {-}\upgamma{\text{RC}}}}} \right) \cdot \left( {\frac{{{\varphi }{\text{Po}}}}{{ 1 -{\varphi }{\text{Po}}}}} \right) \cdot \left( {\frac{{\varPsi {\text{Eo}}}}{{ 1 -\varPsi {\text{Eo}}}}} \right) $$
$$ {\text{PI}}_{\text{total}} {\text{ = PI}}_{\text{ABS }} .\left( {\frac{{\updelta{\text{Ro}}}}{{ 1 -\updelta{\text{Ro}}}}} \right) $$

Additionally, we derived the IP-phase, which reflects the PSI/PSII stoichiometry (Ceppi et al. 2012).

Photosynthetic leaf traits

– Light-saturated net photosynthetic rate (PNmax) and stomatal conductance (gs) were obtained from photosynthesis–irradiance curves (PN-PPFD). Measurements were made on one leaf per tree per season, between 08:00 and 13:00 h, using an open-flow gas exchange system LI-6400XT (LI-COR, USA). The LI-COR 6400 chamber was adjusted to a flow rate of 400 μmol s−1; 400 μmol mol−1 of CO2 concentration, 21 mmol mol−1 of H2O vapor concentration and 31 °C of leaf temperature. PN-PPFD curves were obtained at eleven steps of PPFD: 2000, 1500, 1000, 750, 500, 250, 100, 75, 50, 25 and 0 µmol m−2 s−1 (Santos-Junior et al. 2013). PNmax and gs were extracted at the saturating light value in each PN-PPFD curve. Saturation occurred between 500 and 2000 µmol m−2 s−1. No PN curves showed decline even at 2000 µmol m−2 s−1.

Statistical analysis

– In the cases of non-normality and/or non-homogeneity, the data were log-transformed (Zar 1999). Associations between canopy openness and leaf traits (PNmax, gs, morpho-anatomical traits, leaf N and P content and fluorescence parameters)—and between PNmax versus leaf traits—were obtained by Pearson’s correlation for each season and for the annual average. All analyses were performed using the statistical program R version 3.5.1 (R Core Team 2018).

3 Results

Canopy openness and irradiance availability

– As expected, canopy openness increased exponentially along vertical forest profile and the PPFD difference between understory and the upper canopy was remarkable (Fig. 1). During the cloudy wet season, upper canopy daytime PPFD was typically in the range 150–750 μmol m−2 s−1 and only 14% of daytime records exceeded 1000 μmol m−2 s−1. In the dry season, 62% of daytime records exceeded this threshold. Transition seasons were between these two extremes, with 50% and 46% of daytime records above 1000 μmol m−2 s−1 in the wet-to-dry and the dry-to-wet transitions, respectively. Average understory daytime PPFD was less than 5% of upper canopy daytime irradiance across all four seasons. Over 75% of the daytime irradiance values recorded in the understory were below 25 μmol m−2 s−1 in all four seasons.

Fig. 1
figure 1

Relationship between percent canopy openness and leaf height in Central Amazon forest (a), photosynthetic photon flux density as a function of time of day, measured above the upper canopy (b) and in the understory (c), from means of 5–7 days during each of the four seasons: wet (black diamond), wet/dry transition (white square), dry (white triangle) and dry/wet transition (cross)

Full size image

Leaf traits associated with canopy openness

– The highest and most seasonally consistent correlations were between canopy openness and six leaf traits: PNmax, LMA, P, RE0/ET0, PItotal and IP-phase (Table 1). These were followed by leaf nitrogen which also had seasonally consistent correlation with canopy openness (r = 0.42–0.62). More moderate positive correlations with canopy openness were seen in three out of four seasons for stomatal conductance and RC/ABS. We found moderate positive correlations between canopy openness and leaf thickness, palisade parenchyma thickness and spongy mesophyll thickness. Total chlorophyll and the ratio Chl a:b had no significant relationship to canopy openness in any season, with one exception. Traits with moderate positive correlation to canopy openness but only in one or two seasons of the year were ET0/TR0 and PIABS. FV/FM had a moderate negative correlation with openness, and only for two out of four seasons.

Table 1 Pearson’s correlation between canopy openness and leaf traits in the four seasons and for average of the seasons
Full size table

Leaf traits associated with P Nmax

– PNmax was significantly correlated in all seasons with seven traits that were a mix of morpho-anatomical traits, stomatal conductance, fluorescence parameters and nutrient concentration. The degree of correlation was seasonal (Table 2). In the dry season, the best association was observed between PNmax and gs, while LMA; the JIP-test parameters RE0/ET0, PItotal and IP-phase; and the two nutrients concentrations (N and P) were the traits best correlated with PNmax in the other seasons.

Table 2 Pearson’s correlation between light-saturated photosynthesis and leaf traits in the four seasons and for average of the seasons
Full size table

When the data were lumped without distinguishing season, the best association was observed between PNmax and RE0/ET0. A strong correlation (> 0.60) was also observed for LMA, PItotal, IP-phase, N and P. Leaf and palisade parenchyma thicknesses were moderately correlated with PNmax in the lumped full year of data.

4 Discussion

The novelty of this study lies, first, in reporting changes along the vertical light profile of a Central Amazon upland forest for a suite of morpho-anatomical and physiological leaf traits. Few investigations of this type have been published for the region (Carswell et al. 2000). Second, and more importantly, we show that, when data for several species are combined, photosynthetic rate variation along the vertical profile of tropical forest can be represented by certain traits derived from the chlorophyll a OJIP fluorescence transient, in the place of traditionally utilized traits. Specifically, IP-phase, RE0/ET0 and PItotal—all of which are JIP-test parameters related to the final steps of electron transport—were strongly correlated with PNmax during all seasons.

The ranges found here for leaf anatomical traits, chloroplast pigments, nitrogen and phosphorus, and photosynthetic leaf traits are consistent with those reported in other tropical forest tree species found in different parts of the vertical canopy profile (Kenzo et al. 2006; Cavaleri et al. 2010; Domingues et al. 2013; Kenzo et al. 2015). The range of FV/FM is also compatible with other surveys for tropical tree species (Krause et al. 2012). While few published values of JIP parameters are to be found for tropical trees, our data are within the ranges for tree species in general (Desotgiu et al. 2012).

Amazon forest leaf-level photosynthesis as explained by conventionally employed leaf traits

– Leaf morphology and anatomy are known to change along the vertical light gradient of tropical forests (Kenzo et al. 2006; Cavaleri et al. 2010). Maximum attainable photosynthetic rate (PNmax) was here shown to be greater as canopy openness and morpho-anatomical leaf traits’ values (LMA, LT, PP and SP) increased (Table 1). A high LMA as a consequence of a thicker leaf and thicker palisade parenchyma layers maximizes the utilization of the higher available irradiance of the upper canopy by increasing the number of mesophyll chloroplasts (Kenzo et al. 2004; Oguchi et al. 2005; Kenzo et al. 2006) and, consequently, increases PNmax. Therefore, the moderate associations between PNmax and morpho-anatomical traits are linked by efficient utilization of the light as canopy openness increases.

With few exceptions, neither total chlorophyll nor Chl a:b was related to canopy openness or photosynthesis (Tables 1, 2). The exceptions were weak associations of canopy openness and photosynthesis with Chl a:b during the wet season and between photosynthesis and total chlorophyll during the dry/wet transition season. Chl a:b ratio has been related to the size of the light-harvesting complex and, consequently, the fraction of energy absorbed by leaves (Niinemets 2010). During the wet season, light can be limiting to photosynthesis in tropical forests (Graham et al. 2003) and thus increasing the importance of efficient light harvesting along the vertical forest profile.

Photosynthesis had a moderate to a strong positive correlation with N and P leaf concentrations (Tables 1, 2). Higher nitrogen and phosphorous investments per unit of leaf area ensure higher maximum rates of photosynthesis in the upper canopy—as expected in a high resource environment—according to the leaf economic spectrum (Wright et al. 2004). Increasing photosynthesis under high irradiance has been related to the partitioning of N to enzymes of the Calvin–Benson cycle and other proteins involved in the photosynthetic process, while P is partitioned to inorganic phosphate and intermediate metabolites, such as Ribulose 1,5 bisphosphate (Evans 1989; Hidaka and Kitayama 2009, 2013). The stronger association between phosphorus and photosynthesis when compared with nitrogen (Table 2) can be related to phosphorus being derived from the geologic substrate and thus strongly limiting in highly weathered Amazonian soils (Vitousek and Sanford 1986; Quesada et al. 2012), whereas N is fixed from the atmosphere by biological processes.

Stomatal conductance was the trait most strongly correlated with leaf-level photosynthesis in the peak dry season month of September (Table 2). The September field campaign coincided with a strong drought across a large part of the Amazon Basin (Jiménez-Muñoz et al. 2016), set off by an El Niño-Southern Oscillation event. Particularly, a photosynthesis decrease in this extreme dry season, compared to a “normal” dry season 1 year later, was caused by stomatal closure at our Central Amazon site (Santos et al. 2018). Therefore, the strong association between PNmax and gs during the dry season could be related to the unusually extreme water deficit of 2015.

JIP-test parameters related to PSII are poor predictors of PNmax

– JIP-test parameters, except for FV/FM, increase together with canopy openness (Table 1), and a moderate-to-strong correlation was seen between PNmax and those JIP-test parameters that are related to the final steps of electron transport (PSI) during all seasons (Table 2). In contrast, we found a weak negative relationship between FV/FM (a PSII parameter) and canopy openness. This could be related to dynamic photoinhibition of PSII processes under high irradiance (Demmig-Adams and Adams 2000). Chlorophyll a fluorescence was recorded between 08:00 and 10:00 h when the irradiance reached values of 600–1600 μmol m−2 s−1, enough to trigger photoprotective processes such as thermal dissipation on sun-exposed leaves (Demmig-Adams and Adams 2006; Desotgiu et al. 2012; Krause et al. 2012). Therefore, the negative correlation between FV/FM and canopy openness, and consequently photosynthesis, can arise from a reduction in energy trapping by PSII primary acceptors as a consequence of non-photochemical quenching. Unfortunately, a pre-dawn condition of FV/FM—after dark recovery—was not possible to measure, precluding a definitive conclusion about dynamic and permanent photoinhibition. In conditions of high irradiance exposure, some tropical tree species cannot recover the pre-dawn values of FV/FM, indicating a chronic photoinhibition (Castro et al. 1995; Ribeiro et al. 2005, Quevedo-Rojas et al. 2018). The PIABS index is a combination of three PSII parameters, all of which were poor predictors of photosynthesis. These are FV/FM, the total number of active PSII reaction centers per unit of absorption flux (RC/ABS), and the efficiency that an electron moves further than quinone Q −A (ET0/TR0). RC/ABS was moderate to weakly correlated with canopy openness and photosynthesis, while ET0/TR0 was correlated only in the wet season (Tables 1, 2). Therefore, the poor, and in some cases absent, correlation of PIABS Index with canopy openness and photosynthesis was a result of weak correlation by its components, principally FV/FM and ET0/TR0. However, it should be noted that a moderate correlation was observed between PIABS and photosynthesis in the wet season, provided by the RC/ABS contribution. Consistent with the close association between photosynthesis and Chl a:b during the wet season, the improved wet season association between PNmax and PIABS is further evidence of light limitation acting on the first PSII steps of photosynthesis (Graham et al. 2003). In an extensive survey of European forest stands (Pollastrini et al. 2016), PIABS was the most effective chlorophyll a fluorescence parameter to describe the photochemical properties of tree species and FV/FM was associated with light-saturated photosynthesis (clustered at first PCA axis) in broadleaved tree species.

JIP-test parameters related to the final steps of electron transport are good proxies of photosynthesis

– Though PNmax was correlated with parameters at different stages of electron transport, it was most strongly correlated with the final stages of electron transport (IP-phase and REo/ETo; Table 2). Thus, jointly with photosynthesis, the IP-phase and REo/ETo were greater as canopy openness increased (Table 1). Under high irradiance levels, the PSI in sun leaves of tropical trees has energy dissipation mechanisms that protect it against photo-oxidative damage (Barth et al. 2001) and allow PSI to continue functioning at high efficiency. Moreover, the higher values of the IP-phase parameter detected in the upper canopy are an indirect indication of a higher PSI/PSII ratio (Schansker et al. 2003, 2005) as also observed in well-watered and nutrient-supplied plants (Oukarroum et al. 2009; Nikiforou and Manetas 2011; Ceppi et al. 2012). The IP-phase amplitude increases in leaves of tree canopies exposed to high irradiance in mixed temperate forest stands (Pollastrini et al. 2016, 2017). The sun leaves of Fagus sylvatica L. trees, for example, have a lower capacity to trap electrons at PSII (low FV/FM), but a great efficiency in reducing the PSI end electrons acceptors (REo/ETo) when compared with shade leaves (Cascio et al. 2010; Desotgiu et al. 2012). Thus, in the well-illuminated upper canopy leaves, a higher PSI:PSII ratio combined with the occurrence of effective PSI protection mechanisms may increase the efficiency of electron transfer to the acceptors at the PSI side (REo/ETo) and the overall efficiency index is increased (PItotal).

The reduction power produced in the electron transport chain in the thylakoid is crucial for the Calvin–Benson cycle intermediate reduction stages and, consequently, higher CO2 assimilation (Sharkey et al. 2007). Furthermore, the PSI acceptor side can be limited by the activity of the Calvin–Benson cycle (Schansker et al. 2003). Therefore, the close association between light-saturated photosynthesis and the fluorescence parameters related to the end steps of electron transport arises from the feedback between high supply and demand of reduction power (NADPH) in a well-illuminated—high resource—environment. In plants submitted to nutrient and water limitation (Lin et al. 2009; Campos et al. 2014) and ozone stress (Desotgiu et al. 2012), REo/ETo and PItotal are low, relative to well-supplied plants. This is related to reduced PSI (as derived from PSI/PSII stoichiometry), which reduces CO2 assimilation.

The investigation of the photosynthetic apparatus function by the rapid and noninvasive JIP-test provides information about photosynthetic performance under various environmental conditions. We have here extended this to include the vertical light profile of a tropical forest. Moreover, we have demonstrated that JIP-test parameters related to the final steps of electron transport predict light-saturated photosynthesis as well as or better than the usually utilized traits.