This item is only available as the following downloads:
1 TEMPERATURE RESPONSES OF LEAF DARK RESPIRATION IN THE UPPER TROPICAL FOREST CANOPY AND THEIR IMPLICATIONS FOR TROPICAL FOREST CARBON BALANCE By MARTIJN SLOT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013
2 2013 Martijn Slot
3 To my parents, for valuing unsupervised play
4 ACKNOWLEDGME NTS I thank my advisor, Kaoru Kitajima, for helping me improve my research through out my PhD and for making me a better scientist. I also thank Klaus Winter and Joe Wright at the Smithsonian Tropical Research Institute (STRI) for letting me use their labor atory facilities and research equipment and for providing feedback on my manuscripts Chapter 2 would not have existed if not for the kindness of the anonymous person who turned in my bag to the lost and found department of the Nederlandse Spoorwegen (Dut ch railways); the bag contained my laptop with all data collected in the 2010 field season as well as the only backups of th ose data. Camilo Rey Snchez has been of great help in the field and the lab and without his help I would not have been able to col lect all the data for C hapters 3 and 4 Thanks also go to Phil Nguyen, who helped with data collection for C hapter 4 Grace C r ummer, Camila Pizano and Danielle Palow helped with nitrogen and carbohydrate analyses for C hapters 2, 3 and 4 My fieldwork would not have been possible without the help of crane operators Edwin Andrade, Jos Herrera and Julio Piti. Thanks also go to Julio for taking me to the gas station to drink a few beers in the parking lot in celebration of my 32 nd birthday, back in 2010. That was special I would also like to acknowledge my committee members Jeremy Lichstein, Tom Sinclair, Tim Martin, and Ted Schuur for feedback on my proposal and manuscripts. I am grateful to all the contributors for the inspiring discussions and presentations at PEERS and the journal clubs at University of Florida (UF) and STRI. I would also like to thank past and current lab mates, office mates, and other friends and colleagues at UF and STRI. I particularly enjoyed working on side projects that are not part of this dissertation with Danielle Palow, Gerardo Celis, Kris Callis, Shaji Faisal, and Ori Baber.
5 Life in Gainesville was much improved by a lot of wonderful people: Pepe, Anglica, Walid, Linda, Maninderpal, and Marlene, for their friendship and never en ding backyard barbeques; and John, Jeff, Damion, and the rest of the Ethnoecology garden crowd for Friday evening gardening and relaxation. Whether in Gainesville, Panama, or anywhere else, I always receive friendship and support from my very best friends: my parents, my brother, and Geetha.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 9 LIST OF FIGURES ................................ ................................ ................................ ........ 10 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 1 4 2 FOLIAR RESPIRATION AND ITS TEMPERATURE SENSITIVITY OF TREES AND LIANAS: IN SITU MEASUR EMENTS IN THE UPPER CANOPY OF A TROPICAL FOREST ................................ ................................ .............................. 19 Background ................................ ................................ ................................ ............. 19 Materials and Methods ................................ ................................ ............................ 23 Study Site and Species ................................ ................................ .................... 23 Field Measurements of Leaf Respiration Rates ................................ ................ 23 Functional Trait Data ................................ ................................ ........................ 25 Data Analysis ................................ ................................ ................................ ... 25 Results ................................ ................................ ................................ .................... 27 Discussion ................................ ................................ ................................ .............. 28 Significant Trait Correlations with Dark Respiration over a Small Trait Space 29 Similarity in Respiration Characteristics among Plant Functional Types .......... 31 Differences between Trees and Lianas ................................ ............................ 32 Concluding Remarks ................................ ................................ ........................ 32 3 TRAIT BASED SCALING OF TEMPERATURE DEPENDENT FOLIAR RESPIRATION IN A SPECIES RICH TROPICAL FOREST CANOPY ................... 41 Background ................................ ................................ ................................ ............. 41 Meth ods and Materials ................................ ................................ ............................ 44 Study Site ................................ ................................ ................................ ......... 44 Respiration Measurements ................................ ................................ ............... 45 F unctional Trait Data ................................ ................................ ........................ 45 Quantification of R and Q 10 ................................ ................................ .............. 46 Estimating Stand level Respiration Fluxes ................................ ....................... 47 Statistical Analyses ................................ ................................ .......................... 49 Results ................................ ................................ ................................ .................... 50 Respiration at 25 C ................................ ................................ ......................... 50 Q 10 Values by Species, Plant Functional Type and Growth Form .................... 50 Variance of Respiration and Q 10 ................................ ................................ ....... 51
7 Trait Correlations and Multiple Regression Models for Respiration and Q 10 .... 51 Stand level Leaf Respiratory Carbon Flux ................................ ........................ 52 Discu ssion ................................ ................................ ................................ .............. 53 Species and PFT Differences in Respiration Traits ................................ .......... 53 Multiple Regression of R and Q 10 ................................ ................................ ..... 54 Annual Leaf Respiratory Carbon Flux at the Stand level ................................ .. 56 Significance for Modeling ................................ ................................ ................. 58 4 THERM AL ACCLIMATION OF LEAF DARK RESPIRATION TO EXPERIMENTAL NIGHTTIME WARMING IN TROPICAL CANOPY TREES AND LIANAS ................................ ................................ ................................ .......... 67 Background ................................ ................................ ................................ ............. 67 M aterials and Methods ................................ ................................ ............................ 70 Study Site and Species selection ................................ ................................ ..... 70 In Situ Warming Protocol ................................ ................................ .................. 71 Dark Respiration Measurements ................................ ................................ ...... 71 Functional Trait Data ................................ ................................ ........................ 72 Data Analysis ................................ ................................ ................................ ... 73 Estimating Stand level Respiration Fluxes ................................ ....................... 74 Results ................................ ................................ ................................ .................... 76 Warming Effect on Respiration ................................ ................................ ......... 76 Warming Effects on Other Leaf Traits ................................ .............................. 77 Across species Correlates of Acclimation ................................ ........................ 77 Consequence of Acclimation for Stand level Respiration Fluxes ..................... 78 Discussion ................................ ................................ ................................ .............. 78 Consistent Acclimation of Respira tion to Elevated Nighttime Temperature ...... 78 Correlates of Acclimation ................................ ................................ .................. 79 Acclimation in both Trees and Lianas ................................ ............................... 81 Consequences of Acclimation for Predicted Respiratory Carbon Fluxes from Tropical Forests ................................ ................................ ............................ 82 Significance for Modeling ................................ ................................ ................. 82 Concluding Remarks ................................ ................................ ........................ 83 5 GENERAL PATTERNS OF THERMAL ACCLIMATION OF LEAF DARK RESPIRATION ACROSS BIOMES AND PLANT TYPES ................................ ....... 92 Background ................................ ................................ ................................ ............. 92 Thermal Acclimation of Leaf Dark Respiration ................................ ................. 92 Potential Differences in Warmin g Response of Tropical and Cool Climate Vegetation ................................ ................................ ................................ ..... 95 Methods ................................ ................................ ................................ .................. 97 Data Selection ................................ ................................ ................................ .. 97 Data Analysis ................................ ................................ ................................ ... 98 Acclimation Type ................................ ................................ .............................. 99 Results ................................ ................................ ................................ .................. 100 Treatment and Species Effects on Acclimation of Respiration ....................... 101
8 In Situ Warming ................................ ................................ .............................. 102 Acclimation Type ................................ ................................ ............................ 103 Discussion ................................ ................................ ................................ ............ 104 Considerations for Quantifying Acclimation ................................ .................... 104 Biome dependent Accli mation Potential? ................................ ....................... 106 Near homeostatic Acclimation to Development Temperature ........................ 107 Type I and Type II Acclimation in Pre existing and Newly developed Leaves 108 Acclimation and Climate Warming ................................ ................................ .. 108 6 CONCLUSIONS ................................ ................................ ................................ ... 121 APPENDIX A SUPPLEMENTARY TABLES AND FIGURES TO CHAPTER 2 ........................... 128 B REFERENCES USED IN THE META ANALYSIS IN CHAPTER 5 ...................... 131 LIST OF REFERENCES ................................ ................................ ............................. 134 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 149
9 LIST OF TABLES Table page 2 1 Species used, their plant functional type and their trait values.. ......................... 34 2 2 Comparison of means of physiological traits and leaf mass per unit area between trees and lianas ................................ ................................ .................... 36 2 3 Comparison of leaf dark respiration rates at 25C ( R 25 ) leaf nitrogen content R 25 per unit leaf nitrogen ( R 25 /N) and Q 10 among studies of sun exposed leaves from tropical forest trees and lianas. ................................ ....................... 37 3 1 Species codes, names, families, plant functional type s and means of dark respiration, leaf mass per unit area, photosynthetic capacity, and concentrations of nitrogen, phosphorus and non s tructural carbohydrates ....... 60 3 2 Results from tests of pairwise correlation between respiration at 25C, Q 10 and other leaf traits ................................ ................................ ............................. 62 3 3 Parameter estimates of multiple regression analysis of respiration and Q 10 against leaf phosphorus photosynthetic capacity leaf mass per area carbohydrate concentration and growth form ................................ ..................... 62 3 4 Distribution of stand level foliar respiration flux among plant functional types.. .. 63 4 1 S pecies used their mean respiration rates per unit area and per unit nitrogen and Q 10 v alues in warmed and control leaves; and acclimation parameter s for R 25 and R 25 /N ................................ ................................ ............. 85 5 1 Summary of results of studies of thermal acclimation of respiration in the lab and field. ................................ ................................ ................................ ........... 110 5 2 P values of models of the dependence of Acclim SetTemp and Acclim Homeo on species traits and experimental conditions. ................................ ...................... 111 A 1 Results o f standardized major axis regression analyses that are shown in Fig. 2 3. ................................ ................................ ................................ ............ 128 A 2 Study species and f amilies, plant functional types, and leaf functional trait values ................................ ................................ ................................ .............. 129
10 LIST OF FIGURES Figure page 2 1 Natural log transformed in situ leaf respiration rates plotted against leaf temperature for 13 tree species and 13 liana species. ................................ ....... 38 2 2 Variation of foliar respiration characteristics within and among plant functional types: early successional trees, mid successional trees, late successional trees, and lianas ................................ ................................ ............ 39 2 3 Correlations between respiration rates at 25C and other leaf traits. .................. 40 3 1 Variation of respiration traits within and among species; within and among plant functional types, and within and between growth forms. ............................ 64 3 2 Proportion of variance in Q 10 R A R M and LMA explained by variance within species, among species within p lant functional type (PFT), among PFTs and between growth forms, as determined by partial R 2 analysis. ............................ 65 3 3 Partial residual plots for the best model of respiration per unit leaf area ( R A ) in which R A is regressed against phosphorus content per unit area, photosynthetic capacity, and leaf mass per unit area ................................ ......... 65 3 4 Partial residual plots for the best model of Q 10 in which Q 10 is regressed against TNC Area and growth form ................................ ................................ ........ 66 3 5 Proportional contribution to the 17 year mean total annual leaf respiratory carbon flux by lianas, early successional mid successional and la te successional tree species ................................ ................................ ................... 66 4 1 Example of acclimation, using the actual leaf temperature data and R 25 and Q 10 values for warmed and control leaves of Castilla elastica ............................ 86 4 2 Example of experimental warming set up ................................ ........................... 87 4 3 Leaf dark respiration at 25C in relation to average nighttime leaf temperature in the precedi ng 6 8 days for the tree species Anacardium Castilla and Luehea and the liana species Bonamia and Stigmaphyllon ........... 88 4 4 Respiration at 25C ( R 25 ) in relation to the average nighttime leaf tem perature during the experiment, standardized by the species mean R 25 of control leaves at their average nighttime temperature ................................ ........ 89 4 5 Correlations between species level acclimation other leaf t raits where acclimation is represented by the slope of R 25 versus the mean nighttime leaf temperature ................................ ................................ ................................ ........ 90
11 4 6 E temp erature (1995 2011), and under 4C warming with different acclimation scenarios. ................................ ................................ ................................ ........... 91 5 1 Acclimation of respiration illustrated ................................ ................................ 112 5 2 Explanation of methods o f quantification of acclimation ................................ ... 113 5 3 Relative change in respiration rate at a set temperature following acclimation of 64 plant species, based on 195 sets of respiration measurements at contr asting acclimation temperatures ................................ ............................... 114 5 4 Acclimation parameters Acclim SetTemp and Acclim Homeo in relation to plant tra its and experimental conditions ................................ ................................ .... 115 5 5 Acclimation parameters Acclim SetTemp and Acclim Homeo in relation to experime ntal conditions and leaf traits ................................ ............................. 117 5 6 Interaction ef fect of duration of warming and degree of warming on Acclim SetTemp .. ................................ ................................ ................................ ... 118 5 7 Frequency of Type I and Typ e II acclimation of respiration .............................. 119 5 8 Acclimation parameters Acclim SetTemp and Acclim Homeo under Type I and Typ e II acclimation of respiration ................................ ................................ ............... 120 A 1 Correlations between Q 10 values, determined from species level temperature response curves of area normalized respiration data and mass normalized respiration da ta, and other leaf traits ................................ ................................ 130
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy TEMPERATURE RESPONSES OF LEAF DARK RESPIRATION IN THE UPPER TROPICAL FOREST CANOPY AND THEIR IMPLICATIONS FOR TROPICAL FOREST CARBON BALANCE By Martijn Slot Augu st 2013 Chair: Kaoru Kitajima Major: Botany Global warming could decrease the net primary productivity (NPP) of tropical forests if warming increase s plant respiration more than gross photosynthesis To address the current lack of empirical data on folia r respiration in relation to temperature from tropical forest I measured respiration rates of upper canopy leaves of trees and lianas in a tropical forest in Panama, and analyzed the response of respiration to temperature in the short term (minutes to hou rs) and in the longer term ( days ) Both in situ and laboratory measurements indicate that the short term temperature response of respiration, expressed as Q 10 (the proportional increase in respiration per 10C warming) was higher than what is commonly a ssumed in global models Respiration rates at 25C ( R 25 ) were also high compared to previously reported values for tropical forests. These results suggest that published global models may underestimate leaf respiratory carb on fluxes from tropical forests. To assess whether mature leaves of tropical forest trees and lianas can acclimate to nighttime warming I experimentally warmed upper canopy leaves for 6 8 days R espiration wa s down regulated by on average 2.9% per degree of warming
13 above ambient temperat ure while the Q 10 remain ed unchanged Acclimation was, however, not completely homeostatic. Accounting for acclimation when scaling respiration to the canopy under a 4C warming scenario reduced the respiration flux compared to no acclimation by 16% or 0. 9 Mg carbon ha 1 yr 1 T o identify general patterns in thermal acclimation of respiration I conducted a meta analysis This revealed that : 1) t he greater the degree of warming the less complete respiration acclimate s 2) the longer the duration of warmin g, the more complete respiration acclimates 3) l eaves that develop under the experimental temperature are better acclimated than leaves that are transferred from a lower temperature and 4) growth forms and biomes do not differ in acclimation potential I n conclusion, I found evidence that leaf respiration of tropical forest c anopy trees and lianas exhibit higher sensitivity to short term temperature changes than commonly assumed but that long term warming results in acclimatory down regulation of respira tion in accordance with acclimation in temperate and boreal forest trees This suggests that with gradual warming acclimation of respiration will minimize the potential decrease in NPP with global warming
14 CHAPTER 1 INTRODUCTION There is an urgent need to improve our ability to predict responses of tropical forests to climate change. Tropical forests account for more than 30% of global net primary production (NPP) (Malhi & Grace, 2000; Saugier et al ., 2001; Huston & Wolverton, 2009), but they are though t to be close to their high temperature threshold (Doughty & Goulden, 2008). Warming by several degrees Celsius, as predicted for the coming decades (Diffenbaugh & Scherer, 2011) will push the majority of tropical forests into a climate envelope currently not occupied by closed canopy forests (Wright et al ., 2009 ). Mitochondrial respiration rates increase exponentially with warming Global rise of temperature, especially during nighttime (Easterling et al ., 1997) may thus have major impacts on the NPP of tr opical forests if gross photosynthetic productivity does not increase with warming However, despite the importance of tropical forests for global NPP insufficient empirical data are currently available to predict temperature responses of tropical ecosyste ms (Reed et al ., 2012). Importantly, uncertainty about leaf respiration in tropical forests hinders accurate and precise modeling of carbon fluxes in tropical forests under current and future climates (Malhi et al ., 2009). How tropical forests will respon d to climate warming cannot yet be answered, but in my dissertation research I have generated original empirical data on leaf respiration and its response to temperature over several timescales. These data are critically needed for predicting carbon balanc e of tropical forests under climate warming. Tropical forests characteristically maintain multiple canopy layers, as a result of which understory seedlings only receive 1 2% of above canopy solar radiation (Clark et al ., 1996). In deep shade photosynthesis and respiration are very low (e.g., Kitajima, 1994),
15 so the largest proportion of the leaf respiratory flux in tropical forests comes from leaves in the sun exposed upper canopy (Cavaleri et al ., 2008). In my dissertation I therefore focus on dark respira tion of sun exposed upper canopy leaves in a semi deciduous tropical forest in Panama, where a construction crane enables canopy access. My dissertation consists of four c hapters Chapters 2 4 report on related studies that combine empirical data with simp le models to explore the ecological implications of temperature responses of leaf dark respiration while C hapter 5 synthesizes the current understanding of leaf respiratory acclimation to warming in a meta analysis of published data from a range of growth forms and biomes In C hapter 2, I test the hypothesis that in situ rates of leaf respiration and the temperature response of respiration differ significantly among species and plant functional types in association with differences in other leaf function al traits. To do so, I compare respiration data from 26 species of tree and liana collected with a new protocol for quantifying the short term temperature sensitivity of leaf respiration under ambient temperature changes, using a series of pre darkened lea ves. While the in situ measurements offer important insight s into the temperature response processes under ecologically relevant ambient temperature variation, greater degree of temperature control can be achieved in the laboratory, and individual leaves can be measured repeated ly In C hapter 3, I report leaf level temperature response curves of respiration measured in the laboratory for 123 upper canopy leaves collected from 28 species of tree and liana. From these curves the temperature sensitivity of re spiration, expressed as the Q 10 (proportional increase in respiration with 10C warming) could be determined at the leaf level. Leaf traits can vary considerable across
16 ecological scales (Messier et al ., 2010), so to identify where most of the variance of leaf respiration traits occurs, the chapter explores how the variance in leaf respiration and Q 10 can be decomposed into variance at the level of growth form (trees versus lianas), plant functional type based on tree successional status, species within gro wth form, and leaves within species. Leaf respiration and Q 10 values of canopy trees are challenging to measure, especially in tall and diverse tropical forests. Obtaining reliable estimates from more readily measured traits is therefore desirable. In C h apter 3, I use species level trait averages to develop trait based models of respiration and Q 10 which are subsequently used to estimate respiration and Q 10 values for 24 tree and liana species. These leaf level estimates are then scaled up to the canopy of the study site, making use of multi year temperature data collected at the site and site specific information on species relative abundance and leaf area index. For these calculations of nocturnal respiratory carbon fluxes, information on the instantane ous temperature response of dark respiration (e.g., the Q 10 ) is important but short term temperature responses cannot be translated into long term responses of respiration to climate change. Global warming has been more pronounced at night than during the day (Easterling et al ., 1997; Alward et al ., 1999) and it is likely that the nighttime temperatures will continue to rise asymmetrically (IPCC, 2007). Such asymmetric warming is expected to reduce NPP, as autotrophic respiration fluxes will rise with incr easingly warmer nights, while daily net photosynthesis is unlikely to increase with daytime warming in the tropics (Cunningham & Read, 2002, 2003a; Doughty & Goulden, 2008; Doughty, 2011). Indeed, in several tropical forests, annual tree growth
17 is negative ly correlated with annual means of daily minimum temperature, rather than with average temperature or atmospheric [CO 2 ] (Clark et al ., 2003, 2010; Feely et al ., 2007), suggesting a strong temperature sensitivity of nighttime respiration of these trees. How ever, acclimation of plant respiration to warming may reduce the potential decline in NPP (King et al ., 2006; Smith & Dukes, 2013). Whether nighttime respiration in tropical tree and liana species can acclimate to elevated nighttime temperatures is current ly unknown. The objective of C hapter 4 was to determine whether dark respiration of fully expanded leaves of tropical trees and lianas can acclimate to experimentally elevated nighttime temperatures. I report thermal acclimation responses of respiration in upper canopy leaves of three tree and two liana species after nighttime warming for 6 8 days. Acclimation is assessed by correlating average nighttime leaf temperature against the rate of leaf dark respiration at a set temperature for each species. These correlations are then used to calculate the canopy level annual nighttime carbon flux associated with foliar respiration under several nighttime warming and acclimation scenarios. Despite recent increased interest in thermal acclimation of respiration (e.g ., Chen & Zhuang, 2013; Smith & Dukes, 2013; Wythers et al ., 2013), general patterns in respiratory acclimation and how it impacts carbon fluxes from the terrestrial biosphere are still lacking. In C hapter 5, a meta analysis of the published data from 30 s ources attempts to identify general patterns of thermal acclimation of leaf dark respiration to warming of terrestrial plant species from across the globe. In total 237 temperature contrasts are included, representing 87 species of forbs, graminoids, shrub s, trees and lianas native to arctic and Antarctic, boreal, temperate and tropical ecosystems.
18 Global warming poses significant challenges on tropical forests through direct warming effect s on high temperature stress of photosynthesis (Doughty & Goulden, 2008; Doughty, 2011), effects of warming induced changes in leaf to air vapor pressure deficit and potential drought induced changes in NPP (Zhao & Running, 2010), and the increased burden of nighttime respiratory carbon loss (Clark et al ., 2003,2010; Feel ey et al ., 2008). While the extent of these potential effects is still poorly understood, the field and laboratory measurements, the field experiments, data analyses, and model simulations I employ in this dissertation should contribute to a better underst anding of how tropical forests may respond to climate warming.
19 CHAPTER 2 FOLIAR RESPIRATION AND ITS TEMPERATURE SENSITIVITY OF TREES AND LIANAS: IN SITU MEASUREMENTS IN THE UPPER CANOPY OF A TROPICAL FOREST Background Tropical forests account for more than one third of global terrestrial net primary productivity (NPP) (Malhi & Grace 2000; Saugier et al 2001), but lack of empirical data on leaf dark respiration ( R ) hinders efforts to reliably model carbon fluxes in tropical forests under current and f uture temperature regimes (Malhi et al 2009). R increases with leaf temperature, and increases in R will reduce NPP if gross photosynthetic productivity remains constant or decreases. Global rise of temperature, especially during nighttime (Easterling et al 1997), may thus have major impacts on the NPP of tropical forests (Galbraith et al 2010). This mechanism is implicated in studies that report a negative correlation between tree diameter growth and annual means of daily minimum temperature (Clark e t al 2003 2010). These results suggest that a temperature driven increase in nighttime respiratory carbon loss reduced the net daily carbon gain available for tree growth. To model plant respiratory carbon efflux from tropical forest trees in response to climate warming, both respiration at a set temperature (e.g., R at 25C) and the temperature sensitivity of R need to be known. The latter is commonly described by the Q 10 ; the proportional increase in R with 10C warming Many coupled climate vegetatio n models assume a constant value of 2.0 for Q 10 i.e. R doubles as temperature increases by 10C (e.g. Cramer et al 2001 ; Cox et al 2004 ; Wang et al 2011). However, Q 10 is not constant over a wide range of temperature s Rather, Q 10 is lower when de termined over higher temperature ranges (Tjoelker et al 2001 ; Atkin et al
20 2005). Available data from tropical forests suggest that R and Q 10 differ widely among tree species (Meir et al 2001) and growth forms (Cavaleri et al 2008). Unfortunately, r eliable species level measurements of R and Q 10 are extremely scarce in tropical forests, hindering efforts to generalize patterns of R and its temperature sensitivity for tropical trees and lianas. Species level data on R and Q 10 will be valuable both to parameterize carbon flux models, and to identify their relationships with other leaf functional traits. If R and Q 10 correlate strongly with commonly measured traits, such as photosynthetic capacity, leaf mass per unit area (LMA), nitrogen or phosphorus c ontent, and leaf lifespan, these relationships may be used to predict R and Q 10 in species rich tropical forests. Global analyses show that R per unit leaf mass at 25C ( R 25 Mass ) correlates positively with leaf nitrogen content and with photosynthetic cap acity (Reich et al 1998 ; S. J. Wright et al 2004), but it remains unknown whether Q 10 correlates with leaf functional traits and also whether R correlates with functional traits within the limited range of trait values observed among upper canopy leave s in tropical forests. Identification of leaf traits that correlate with respiration characteristics would facilitate scaling up from leaf to ecosystem and biome level processes, and may enable trait based vegetation modeling in which traits rather than sp ecies identity are the starting point of the analysis (Van Bodegom et al 2012). Generalizable differences in R among plant functional types (PFTs) provide another means to scale up carbon flux estimates from individual leaves to the canopy in species ri ch tropical forests. PFTs in tropical forests, such as lianas or tree species of different successional status, can be identified with remote sensing techniques (e.g.,
21 Lefsky et al 2002 ; Kalacska et al 2007 ; Alvarez Aorve et al 2012). If generalizab le patterns of R and Q 10 exist among PFTs, estimation of carbon fluxes of a given tropical forest could be facilitated when the relative abundance of different PFTs is known. Leaf traits of tree species differ among PFTs defined by their successional statu s, from the trait syndrome associated with high metabolic activity in fast growing early successional trees, to the syndrome associated with the conservative growth strategy of late successional trees (Reich et al 1995; Kitajima & Poorter 2008). Lianas represent an important PFT in tropical forest canopies, and their abundance has increased in several tropical forests over recent decades (Phillips et al 2002 ; I. J. Wright et al 2004 ; Ingwell et al 2010 ; Schnitzer & Bongers 2011). Many lianas are fast growing because they invest proportionally less in structural support and more in metabolism than trees. Indeed, in a Costa Rican rainforest, Cavaleri et al (2008) found that lianas had on average higher R per unit leaf area at a given temperature t han trees. Their data also showed that lianas had a marginally lower Q 10 than trees, suggesting that lianas may have the competitive advantage of lower respiratory carbon loss at higher temperatures. However, these differences between trees and lianas are yet to be confirmed in another tropical forest. The main objective of this study was to quantify in situ leaf respiration rates of trees and lianas in the upper canopy of a tropical forest, in relation to leaf temperature and leaf functional traits, and t o determine whether plant functional types differ in respiration characteristics. More specifically, the following hypotheses were tested: 1) Leaf respiration rates and Q 10 values vary among 26 species of tree and liana, showing significant differences amo ng plant functional types. 2) Species differences in R and Q 10
22 are associated with differences in other leaf functional traits. We hypothesized that early successional species would have higher R than later successional species, in concordance with general successional status, and that lianas would have higher R than trees, in accordance with Cavaleri et al (2008). We further hypothesized that R would correlate with leaf traits according to the predictions of the leaf economic spectrum ( I. J. Wright et al 2004). Species differences in Q 10 may be caused by differences among species in factors controlling the rate of R at different temperatures (Atkin & Tjoelker 2003), such that fast growing species with hi gh demand for respiratory products (energy and carbon skeletons) are likely to exhibit a greater increase in R with temperature than slow growing species with lower respiratory demand. We thus expected Q 10 to vary among PFTs and correlate with leaf traits associated with growth and metabolism, such as R and photosynthetic capacity. We developed a new protocol to conduct in situ measurements of R of intact and pre darkened leaves equilibrated with ambient temperature in the upper canopy. Leaf respiration of tropical canopy trees is frequently measured on cut off branches taken to the laboratory (e.g. Cavaleri et al ., 2008 ; Metcalfe et al 2010 ; Van de Weg et al 2012). Many such studies confirm the adequacy of laboratory measurements with cut branches with a pilot study, but the magnitudes of potential artifacts are rarely reported in the literature. Furthermore, many large tropical leaves cannot be enclosed entirely in a standard sized gas exchange chamber, and warming only the portion of the leaf enclosed in the chamber may cause artifacts. The in situ measurement protocol used in
23 the study avoided these problems, and allowed sampling of a large number of leaves in ambient conditions of temperature and humidity. Materials and Methods Study S ite and Spec ies The study was conducted in Parque Natural Metropolitano (PNM, 859'N, 7933'W, 100 m a.s.l.), a seasonally dry tropical forest on the Pacific coast of the Republic of Panama, near Panama City. Annual rainfall at the site averages 1740 mm, most of which falls during the rainy season from May through December. The park is a 256 hectare natural reserve consisting of 80 150 y ea r old secondary forest with tree heights up to 40 m. A 42 m tall construction crane with a 51 m long jib (Parker et al 1992) allow ing repeated non de structive measurements of upper canopy leaves (Kitajima et al 2005). Tree and liana species were selected to represent a variety of functional types defined in terms of growth form (trees and lianas) and successional status (early mi d and late successional tree species) (Prez & Condit 2012). Thirteen tree species and 13 liana species were selected from the upper forest canopy (Table 2 1). Additional leaf trait data had been collected previously at the same site for most of these sp ecies (S. J. Wright. Unpublished ). Together these 26 species cover > 70% of the canopy area in reach of the crane (Avalos & Mulkey 1999) Field M easurements of L eaf R espiration R ates All measurements were made in the wet season of 2010 (between late Septemb er and late November) when all study species had mature and non senescing leaves. For each species two to five sun exposed terminal shoots were selected on one to three individuals. Prior to sunset (6.00 p.m.) on the day before respiration measurements, 10 to 34 recently matured leaves were covered with thin aluminum foil,
24 so that they would not be exposed to any sunlight till the measurements. The abaxial side was not completely covered to allow free gas exchange during night. The following morning respira tion was measured once on each leaf between 5.45 and 11.00 a.m. at ambient temperature. Before each measurement, we measured the leaf and ambient air temperature with a thermocouple. The air temperature in the upper canopy rose from 23 24C pre dawn to 28 31C at 11.00 a.m., and the temperature of leaves that were kept covered with aluminum foil closely followed the ambient air temperature. Because the leaves were covered overnight our measurements avoided the effects of light enhanced dark respiration (Atk in et al 1998), and light induced metabolites and respiratory gene expression (Florez Sarasa et al 2012) on respiration. Thus, our measurements attempted to estimate nighttime dark respiration in a manner to minimize the effects of respiration associat ed with carbohydrate exports and temporal variation in substrate availability during the night (Noguchi 2005). Leaf R was measured as CO 2 release rates with a portable infrared gas analyzer (LI 6400, Licor Lincoln, NE, USA), at ambient humidity (70 90%) and a set CO 2 concentration of 400 ppm maintained with the built in CO 2 mixer. The block temperature of the LI 6400 was set to equal the ambient leaf temperature just before the measurement. Thus, during the respiration measurements, the leaf portion insi de the gas exchange cuvette had the same temperature as the whole leaf, avoiding the potential measurement artifacts associated with warming a single leaf or leaf portion opposed to warming of the whole plant (Atkin et al 2000) or stand (Griffin et al 2002a). After the measurement at a single temperature, each leaf was harvested and brought back to the lab for additional trait measurements.
25 Functional T rait D ata Photosynthetic capacity (A max ) was measured on a separate set of 3 6 leaves per species, sim ilar in sun exposure and apparent leaf age, between 9.00 and 10.00 2 s 1 and 400 ppm CO 2 at ambient temperature (range 26 29C). After measurement of the leaf area with a LI 3000 leaf area meter ( Licor ), and leaf mass per area (LMA). Five randomly selected leaves per species on which respiration was measured were ground for analysis of nitrogen (N) concentration with an elemental analyze r (Costech Analytical, Los Angeles, California, USA). We report R on an area basis ( R Area ) unless otherwise specified, but we also calculated respiration per unit leaf mass ( R Mass ) by dividing R by LMA of each leaf. The median leaf lifespan and mean leaf phosphorus concentrations had been collected independently at the species level from the same site (S.J. Wright U npublished). The species means for A max N and LMA in this independent data set match well with the species means determined in the current st udy ( R 2 > 0.65 for each trait comparison). Data A nalysis In our protocol, each leaf was measured only once at one specific temperature, and we estimated temperature dependence (Q 10 ) and respiration at a standardized temperature of 25C at the species level by pooling all measurements from each species. The temperature dependence of R was evaluated for each species from least square regression of natural log transformed R on leaf temperature: ln( R ) = a + bT leaf ( 2 1 ) In equation 2 1, a (the inter cept) and b (slope) are species specific constants, used to estimate species specific Q 10 as:
26 Q 10 = e 10b ( 2 2 ) We report Q 10 values calculated from the temperature response of R Area but also calculated Q 10 of R Mass (Table A 1). Confidence interv als for the Q 10 estimates were calculated from the confidence intervals associated with the parameter b in equation 2 1. Subsequently, R 25 of each leaf was calculated as: ( 2 3 ) where R and T were the actual measurements for each leaf. The sample size varied from 10 to 34 leaves per species (Table 2 1), with larger sample size associated with species that could be sampled from multiple trees under the crane. For four out of 26 species the temperature range of R measurements di d not include 25C (with minimum measurement temperature between 25 and 27C). Thus, we also calculated R 27 in a similar manner as R 25 because 27C fell within the temperature range for all species (Table A 1). However we report only R 25 values, because they are widely reported in the literature, and because trait correlations and comparisons among PFTs were similar for R 27 and R 25 Trait correlations were analyzed with Standardized Major Axis Regression using the SMATR package (Warton et al 2012) in R (R Development Core Team, 2011). Slopes of trait correlations were compared between trees and lianas, but because no differences were found, all species were pooled. Not enough data were available for comparison of trait correlations among PFTs. Comparison s among species, PFTs, and growth forms were made using one way ANOVAs and Tukey HSD post hoc tests, or tests. Data were transformed to improve normality and homogeneity of
27 variance where necessary. All statistical analyses were performed in R version 2.14.1 (R Development Core Team, 2011). Results Respiration increased significantly with temperature for 22 of the 26 species (Fig. 2 1). The four species for which R did not increase significantly with temperature were included in the analyses to avoid biasing against low Q 10 values. Both the elevation and the steepness of the slopes of the temperature response curves of ln R varied across species, resulting in considerable variation in R 25 and Q 10 (Table 2 1). The average R 25 per unit leaf area ( R 25 Area ) did not differ significantly among plant functional types (PFTs) (Fig. 2 2 A ). R 25 expressed on a mass basis ( R 25 Mass ) was higher in lianas than in trees (Table 2 2), but differences among early mid and late successional tree species were not significant (Fig. 2 2 B ). The Q 10 varied widely among species (range 1.24 3.66, mean 2.19; Table 2 1), exhibiting similar patterns across PFTs for Q 10 derived from area and mass based R Q 10 did not differ significantly between trees and lianas (Table 2 2 ), but Q 10 values were higher for early successional trees than for each of the other PFTs (Fig. 2 2 C ). The mean Q 10 for all 26 species was marginally greater than 2.0 (Q 10 Area = 2.18, P = 0.09; Q 10 Mass = 2.26, P = 0.03. t test), although the 95% confide nce intervals of the Q 10 of individual species included 2.0 for 25 of 26 species. R 25 Area correlated positively with photosynthetic capacity (A max ), LMA, and concentrations of N and P ( P < 0.01 for all) (Fig. 2 3 A D ; Table A 1). R 25 Mass showed positive c orrelations with A max N, P, and negative correlations with LMA and leaf lifespan (Fig. 2 3F J ; Table A 1 ). When analyzed separately the slopes of these correlations were the same for trees and lianas ( P > 0.1 for all comparisons of slopes).
28 Whereas R 25 s howed consistent and strong correlations with other leaf functional traits, Q 10 values did not correlate with R or any of the other leaf traits ( Fig A 1 ). Weighted regression analysis, in which species were weighted by the goodness of fit of the respirati on temperature response curves improved the correlations, but did not produce significant results either (data not shown). Discussion In this study we determined in situ leaf dark respiration rates and Q 10 values for tropical trees and lianas in the upper canopy of a seasonal tropical forest, using a new protocol to measure leaves darkened overnight. In our method, the whole leaf was equilibrated to the ambient temperature, and activation of photosynthetic metabolism was unlikely given continuous leaf darke ning overnight till the measurement time. S pecies level estimates of R 25 from our study are relatively high when compared with the available data of R measured on upper canopy leaves of other tropical trees (Table 2 3). Variations in ecosystem characteris tics and methods of respiration measurement may both have contributed to the differences among studies. R tends to be higher for trees on fertile soils than infertile soils when foliar N and P contents are accordingly high (Meir et al 2001 ; Turnbull et a l 2005). The soil at PNM is less acidic and more fertile than many tropical rainforests (pH = 7.01; total exchangeable cations = 4.3 g kg 1 ; available P = 5.81 mg kg 1 B L Turner, p ers c omm 1 ). However, mean N content per unit leaf mass was comparable to that at most other tropical sites, with the exception of nitrogen poor sites in Venezuela (Reich et al 1998) and Australia (Pearcy 1987). Consequently, R 25 per unit leaf N was considerably higher at our site than at 1 Email correspondence on 11 October 2012. Averages based on unpublished data.
29 most other tropical sites (Table 2 3), but comparable to the mean R 25 /N from the eight tree species in Venezuela (Reich et al 1998). Another explanation for the relatively high rates of R we measured may be the large number of early and mid successional species that are fast growing a nd metabolically active. However, R 25 of early successional species was not systematically higher than in late successional species on a unit area basis (Fig. 2 2 A ) and not significantly so on a per unit mass basis (Fig. 2 2 B ). It is also possible that mea suring intact sun exposed top canopy leaves contributed to the high rates of R in our study, contrasting with studies that used detached branches, or that measured leaves from a tower with which the most sun exposed position of the upper canopy may not hav e been accessible. Studies that use detached branches generally report that tests have been made to assure that measurements of attached and detached leaves yield the same results (e.g. Cavaleri et al 2008), but our measurements were also higher than in situ measured R (Table 2 3). Significant Trait C orrelations with D ark R espiration over a S mall T rait S pace R 25 only varied by a factor of 2.5 in our study, but nevertheless we found trait correlations similar to those of the mass based leaf economics spec trum ( I. J. Wright et al 2004 ) in which R varied by more than an order of magnitude. Furthermore, these trait correlations also exist on a leaf area basis, with coefficients of correlation comparable to those found for leaf area based correlations report ed in I. J. Wright et al (2004 ). R 25 Area ranged from 0.7 to 1.8 mol m 2 s 1 (Table 2 1) and this small interspecific variation correlated with other leaf traits that themselves cover a relatively small (2 4 fold) range of trait values (Table 2 1, Table A 2 ). Although the percent of interspecific variation in R 25 Area explained by other leaf traits was modest, the
30 relationship was robust despite the small range of R 25 among species. This result thus offers potential for a trait based approach to vegetatio n modeling (Van Bodegom et al 2012). Q 10 did not correlate with any of the other functional traits we determined. Previous observations in which the temperature sensitivity of R correlated with leaf traits were made within species across depths in the ca nopy (Griffin et al 2002b ; Turnbull et al 2003). In both cases Q 10 correlated positively with leaf N content, but while Griffin et al (2002 b ) found that lower canopy leaves of Populus deltoides had higher Q 10 values and higher N (per unit leaf mass), Turnbull et al (2003) found for several species that Q 10 values were higher in top canopy leaves, as was N (per unit leaf area). There is no clear mechanistic explanation for why the relationship between leaf N and respiratory temperature response exists. Similar to our study, Bolstad et al (1999) found no correlation between Q 10 and leaf N for 18 deciduous temperate tree species, despite considerable variation in shade tolerance and associated photosynthetic properties of the species. Interestingly, the average of the Q 10 values across 26 species in our study was considerably higher than the 2.0 that is often used in coupled climate vegetation models. Because the Q 10 changes with the temperature interval over which it is measured Atkin et al (2005) desc ribed a temperature dependent Q 10 : Q 10 = 3.09 0.043*T, where T is the measurement temperature. According to this relationship, a Q 10 of 1.9 is predicted at 27C, the average mid point of the temperature intervals over which we determined the Q 10 Our Q 10 d ata are comparable to Q 10 measured on other tropical forest canopy leaves (Table 2 3). Both the fixed Q 10 of 2.0 as used in many
31 coupled climate vegetation models and the prediction from the temperature dependent Q 10 of Atkin et al might thus be underesti mating the Q 10 for tropical species. Similarity in R espiration C haracteristics among P lant F unctional T ypes We found limited patterns of Q 10 across 26 species of tree and liana. Similarly Bolstad et al (1999) found that interspecific variation in Q 10 amon g 18 temperate tree species was apparently independent of PFT. We did, however, find that early successional trees exhibited higher Q 10 than later successional trees. There is no obvious mechanistic reason for this difference. Fast growing, light demanding species, such as the early successional tree species in our study, have high demand for respiratory products, and they can achieve high R when respiratory substrate is available (Noguchi & Terashima 1997). Conditions of high demand for respiratory produc ts under non limiting substrate supply can also result in high Q 10 values (Slot et al 2008). However, R 25 was not significantly higher in early successional species so it seems unlikely that the high Q 10 we observed was solely driven by differences in me tabolic demands among PFTs. More research will be needed to establish the generality of the pattern we observed. R measurements were made on a relatively large number of leaves per species, but because temperature dependence was not determined at the indiv idual leaf level, wide 95% confidence intervals are associated with some of the Q 10 estimates (Table 2 1). Future work should include leaf level measurements of Q 10 repeated within species so the intra specific variation in Q 10 as well as of R can be consi dered when analyzing patterns across large numbers of species and PFTs.
32 D ifferences between T rees and L ianas R 25 Mass was higher, while Q 10 values slightly lower in lianas than in trees, similar to the trend found by Cavaleri et al (2008) in Costa Rica. The Q 10 differences were, however, small and not statistically significant (Table 2 2). Given that the global temperature has risen just about 0.2C per decade over the last 30 years (Hansen et al 2006) the observed differences in respiration characteri stics between trees and lianas were too small to suspect that temperature responses of respiration could have contributed to the increase in liana dominance in tropical forests over the last 30 years. Meanwhile, the similarity in area normalized physiologi cal traits of trees and lianas suggests that vegetation models could justifiably ignore growth form and consider all canopy leaves as equal, but use leaf traits to fine tune the model to account for underlying physiological differences among leaves of diff erent species, plant functional types, and growth forms. Concluding R emarks Given the species richness of tropical forests, the lack of species level respiration data from tropical canopy species has been striking; indeed, the paucity of respiration data i s a major source of uncertainty in modeling carbon fluxes in tropical forests (Malhi et al 2009). The R 25 data generated from a total of 461 leaves measured in situ across 26 species of trees and lianas in the current study are a valuable addition. Furth ermore, the results support that R 25 can be estimated from easy to traits that are widely measured for tropical canopy trees, similar to the global trends ( I. J. Wright et al 2004). On the other hand, trees and lianas are similar in R 25 an d in R 25 trait correlations and large variation in R and leaf traits exists within each PFT. These results strongly support the notion that trait based vegetation modeling is more promising than PFT
33 based modeling of ecosystem atmosphere fluxes (Van Bodego m et al 2012). Especially in diverse tropical forests, bypassing species as a working unit and instead using trait data to model ecosystem processes would greatly simplify data collection. The mean Q 10 across 26 species was greater than 2.0, the value of ten assumed in coupled climate vegetation models. This significantly higher Q 10 over the temperature range representative for current and near future nighttime in the tropics should be considered for modeling carbon fluxes in tropical forests under climate warming scenarios.
34 Table 2 1. Species used, their plant functional type (PFT, ES: early successional, MS: mid successional, LS: late successional, L: liana) and their trait values. Respiration ( R ) at 25C ( R 25 ) and Q 10 were determined from least square regression of in situ measurements of R of n leaves per species. Confidence intervals (CI) for Q 10 were determined from the CI of the slope of the temperature response curves (Fig. 2 1). CIs of R 25 were determined assuming normal distribution, as mean R 25 1.96 *(standard error of R 25 ). Also reported are photosynthetic capacity (A max ) leaf mass per unit leaf area (LMA), % leaf nitrogen (N) determined for 5 leaves per species. Phosphorus (P) and leaf lifespan data were obtain ed from S.J. Wright (Unpublishe d ). nd : no data available. Species PFT n R 25 Area (95% CI) R 25 Mass (95% CI) Q 10 Area (95% CI) A max LMA N P Lifespan mol m 2 s 1 nmol g 1 s 1 mol m 2 s 1 g m 2 % % days Albizia guachapele (Kunth) Harms ES 18 0.76 (0.7 0.8) 8.8 (8.0 9.4) 3.66 (1. 8 7. 6 ) 12.4 87 3.3 0.10 183 Annona spraguei Saff ES 11 0.95 (0. 9 1.0) 10.7 (10.2 11.2) 2.79 ( 2.0 4.0 ) 13.5 89 2.8 0.17 216 Cecropia peltata L ES 13 1.25 (1.1 1.4 ) 12.5 (10.9 14.4) 3.14 (1.4 6.8) 19.8 100 2.8 0.17 114 Pittoniotis trichantha Griseb ES 12 1.1 9 (1.1 1.3) 17.5 (15.6 19.3) 2.58 (1.5 4.2) 13.5 68 2.3 0.14 180 Astronium graveolens Jacq. MS 22 1.22 (1.2 1.3 ) 13.8 (13.1 14.9) 1.24 (0.8 1.8) 12.8 88 3.0 0.20 271 Castilla elastica var.costaricana, (Liebm ) CC. Berg MS 29 1.15 (1.1 1.2 ) 11.0 (10.5 11.5 ) 2.84 (2.2 3.6) 19.5 104 2.8 0.17 180 Ficus insipida Willd MS 13 1.45 (1.4 1.6 ) 9.9 (9.2 10.6) 1.70 (1.1 2.8) 23.4 147 2.3 0.17 92 Luehea seemannii Triana & Planch. MS 34 1.14 (1.1 1.2 ) 10.9 (10.5 11.3) 2.12 (1.6 2.8) 19.8 105 2.4 0.15 186 Pseudobomba x septenatum ( Jacq.) Dugand MS 22 1.17 (1.1 1.3 ) 11.5 (10.7 12.1) 1.51 (1.1 2.1) 16.3 102 2.3 nd nd Spondias mombin L MS 16 1.48 (1.4 1. 6 ) 14.5 (13.8 15.0) 1.73 (1.3 2. 3 ) 16.5 102 2.7 0.13 173 Zuelania guidonia (Sw.) Britt. & Millsp. MS 13 1.36 (1.3 1. 5 ) 11.0 (10.1 11.8) 2.42 (1.5 3.9) 17.3 124 2.0 nd nd Anacardium excelsum (Bertero & Balb ex Kunth) Skeels LS 23 1.22 (1.2 1.3 ) 11.2 (10.8 11.5) 1.99 (1.7 2.3) 13.8 109 1.4 0.14 280 Chrysophyllum cainito L LS 25 0.81 (0.8 0.9 ) 7.0 (6.7 7.3) 1.94 (1.6 2.3 ) 14.1 116 1.8 0.10 153 Amphilophium paniculatum ( L .) K unth L 23 0.84 (0.8 0.9 ) 12.8 (12.0 13.4) 1.75 (1.3 2.3) 10.9 66 2.8 0.17 122 Aristolochia tonduzii OC. Schmidt L 20 1.02 (1.0 1. 1 ) 12.6 (11.9 13.4) 2.19 (1. 7 2.9) 12.1 81 3.1 0.15 173 Bonamia trich antha Hallier f. L 18 0.80 (0.8 0. 9) 11.7 (10.8 12.6) 1.82 (1. 1 3.1 ) 10.4 69 3.2 0.18 179
35 Table 2 1. Continued. Species PFT n R 25 Area (95% CI) R 25 Mass (95% CI) Q 10 Area (95% CI) A max LMA N P Lifespan mol m 2 s 1 nmol g 1 s 1 mol m 2 s 1 g m 2 % % days Cissus erosa Rich L 13 1.30 (1.2 1.4) 22.7 (20.4 24.8) 2.45 (1.6 3.8) 17.2 57 3.2 nd 89 Combretum fruticosum (Loefl.) Stuntz L 13 1.09 (1.0 1.2) 14.5 (10.3 14.3) 2.19 (1.2 4.0) 17.7 75 2.3 0.16 154 Forsteronia myriantha Donn. Sm. L 11 0.99 (0.9 1 .1) 14.8 (13.4 16.2) 2.46 (1.2 5.1) 17.5 67 2.9 0.16 107 Gouania lupuloides (L.) Urb. L 22 0.90 (0.8 1.0) 18.9 (16.7 21.7) 1.49 (1.0 2.3) 12.7 48 4.0 0.18 92 Mikania leiostachya Benth. L 10 0.72 (0.6 0.8) 12.4 (10.8 13.9) 2.58 (1.1 6.1) 9.8 59 2.4 0.13 2 13 Phryganocydia corymbosa (Vent.) Bureau ex K. Schum L 17 1.79 (1.6 2.0) 18.3 (16.7 19.5) 1.69 (1.1 2.6) 12.2 98 3.7 0.15 89 Serjania mexicana (L.) willd L 11 1.01 (0.9 1.1) 10.9 (9.8 11.9) 2.52 (1.4 4.7) 12.9 93 2.4 0.15 nd Stigmaphyllon lindenianum A Juss. L 12 1.05 (1.0 1.1) 13.2 (12.5 14.0) 1.62 (1.3 2.0) 16.3 80 2.7 0.14 150 Trichostigma octandrum (L.) H. Walt. L 23 1.15 (1.1 1.2) 20.1 (18.6 21.5) 2.13 (1.3 3.6) 15.1 57 4.1 0.32 57 Vitis tillifolia Humb. & Bonpl. ex Roem. & Schult L 17 0.94 ( 0.9 1.0) 14.3 (12.9 15.7) 2.34 (1.0 5.5) 13.0 66 2.1 0.13 66
36 Table 2 2. Comparison of means ( 1 standard deviation) of physiological traits and leaf mass per unit area (LMA) of trees (13 species) and lianas (13 species). Respiration at 25C ( R 25 ), Q 10 and photosynthetic capacity (A max ) are shown A M basis. The ratio of respiration over photosynthetic capacity is also shown. Trees Lianas R 25 Area (mol m 2 s 1 ) 1.17 0.22 1. 05 0.27 R 25 Mass (nmol g 1 s 1 )* 11.56 2.63 15.17 3.65 Q 10 Area 2.28 0.67 2.10 0.37 Q 10 Mass 2.36 0.60 2.16 0.55 A max Area (mol m 2 s 1 )* 16.37 3.43 13.36 2.64 A max Mass (nmol g 1 s 1 )* 160.0 26.3 197.0 57.6 R 25 / A max 0.07 0.01 0.08 0.02 LMA (g m 2 )§ 103.2 19.3 70.4 14.6 Trees and lianas different ( P < 0.05), § P < 0.01
37 Table 2 3. Comparison of leaf dark respiration rates at 25C ( R 25 ), leaf nitrogen content (N), R 25 per unit leaf nitrogen ( R 25 /N) and Q 10 amo ng studies of sun exposed leaves from tropical forest trees and lianas (Means standard deviations of n species). Location Growth form n R 25 N R 25 /N Q 10 (nmol g 1 s 1 ) (mg g 1 ) (mol (g N) 1 s 1 ) Panama 1 (PNM This study) Trees 13 11. 6 2.6 24.5 5.2 0.49 2.28 Lianas 13 15.2 3.7 29.9 6.4 0.51 2.10 Costa Rica 2 (La Selva) Trees 19 3.8 nd 0.32 2.34 Lianas 6 6.8 nd 2.14 Venezuela 3 (San Carlos) Trees 8 8.6 4.4 16.2 4.3 0.53 nd Bra zil 4 (Jaru) Trees 6 2.8 25.0 0.11 2.30 Brazil 5 (Caxiuaa reserve) Trees 8 4.6 1.1 nd Brazil 6 (Flona Tapajs) Trees 6 7.4 4.9 21.2 6.4 0.35 nd Lianas 6 7.2 2.9 20.8 3.2 0.35 nd Cameroon 4 (Mbalmayo) Trees 6 4.7 25.7 0.18 2.00 Australia 7 (Curtain Fig NP) Trees 1 4.4 1.6 16.7 4.1 0.26 nd Indonesia 8 (Bogor) Trees 3 nd nd nd 2.00 1 current study; 2 Cavaleri et al (2008), R measured on detached leaves; 3 Reich et al (1998); 4 Meir et al (2001); 5 Metcalfe et al (2010); 6 Dominquez et al (2007), R obtained from photosynthetic light response curves; 7 Pearcy (1987); 8 Stocker (1935). Where necessary, R values were converted into R 25 using a Q 10 of 2.0. nd : no da ta available.
38 Figure 2 1. Natural log transformed in situ leaf respiration rates (ln ( R a rea )) plotted against leaf temperature (T leaf ) for 13 tree species ( A M ) and 13 liana species ( N Z ). Species are indicated by their genus name alone; see Table 1 for full species names and plant functional type the species belong to. Solid black lines represent the linear least square fits (significance of fit is indicated with P values); solid gray lines indicate the 95% confidence intervals of the fits; dashed black lines represent the 95% confidence intervals of the data.
39 Figure 2 2 Variation of foliar respiration ( R ) characteristics within and among plant functional types (PFTs): early successional trees (ES, 4 species), mid successional trees (MS, 7 species), l ate successional trees (LS, 2 species), and lianas (13 species). A) R per unit leaf area at a temperature of 25C ( R 25Area ) B ) R per unit leaf mass at 25C ( R 25 Mass ) C ) Q 10 calculated from R per unit area. The box plots indicate the median, 25 th and 75 t h percentile for each PFT. Whiskers extend to 1.5 times the interquartile range. Different letters indicate groups that are significantly different from one another ( P < 0.05) (One way ANOVA with Tukey post hoc testing).
40 Figure 2 3. Correlations between respiration rates at 25C ( R 25 ) and other leaf traits A E) Traits per unit leaf area. F J) Traits per unit leaf mass All axes are on a natural log scale. Each data point represents one species. A F ) R 25 is correlated with photosynthetic capacity ( A max ). B,G) R 25 vs. leaf mass per unit area (LMA ) C,H) R 25 vs. leaf nitrogen content (N) D,I) R 25 vs. leaf phosphorus content (P) E,J) R 25 vs. leaf lifespan. Solid lines represent significant correlation of the traits (trees and lianas combined) ( P < 0.05), as determined by standardized major axes regression. The dashed line in e indicated P < 0.1.
41 CHAPTER 3 TRAIT BASED SCALING OF TEMPERATURE DEPENDENT FOLIAR RESPIRATION IN A SPECIES RICH TROPICAL FOREST CANOPY Background Tropical forests account for more than one third of global terrestrial gross primary productivity (GPP) (Beer et al ., 2010), but 30% of this photosynthetically fixed carbon is released back into the atmosphere by leaf respiration ( R ) (Chambers et al ., 2004; Malhi et al ., 2011). Global rise of temperature, especially during nighttime (Kukla & Karl, 1993; Easterling et al ., 1997) may have major impacts on the net primary productivity (NPP = GPP autotrophic respiration), as autotrophic respiration increases with temperature. Tropical forests are likely to experience unprecedented warming within the next two decades (Diffenbaugh & Scherer, 2011), but as long as lack of empirical data on R and its temperature sensitivity hinders efforts to reliably model current carbon fluxes in tropical forests (M alhi et al ., 2009), quantitative predictions of changes in these fluxes under future climates will remain ambiguous at best. R of canopy trees is challenging to measure, especially in tall and diverse tropical forests. Eddy covariance techniques capture n octurnal respiration fluxes poorly (Goulden et al ., 1996; Lavigne et al ., 1997; Loescher et al ., 2003) and do not allow for straight forward partitioning of ecosystem respiration to its component sources. Thus, direct measurements of respiration are necess ary for mechanistic prediction of forest canopy respiration. In tropical forests R is considered to be the largest component of autotrophic respiration (Chambers et al ., 2004; Malhi et al ., 2011; Malhi, 2012) and R is highest at the top of the canopy where leaves are exposed to full sun (Meir et al ., 2001; Cavaleri et al ., 2008). Therefore, to quantify the leaf respiratory carbon flux from tropical forests, fully exposed upper canopy leaves should be measured. Such data at the
42 species level are very scarce considering the diversity and importance of tropical forests, and the uncertainty associated with R flux estimates is consequently large (Malhi et al ., 2009). Furthermore, to calculate the respiratory carbon flux from tropical forest canopies both R at a s et temperature (e.g., R at 25C; R 25 ) and the short term temperature sensitivity of R need to be known. R roughly doubles with 10C temperature increase; i.e., it has a Q 10 of 2.0, although published values range from 1.1 to 4.2 (e.g., Azcn Bieto & Osmond 1983; Larigauderie & Krner, 1995; Tjoelker et al ., 2001). Available data from tropical forests suggest that R and Q 10 differ widely among species (Meir et al ., 2001; Slot et al ., 2013) and growth forms (Cavaleri et al ., 2008), but not enough data are c urrently available to identify systematic patterns in Q 10 among tropical forest trees and lianas The leaf economics spectrum ( I. J. Wright et al ., 2004) describes general correlations among photosynthesis (A max ) and R and leaf structural and chemical pr operties, including leaf mass per unit leaf area (LMA), leaf longevity, and concentration of nitrogen (N) and phosphorus (P). Because these structural and chemical traits are easier to determine reliably than photosynthesis and respiration, these trait cor relations are useful for estimating carbon (C) flux in species rich forest canopies. Further, in many global vegetation models R is defined as a fixed fraction of maximum photosynthesis (e.g., HyLand Levy et al ., 2004; IBIS, Foley et al ., 1996; LPJ, Sitc h et al ., 2003), or as a function of leaf N ( e.g., HYBRID, Friend et al ., 1997 ; NCAR LSM, Bonan et al ., 2003). What is less certain is whether correlations with other leaf traits can be identified for the temperature sensitivity of R in particular for the Q 10 which is widely used in C flux models.
43 Our ability to estimate R fluxes from tropical forests would greatly improve if the Q 10 could be captured by a simple trait based model in the way that R can be modeled from N or photosynthesis. Observational ev idence suggests that Q 10 values vary among species and with environmental conditions (Griffin et al ., 2002 b ; Turnbull et al ., 2003; Slot et al ., 2008). It remains unknown whether this variation is predictable. The factor that restricts the rate of R change s with temperature (Atkin & Tjoelker, 2003) At low temperature the respiratory capacity is most limiting (because enzymatic reaction rate s are low), while at high temperature R is more likely to be limited by the availability of respiratory substrate (pri marily carbohydrates), or the demand for respiratory product ( e.g., ATP : when there is low demand for ATP, the rate of respiratory electron transport becomes limited by the availability of ADP. Noguchi & Terashima, 1997). Two species with the same respirat ory capacity but different substrate availability or demand for respiratory products are likely to have different temperature response curves. Furthermore, the factor that controls R may differ among species in relation to their plant functional type (PFT) For example, R in fast growing, light demanding species is more likely to be limited by substrate availability, while slow growing, shade tolerant species are more likely to be limited by the demand for respiratory products (Noguchi & Terashima, 1997). I f species differ in the factor that controls R based on their respiratory substrate content and their PFT, it is likely that variation in Q 10 values can be explained by traits associated with respiratory control and life history. Here we set out to 1) quan tify R 25 and its temperature sensitivity for canopy leaves of a large number of common tropical forest trees and lianas, 2) statistically explain interspecific variation in R 25 and Q 10 in relation to leaf chemical and structural
44 traits, and 3) estimate the annual flux of R at the stand level in our study forest, using trait based models identified under the second objective. To achieve these goals we determined temperature response curves of respiration at the leaf level for 123 leaves of 28 species of tre e and liana from the upper canopy of a tropical forest in Panama, from which we calculated R 25 and Q 10 for each leaf We chose to use R 25 because it is widely used for comparison of respiration rates of plants from different biomes (e.g., Wright et al ., 20 06). We hypothesized that traits associated with the global leaf economics spectrum would also explain variation in respiration at the local scale. Methods and Materials Study S ite The study was conducted in Parque Natural Metropolitano (PNM, 859'N, 793 3'W, 100 m a.s.l.), a semi deciduous moist tropical forest near Panama City on the Pacific coast of the Republic of Panama Annual rainfall at the site averages 1740 mm, most of which falls during the rainy season (May through December). The park is a 256 hectare natural reserve consisting of 80 150 year old secondary forest with tree heights up to 40 m. A 42 m tall construction crane with a 51 m jib enables access to canopy leaves. Twenty eight species were selected from the upper forest canopy, representi ng a mix of plant functional types based on growth form and successional status; lianas (14 species), and trees classified as early successional (5 species), mid successional (7 species) and late successional (2 species) (Table 3 1). Together these 28 spec ies cover > 75% of the canopy area in reach of the crane (Avalos & Mulkey, 1999).
45 Respiration M easurements For each species three to five sun exposed terminal shoots were selected, where possible from multiple individuals. Twigs were collected pre dawn at ca 6 a.m., cut under water, or cut and immediately re cut under water, and brought back to the laboratory in darkness and stored cool until measured. Dark respiration was measured on 3 8 whole, fully mature leaves per species at 5 7 temperatures between 2 0 and 32C, in a Walz gas exchange cuvette with Peltier temperature control (GWK 3M, Walz, Mess und Regeltechnik, Eiffeltrich, Germany), connected to a LI 6252 infrared gas analyzer (Licor, Lincoln, NE, USA). Petioles were cut under water and sealed with Parafilm in a 5 ml glass vial with water to prevent the leaves from drying out during the measurement. These 5 7 respiration measurements per took ca. 60 90 minutes to complete. All leaves were measured within 10 hours of collection. No trend in respiratio n rates with time since collection was detected within this period. All measurements were made in the wet season, between late August and early October, when all selected species had mature but non senescing leaves. Functional T rait D ata Leaf area was meas ured with a LI 3000 leaf area meter (Licor), and leaves were dried at 60C to a constant mass. Leaf N was determined with an elemental analyzer (Costech Analytical, Los Angeles, California, USA). Concentrations of soluble carbohydrates (simple sugars and starch) were determined following Dubois (1956) with modifications. In short, simple sugars were extracted in 80% (v/v) ethanol by shaking at 27C, followed by 2 two hour incubations at 30C. For each sample the supernatant from the three incubations was c ombined in a volumetric flask and brought to 10 ml. Glucose concentrations were determined colorimetrically at 487 nm via the phenol
46 sulfuric acid method. Starch was hydrolyzed to glucose from the pellet in 1.1% hydrochloric acid at 100C. Starch concentr ations were determined as glucose equivalents. We measured photosynthetic capacity (A max ) on a separate set of 3 6 sun 2 s 1 400 ppm CO 2 at ambient temperature (range 26 29 C) w ith an LI 6400 (Licor) All photosynthesis measurements were taken between 9.00 and 10.00 a.m. before mid day stomatal closure was observed. Species means of foliar P were collected previously at the site, along with N, A max and LMA (S.J. Wright. U npublis hed). Species means for N, A max and LMA in this independent data set correlated strongly with the species means determined in the current study ( R 2 > 0.65 for each trait). Quantification of R and Q 10 For each leaf, a linear regression line was fit to the l og 10 transformed leaf R versus leaf temperature (T Leaf ) according to: log 10 R = a + bT Leaf ( 3 1 ) where a and b, respectively the intercept and the slope of the response curve, are leaf specific constants. Q 10 values were calculated from these equ ations as: Q 10 = 10 10b ( 3 2 ) Subsequently R 25 of each leaf was calculated as: ( 3 3 ) where R Tset is R measured at the set cuvette temperate ( T set ). We averaged R 25 over the 5 7 set cuvette temperatures to get a l eaf level R 25 to use in our analyses. Relationships of R 25 and Q 10 with other traits were examined with multiple regression models using species mean trait values. We developed regression models
47 for R 25 and Q 10 using leaf traits averaged at the species le vel rather than using the leaf level data. This is a conservative approach that avoids pseudo replication within species (as multiple leaves within a species are not independent data points). To identify the best combination of predictor variables we used the subset method of variable selection in multiple regression (Miller, 2002) from the Leaps package version 2.9 in R. In this method the best combination of predictors for a subset of n predictors is identified (using R 2 R 2 adjusted p (Mallo ws, 1973)) and this is repeated for all possible subset sizes. Estimating S tand level R espiration F luxes Our unpublished monitoring data at the study site show that leaf temperature (T Leaf ) is coupled closely to air temperature (T Air ), especially at night ( T Leaf = 0.999 T Air R 2 = 0.91; T Leaf of 5 tree species and 3 liana species monitored 5 8 days each). To estimate stand level nighttime R flux we could therefore use a 17 year air temperature record from the site (data collected at 25m height on the cra ne at 60 minute (1995 2005) or 15 minute (2006 2011) intervals. http://biogeodb.stri.si.edu/physical_monitoring /research/metpark ). We used four approaches to scale respiratio n of canopy leaves to the stand level, which differed in how estimates of R A and Q 10 were obtained and averaged across species. In model 1A, R A and Q 10 were estimated at the species level from their multiple regression relationships with other leaf traits. Model 2A, in contrast, used species averages of measured R A and Q 10 To assess the functionality of a PFT level flux prediction, both models were also run after substituting species level estimates with R A and Q 10 values modeled from leaf traits averaged by PFT (1B), or with measured R A and Q 10 values averaged by PFT (2B). In all models, CO 2 efflux from
48 R was calculated by species for every 15 or 60 minute interval between 6 p.m. and 6 a.m. Where available, we used species specific estimates of leaf area index (LAI; total leaf area per unit ground area, in m 2 m 2 ; five species in Kitajima et al ., 2005) to calculate total CO 2 flux per unit ground area. For the remaining species we used mean LAI per growth form from Clark et al (2008). Dark respiration in shade leaves is reduced compared to that of fully exposed upper canopy leaves, and we used site specific data to estimate the degree of reduction in respiration of shade leaves in relation to the LAI of the trees. Transmittance of photosynthetically active radiation (PAR) decreases exponentially with LAI in the canopy of five of our focal tree species (Kitajima et al ., 2005) such that 0.41 x LAI above ( 3 4 ) w here LAI above R decreases linearly with a decrease in % daily PAR such that % of R at full sun ( 3 5 ) (C. Rey Sn chez & M. Slot. U npublished). Combining equation 3 4 and equation 3 5 gives % of R at full sun + 78e 0.41xLAI above ( 3 6 ) L eaves in the second and third leaf layer of a tree canopy thus respire at 74% and 56% of leaves in the first (sun expose d) layer respectively. For models 1A and 1B, however, the decrease in R with depth in the canopy was based on the change in the leaf traits that were predictors in the trait based model of R To extrapolate to annual fluxes evergreen species were assumed to maintain LAI year round, deciduous species had LAI = 0 in the 4 month dry season from January
49 through April, and semi deciduous species maintained LAI for 10 months. For tropical forests both decreases and increases in dry season R have been reported (D omingues et al ., 2005; Miranda et al ., 2005; Stahl et al ., 2013), so because dry season data were not available R was kept constant throughout the year. R was scaled to the stand level by determining the relative abundance of the canopy trees from their ba sal area in 2010 census data [Census data from PNM are collected by Smithsonian Tropical Research Institute Forest Dynamics project ( Condit 1998; Hubbell et al ., 2005; Condit et al ., 2013)]. Because liana abundance was unknown at the species level, liana r elative abundance was equally divided among species, and summed up to cover 30% of the crown area (Avalos & Mulkey, 1999). For each calendar year species level R per hectare was summed up and multiplied by the molecular mass of C to get Mg C respired at ni ght ha 1 yr 1 We also estimated R during the day ( R Day ). Light inhibits R so R Light is lower than R Dark at a given temperature (Sharp et al ., 1984; Krmer 1995). We determined ( R Light ) for saplings of four of our focal species using the Kok method (Shar p et al ., 1984) and found an average reduction of respiration of 46% compared to R Dark at a given temperature (n = 17 leaves). Light also reduces the Q 10 but the extent of the reduction is variable (Atkin et al ., 2000; Pons & Welschen, 2003). For calculati on of R Day we reduced the Q 10 by 25% compared to the Q 10 of R Dark The C flux from R Day was calculated using day time temperature data and a 46% reduction in R and 25% reduction in Q 10 for all species. Statistical A nalyses Comparisons among species, plant functional types, and growth forms were made using one way ANOVAs and Tukey HSD post hoc tests.
50 Where necessary, data were transformed to improve normality and homoscedasticity. The variance in respiration and Q 10 data was broken down to variance at the species, PFT and growth form level using partial R 2 analysis. All statistical analyses were performed in R version 2.14.1 (R Development Core Team, 2011) Results Respiration at 25 C Respiration expressed on an area basis and standard ized at 25C ( R A ) differed significantly among species (Fig 3 1 ; Table 3 1 ), exhibiting 3 fold variation from the early successional tree Cecropia peltata (1.99 0.06 mol m 2 s 1 ; mean SD) to the liana Cissus erosa (0.60 0.15 mol m 2 s 1 ). R A of li anas (0.77 0.13 mol m 2 s 1 ) was lower than R A of trees (1.11 0.41 mol m 2 s 1 : t test P = 0.007). On average lianas had lower R A than early successional tree species, but other differences among PFTs were not significant. R 25 expressed on a mass ba sis ( R M ) also differed widely among species and among PFTs. Lianas and early successional tree species had significantly higher R M than late successional species, with mid successional species showing intermediate values that were not significantly differe nt from the other PFTs (Fig. 3 1) Q 10 V alues by S pecies, Plant F unctional T ype and G rowth F orm Q 10 values varied among species (mean 2.39, range 2.01 2.93), but did not differ systematically between trees (2.35 0.18, mean SD) and lianas (2.45 0.2 6) (Fig. 3 1). Different successional stages did not differ in mean Q 10 values either, with early successional (2.30 0.19), mid successional (2.36 0.19) and late successional species (2.26 0.03) all falling within a narrow range. For all PFTs Q 10 val ues were significantly higher than 2.0 (t test, P < 0.05 for all PFTs). For 19 of the 28 species Q 10 was also significantly greater than 2.0 at the species level.
51 Variance of R espiration and Q 10 Variation in R A and R M was considerable (see Fig. 3 1), with leaf level values ranging more than 6 fold, and species means ranging 3 fold. Most of the variance existed among species (Fig 3 2), but variation among leaves within species also explained 20 30% of the total variance. By comparison, Q 10 values were less variable (leaf level values ranged 2 fold and species means 1.5 fold). About 45% of the Q 10 variance was explained by variation in Q 10 values among leaves within species, with a similar percentage of variance explained by Q 10 differences among species with in PFT (Fig. 3 2). LMA, in contrast, differed significantly between trees and lianas (trees 106 19, lianas 68 15; P < 0.001. t test) and growth form explained most of the variance in this trait. Trait C orrelations and M ultiple R egression M odels for R e spiration and Q 10 Significant pair wise correlations were found between R A and area based A max N, P, total non structural carbohydrates (TNC Area ), and LMA (Table 3 2). To avoid overfitting of the model we chose to have maximally three predictors plus inte rcept in the model. The best three parameter model for R A contained the significant predictors P, A max and LMA: R A = 0.14 + (0.718 P Area ) + (0.042 A max ) (0.009 LMA) where R A and A max are in mol m 2 s 1 P Area is in mg m 2 and LMA in g m 2 This model accounted for 64% of the variance in R A in a subset of 24 species (12 tree, 12 liana species) for which data on all leaf traits were available (Table 3 fit is illustrated by the partial residual plots in Fig. 3 3. The same set of parameters, when expressed on a mass basis ( R M and A max in nmol g 1 s 1 P in mg g 1 ), constituted the best model for R M :
52 R M = 12.8 + (504 P) + (0.005 A max ) (0.18 LMA) Q 10 was best modeled from two independent predictors (Table 3 3): Lianas: Q 10 = 2.21 + (0. 021 TNC Area) T rees : Q 10 = 2.21 + (0. 0 21 TNC Area ) 0.281 where TNC Area is in mg m 2 Fig. 3 4 illustrates the combined effects of TNC Area and growth form on Q 10 Pair wise correlations were not significant for Q 10 (Table 3 2) and models w ith more than two predictors included non significant predictors There was a marginally significant interaction between TNC Area and growth form ( P = 0.083, ANCOVA). Stand level L eaf R espiratory C arbon F lux M ultiple regression models of R A explained more o f the variance in respiration than models for R M so we scaled R to the stand level using R A Mean annual flux of nocturnal respiration between 1995 and 2011 was 4.5 0.34 Mg C ha 1 (mean SD across years) when calculated using R A and Q 10 estimated from multiple regression models (model 1A), and 4.1 0.33 Mg C ha 1 when using leaf traits averaged by PFT (model 1B) (Fig. 3 5; Table 3 4). The annual flux estimate based on species specific R A and Q 10 measurements (model 2A) was 5.5 0.35 Mg C ha 1 and 5. 3 0.34 Mg C ha 1 when PFT level means were used. In all estimates, circa 95% of the respiratory carbon flux came from trees and 5% from lianas. Of the C flux from trees 4 7% (range across the four models) came from early successional species, 46 52% from mid successional species and 42 50% from late successional species. R Day was 2.9 0.13 Mg C ha 1 yr 1 when R was estimated from leaf phosphorus content, A max and LMA, and 2.7 0.11 Mg C ha 1 yr 1 when using PFT averages of these leaf traits (Table 3 4) When species level measurements of R A and Q 10 were
53 used R Day was 2.6 0.09 Mg C ha 1 yr 1 while the R Day flux estimated using PFT averages of measured R A and Q 10 was 3.4 0.11 Mg C ha 1 yr 1 (Table 3 4). Estimates of the total carbon flux from foliar respiration ranged from 6.7 (model 1B) to 8.9 Mg C ha 1 yr 1 (model 2A). Discussion In this study we determined leaf level dark respiration rates and Q 10 values of upper canopy leaves of tropical trees and lianas, and used these data to parameterize multi ple regression models, enabling us to scale leaf respiration to annual stand level carbon efflux from other leaf functional traits. Annual carbon fluxes associated with R were comparable to those reported for lowland tropical forests in the Brazilian Amazo n (Malhi et al ., 2009). Species and PFT D ifferences in R espiration T raits R A of lianas was lower than R A of trees (Fig. 3 1), which is the opposite of what Domingues et al (2007) and Cavaleri et al (2008) found in other Neotropical forests. The LMA of li anas in the current study was, however, significantly lower than the LMA of trees, which was not the case in the above mentioned studies. Liana leaves in the current study may thus simply have contained less metabolically active tissue per unit leaf area t han trees. R M of lianas wa s equal to that of early successional species, confirming that lianas have comparatively high metabolic activity. In accordance with recent in situ measurements ( Chapter 2 ) Q 10 values were significantly higher than 2.0, the common ly assumed value that has been adopted in ecosystem process models (e.g., Thornton et al ., 2002; Wang et al ., 2009). Moreover, for tropical climates Q 10 values even slightly lower than 2.0 have been predicted, based on the observation of declining Q 10 wi th rising temperature interval over which R is
54 measured (Tjoelker et al ., 2001; Atkin & Tjoelker, 2003). According to the linear function reported in Atkin & Tjoelker a Q 10 of 1.97 is predicted at 26C, the mid point of the temperature interval over which we measured R The temperature dependent Q 10 has been incorporated in global vegetation models and ecosystem models as an improvement to the static Q 10 (e.g., Wythers et al ., 2005, 2013; Atkin et al ., 2008; Ziehn et al ., 2011; Chen & Zhuang, 2013). While i t may be appropriate to expect the Q 10 to decrease in relation to temperature within a plant, or of in plants within a climatic region (as originally suggested by Tjoelker et al ., 2001), this observation cannot be extrapolated to predict Q 10 values of trop ical vegetation. The implication of doing so is that a low Q 10 is predicted for the tropics, and that is clearly not supported by our data. Especially if the model uses a reference temperature of respiration that is lower than the mean temperature at the s tudy site, underestimation of Q 10 values can result in considerable underestimation of the calculated respiratory carbon efflux from the forest. Multiple R egression of R and Q 10 We tried to identify trait based models to predict R A and Q 10 that could be o f use in modeling of carbon fluxes in tropical forests. Individual traits accounted for 21 47% of the explained variation in R A and the best three trait regression explained 64% of the variation. The best Q 10 model uses total non structural carbohydrate co ntent per unit leaf area and growth form to explain 26% of the variation in Q 10 while single trait correlations were non significant. While N and A max both correlate with R the fit is relatively poor across these co occurring tropical species. Interestin gly, P turned out to be a much better correlate of R Meir et al (2001) found that in a P limited forest in Jar, Brazil, P correlated more strongly with R than in a less P limited system in Cameroon. Cavaleri et al (2008)
55 working in a P limited lowland rainforest in Costa Rica therefore hypothesized P to be a better correlate of R than N, which is less limiting at the site. While R did correlate positively with P, the correlation was not stronger than the correlation of R with N, suggesting that P did no t limit R more than N did. Plant available P at PNM is on average 5.8 mg kg 1 (B L Turner, p ers c omm 1 ), which is relatively high for a tropical forest soil. Indeed, the P Area values in the current study are considerably higher than those in Meir et al (2001) and Cavaleri et al (2008). The greater strength of the correlation between R and P than between R and N is thus unlikely to reflect P limitation of this forest, but may instead reflect interspecific variation in non metabolic N use in leaves, obsc uring the correlation between R and N The best three parameter model with P, A max and LMA explained 64% of the variance in R A Meir et al (2001) used stepwise regression to model R A of leaves of tropical forest trees, and their best models also included P and LMA, but A max was not considered in their analysis. With our data the full model of R A as a function of P Area and LMA alone would be significant and have an R 2 of 0.52, but LMA would not be a significant predictor in this model. N and LMA together ex plained 50 80% of the variance in R across sites and biomes, including tropical forest (Reich et al ., 1998), but this study did not consider P and A max Although the observation that R correlates with P is not new, the fact that N was not included in our b est models suggests that perhaps especially in tropical forests, P should be considered as a predictor of R even if N is the better correlate across biomes ( I. J. Wright et al ., 2004). 1 Email correspondence on 11 October 2012. Averages based on unpublished data.
56 A maximum of 26% of the observed variance in Q 10 can be explained by TNC and growth form. When accounting for growth form, higher Q 10 values are found in species with higher of TNC per leaf area. Thus, R of species with high TNC increases more with temperature than species with low TNC, suggesting that R in species with low TNC becomes substrate limited at high temperature. Interestingly, however, substituting the concentration of simple sugars (the more immediate substrate for R ) for TNC in the model causes the trend to be lost (i.e., no increase in Q 10 with increase in the concentration of simple sugars). Furthermore, neither the concentration of simple sugars, nor TNC correlated significantly with Q 10 in bivariate correlations. These factors may argue against the above explanation of progressive substrate limitation at hig h temperatures. While important for R and potentially for the Q 10 of respiration, leaf sugar content is not a commonly measured trait in forest ecology. This calls into question the utility of a Q 10 model that requires TNC and growth form to account for 26 % of the interspecific variation in Q 10 when Q 10 values of 85% of the species fall within the relatively narrow range between 2.15 and 2.70 and are not systematically different among PFTs. Annual L eaf R espiratory C arbon F lux at the S tand level Estimates o f annual nocturnal leaf respiratory carbon release were considerably lower for trait based models that used multiple regression models to estimate R A and Q 10 than for models that used measured values of R A and Q 10 despite the fact that the regression mode ls were parameterized on the measured values. These differences are largely based on flux estimates for Anacardium excelsum the species with the highest relative abundance at the site, and the largest single contributor to the stand level C flux. Anacardi um has average A max and P concentration, but one of the highest LMA
57 values of all species (121 g m 2 ; Table 3 1). Because LMA has a negative effect on R A in the multiple regression model, R A of Anacardium and therefore of the study area, is underestimated The use of PFT means, either of measured or modeled respiration traits resulted in a small reduction in flux estimates compared to the use of species level data. In the following discussion we will use the flux calculations based on species level measure ment of R A and Q 10 to represent the best estimate of the respiratory carbon flux in the study forest. Per year, nighttime R of canopy trees and lianas at PNM releases 5.5 Mg C ha 1 This is almost identical to the nighttime R efflux of 5.6 Mg C ha 1 yr 1 t hat Malhi et al (2009 With R data from Domingues et al ., 2005), report for Tapajs in the Brazilian Amazon, but higher than the mean of 4.0 Mg C ha 1 reported for the Caxiuan reserve in the eastern Amazon (Metcalfe et al ., 2010). Although the dry season length at Tapajs is comparable to that at our site, the LAI estimates are higher (5.44 m 2 m 2 ) than the values we used (mean for trees 3.66, with 30% cover of lianas with LAI of 0.73 m 2 m 2 ). Furthermore, we conservatively assumed the dry season deciduou s species (11 of 14 tree species) to be leafless for the full four months of the dry season, whereas the Tapajs site shows increased metabolic activity during dry season leaf flush (Huete et al ., 2006). These calculations suggest that the nighttime R flux we calculated may be conservative, even though it is comparable to fluxes in other tropical forests. Leaf level respiration rates at PNM are at the high end of the spectrum of values reported for tropical forest trees (Slot et al ., 2013), which supports t he notion that the nocturnal leaf respiratory C flux we report may be a slight underestimate of the true value for the study site.
58 R Day added 3.5 Mg C ha 1 yr 1 bringing the total R flux to 8.9 C Mg ha 1 yr 1 The daytime flux was higher than reported in Malhi et al (2009), who assumed R Light to be 67% lower than R Dark based on measurements on Snow Gum ( Eucalyptus pauciflora Sieb. ex Spreng) (Atkin et al ., 2000). Our estimate of 46% reduction of R in the light for four tropical tree species was almost ide ntical to the 47% reduction Pons & Welschen (2003) reported for seedlings of Eperua grandiflora (Aubl.) Benth another tropical tree species. Not enough data on R Light are currently available to generalize the extent of reduction of R Light relative to R Da rk in tropical versus temperate species. Total C flux was comparable to that at Tapajs (7.4 4.0 Mg C ha 1 yr 1 ). For Caxiuan, the site where Melcalfe et al (2010) estimated nighttime R to be 4.0 Mg C ha 1 yr 1 Malhi et al (2009) report a total leaf respiratory carbon flux of 8.9 4.0 Mg C ha 1 yr 1 This suggests either a very high daytime R flux, or, more likely, it illustrates how discrepancy in estimation methods results in grossly different flux estimates. Near Manaus, at a site with a shorter d ry season than PNM the total annual flux is 10.1 4.0 Mg C ha 1 yr 1 (Chambers et al ., 2004; Malhi et al ., 2009). Due to the large uncertainty in the estimates, it is currently not possible to conclusively state forest differences in total respiration flu xes. Significance for M odeling Our results suggest that within a tropical forest canopy, estimating R from either N or photosynthesis alone leaves a lot of variation unexplained. While across biomes these correlations may work well, within a single biome, or indeed a single forest, the value of these bivariate correlations is limited. We also looked at whether it was feasible to link Q 10 to other leaf traits to facilitate future modeling efforts. The concentration of TNC and growth form together explained 26% of the variance in Q 10 but the physiological underpinning of this model is not clear and the explained variance is
59 modest. Rather than trying to model Q 10 from other leaf traits, it will be important to improve our understanding of the dynamic nature of the Q 10 in relationship to temperature, which currently appears to underestimate Q 10 values of tropical trees. Our estimates of stand level leaf respiratory carbon flux based on R A and Q 10 from trait based models were comparable to fluxes reported for o ther tropical forests, suggesting that trait based modeling indeed has potential. The parameterization of such trait based models will, however, need to be done locally, or regionally to assure accurate predictions.
60 Table 3 1 Species codes, names, famil ies, plant functional type (PFT. ES: e arly s uccessional, MS: m id s uccessional, LS: l ater s uccessional, L: l ianas), and the number of leaves measured (n), dark respiration at 25C per unit area ( R A ) and mass ( R M ), leaf mass per unit area (LMA), photosynthet ic capacity (A max ), and concentrations of n itrogen (N), p hosphorus (P) and total non structural carbohydrates (TNC). Code Species Family PFT n R A R M Q 10 LMA A max N P TNC mol m 2 s 1 nmol g 1 s 1 g m 2 mol m 2 s 1 % % mg g 1 ALBG Albizia guacha pele (Kunth) Harms Fabaceae ES 4 0.76 8.2 2.50 94 12.4 3.8 0.10 159 ANNS Annona spraguei Saff. Annonaceae ES 3 0.98 11.2 2.15 88 12.8 2.6 0.17 241 CECL Cecropia longipes L Urticaceae ES 3 1.82 19.5 2.20 94 20.6 2.5 0.20 155 CECP Cecropia peltata L. Urt icaceae ES 3 1.99 16.8 2.60 119 19.8 2.7 0.17 245 PITT Pittoniotis trichantha Griseb Rubiaceae ES 4 0.73 10.2 2.04 74 13.5 2.3 0.14 158 ASTG Astronium graveolens Jacq Anacardiaceae MS 3 1.27 12.9 2.15 98 12.6 2.4 0.20 160 CAS4 Castilla elastica (Lieb m.) C.C. Berg Moraceae MS 6 1.23 11.2 2.19 110 19.4 2.6 0.17 207 FIIS Ficus insipida Willd Moraceae MS 9 1.44 10.8 2.30 133 23.4 2.8 0.17 161 LUE Luehea seemannii Triana & Planch Tiliaceae MS 6 0.92 8.5 2.40 108 19.4 2.1 0.15 191 PSES Pseudobombax se ptenatum (Jacq.) Dugand Malvaceae MS 5 0.90 7.0 2.70 127 16.2 2.0 nd 188 SPOM Spondias mombin L. Anacardiaceae MS 4 0.93 12.1 2.48 78 16.5 2.6 0.13 109 ZUEL Zuelania guidonia (Sw.) Britt. & Millsp Salicaceae MS 3 0.91 8.1 2.30 114 17.3 2.0 nd 164 ANAE Anacardium excelsum Bertero & Balb. ex Kunth Skeels Anacardiaceae LS 8 0.97 8.1 2.24 121 13.8 1.8 0.14 166 CHRC Chrysophyllum cainito L Sapotaceae LS 4 0.63 5.1 2.28 124 17.1 1.9 0.10 177 AMPP Amphilophium paniculatum ( L .) kunth Bignoniaceae L 4 1.00 1 5.4 2.19 64 9.5 2.9 0.17 220 ARIC Aristolochia tonduzii O.C. Schmidt Aristolochiaceae L 3 0.91 15.6 2.39 62 12.1 3.3 0.15 239 BONT Bonamia trichantha Hallier f. Convolvulaceae L 7 0.73 10.0 2.93 74 10.4 2.3 0.18 266 CISE Cissus erosa Rich Vitaceae L 5 0.60 10.9 2.61 54 17.2 2.7 nd 210 COMF Combretum fruticosum (Loefl.) Stuntz Combretaceae L 5 0.80 12.0 2.75 69 17.7 2.8 0.16 196 GOUL Gouania lupuloides (L.) Urb. Rhamnaceae L 4 0.68 14.8 2.30 47 12.7 4.2 0.18 233 MIKL Mikania leiostachya Benth. Ast eraceae L 3 0.66 10.7 2.01 63 9.8 2.1 0.13 102
61 Table 3 1. Continued. Code Species Family PFT n R A R M Q 10 LMA A max N P TNC mol m 2 s 1 nmol g 1 s 1 g m 2 mol m 2 s 1 % % mg g 1 ODOM Odontadenia macrantha Willd. ex Roem & Schult Apocynaceae L 3 0.64 8.8 2.59 74 nd 2.1 0.09 128 PASV Passiflora vitifolia Kunth Passifloraceae L 3 0.73 11.2 2.37 62 6.7 3.4 0.20 189 PHRC Phryganocydia corymbosa ( Vent.) Bureau ex K Schum. Bignoniaceae L 3 0.76 6.9 2.53 110 12.2 3.2 0.15 94 PITC Pithecoctenium cr ucigerum (L.) A.H. Gentry Bignoniaceae L 3 0.84 11.7 2.51 72 13.4 2.9 0.14 167 SERM Serjania mexicana ( L .) willd. Malpighiaceae L 4 0.64 9.0 2.56 71 12.9 2.9 0.15 177 STIH Stigmaphyllon hypargyreum Triana & Planch. Bignoniaceae L 4 1.01 15.3 2.43 67 1 6.3 2.3 0.14 126 VITT Vitis tiliifolia Humb. & Bonpl. ex Roem. & Schult. Vitaceae L 7 0.83 14.2 2.22 59 13.0 2.4 0.13 158 nd: no data available
62 Table 3 2. Results from tests of pairwise correl ation between respiration at 25 C per unit l eaf area ( R A ), per unit mass ( R M ), and Q 10 and other traits: leaf phosphorus (P) and nitrogen (N) concentration, photosynthetic capacity (A max ), leaf mass per unit area (LMA), and total non structural carbohydrate concentration (TNC). R A and Q 10 are corre lated with area based leaf traits; R M with mass based traits. R A R M Q 10 R 2 P n R 2 P n R 2 P n P 0.47 < 0.001 24 0.23 0.018 24 0.0 ns 24 N 0.22 0.019 28 0.1 ns 28 0.0 ns 28 A max 0.33 <0.01 27 0.04 ns 27 0.0 ns 27 LMA 0.21 0.013 28 0.17 0.0 28 28 0.0 ns 28 TNC 0.26 <0.01 28 0.06 ns 28 0.1 ns 28 Growth form 0.25 <0.01 28 0.03 ns 28 0.1 ns 28 Table 3 3. Parameter estimates of multiple regression analysis of R A (mol CO 2 m 2 s 1 ), R M (nmol CO 2 g 1 s 1 ) and Q 10 against leaf phosphorus ( P), photosynthetic capacity (A max ), leaf mass per area (LMA: g m 2 ), TNC and growth form (Liana (L) vs. Tree (T)), where Response = Intercept + a P + b A max + c LMA + d TNC + e (Growth form). R A is regressed on area based leaf traits (P, TNC: mg m 2 A max : mo l CO 2 m 2 s 1 ); R M on mass based traits (P, TNC: mg g 1 A max : nmol g 1 s 1 ). In the Q 10 model TNC is expressed in g m 2 Number of species (n), overall (multiple) R 2 and model significance or also shown. Response Intercept P A max LMA TNC Growth form n R 2 P a b c d e R A 0.14 0.72** 0.042* 0.009* 24 0.64 <0.001 R M 12.8** 504* 0.005* 0.182** 24 0.56 <0.001 Q 10 2.21*** 0.021* L: 0 T: 0.28** 28 0.26 <0.05 *, ** and *** indicate significance of the variable with P < 0.05, P < 0.01 and P < 0.001.
63 Table 3 4. Distribution of stand level foliar respiration carbon flux (Mg C ha 1 yr 1 ) during night and day among plant functional types ( L ianas ES: e arly s uccessional, MS: m id s uccessional, LS: l ater s uccessional trees ). Estimate s are made using modeled and measured respiration ( R A ) and Q 10 at the species level and the level of the plant functional type (PFT). Trait based estimates of R A and Q 10 Measured R A and Q 10 Species means PFT means Species means PFT means Night Da y Night Day Night Day Night Day Lianas 0.26 0.17 0.26 0.17 0.25 0.16 0.26 0.16 ES trees 0.17 0.11 0.20 0.13 0.18 0.11 0.28 0.18 MS trees 2.20 1.46 2.06 1.37 2.42 1.54 2.59 1.65 LS trees 1.83 1.20 1.53 1.01 2.61 1.64 2.16 1.36 S um 4.46 2 .94 4.05 2.68 5.46 3.45 5.29 3.35
64 Figure 3 1. Variation of respiration traits within and among species (species codes as in Table 1); within and among plant functional types (PFT), and within and between growth forms. A) Respiration per unit leaf a rea ( R A ). B) Respiration per unit leaf mass ( R M ). C) Temperature sensitivity of respiration ( Q 10 ) The overall means are indicated by the dashed lines. T he gray line in C indicates Q 10 = 2.0, the value widely used in C flux models. The tree species are earl y successional (ES, 5 species), mid successional (MS, 7 species), and late successional (LS, 2 species). The box plots indicate the median, 25th and 75th percentile for each species, PFT and growth form. Whiskers extend to 1.5 times the interquartile range Boxplots for PFT and growth form are calculated from species means. Different letters indicate groups that are significantly different from one another ( P < 0.05) (One way ANOVA with Tukey post hoc testing).
65 Figure 3 2. Proportion of variance in Q 10 R A (mol m 2 s 1 ), R M (nmol g 1 s 1 ) and LMA (g m 2 ) explained by variance within species, among species within plant functional type (PFT), among PFTs and between growth forms, as determined by partial R 2 analysis. Figure 3 3. Partial residual plots for the best model of respiration per unit leaf area ( R A ) in which R A is regressed against phosphorus content per unit area (P Area ), photosynthetic capacity (A max ), and leaf mass per unit area (LMA). These plots show the predictor variable on the x axis, and on the y axis the residuals of the full model plus the partial regression coefficient of the predictor variable multiplied by the predictor variable. A) Phosphorus concentration per unit leaf area (P). B) Photosynthetic capacity (A max ). C) Leaf mass p er unit area. Trees and lianas are combined in these analyses as growth form was not a significant predictor
66 Figure 3 4. Partial residual plots for the best model of Q 10 in which Q 10 is regressed against TNC Area and growth form. A) The residuals of th e full model + the regression coefficient of TNC Area multiplied by TNC Area are together plotted against TNC Area B) T he sum of the full model residuals plus the regression coefficient of growth form multiplied by growth form is regressed against growth for m on the x axis. Figure 3 5. Proportional contribution to the 17 year mean total annual leaf respiratory carbon flux by lianas, early successional (ES), mid successional (MS) and late successional (LS) tree species. R A and Q 10 were either modeled from leaf traits (models 1A,B. see main text) or measured values were used (models 2A,B).
67 CHAPTER 4 THERMAL ACCLIMATION OF LEAF DARK RESPIRATION TO EXPERIMENTAL NIGHTTIME WARMING IN TROPICAL CANOPY TREES AND LIANAS Background Contemporary tropical forests e xist in a narrow temperature range (Janzen, 1967; Wright et al ., 2009), close to what may be a high temperature threshold (Doughty & Goulden, 2008). Models predict unprecedented warming in the tropics over the current century (Diffenbaugh & Scherer, 2011), which will push the majority of tropical forests into a climate envelope currently not occupied by closed canopy forests (Wright et al ., 2009). The capacity of organisms to acclimate to a change in temperature is less likely to be under natural selection in thermally stable conditions such as tropical forests than in temperate and boreal biomes (Janzen, 1967; Cunningham & Read, 2002; Ghalambor et al ., 2006). Tropical forests are currently considered to be an important carbon sink (Phillips et al ., 1998; B aker et al ., 2004; Lewis et al ., 2009), but how they will respond to climate warming is uncertain, as illustrated by the large discrepancy in predictions among dynamic global vegetation models (DGVMs) and earth system models (Ahlstrm et al ., 2012; Cox et al ., 2013). The physiological response of vegetation to temperature and CO 2 are the largest source of uncertainty in such models (Arneth et al ., 2012; Booth et al ., 2012 ; Huntingford et al ., 2013 ), in part because thermal acclimation of photosynthesis and respiration of the vegetation is only addressed by some models, but not by others (Smith & Dukes, 2013). More importantly, the current lack of empirical data on thermal acclimation of tropical forest species hinders the improvement of the models and the li kelihood of achieving predictions for tropical forests that are consistent across models. Leaf dark respiration (non photorespiratory mitochondrial respiration) is highly sensitive to changes in temperature, generally doubling with a 10C rise in
68 temperatu re (i.e., Q 10 = 2.0) (e.g., Amthor, 1984). Photosynthesis also increases with temperature, but peaks at a lower temperature than respiration (Hve et al ., 2011). Consequently, without perfect physiological acclimation, further temperature rise in tropical forests may reduce photosynthesis (Doughty, 2011) while increasing respiration, thus reducing NPP and the size of the potential carbon sink of tropical forests. Respiration is essential to growth and survival of plants, as it provides the energy and carbon skeletons for biosynthesis (Penning de Vries, 1975). However, the increase in respiration with temperature is primarily attributable to increased demand for cellular maintenance (Amthor, 1984; Ryan, 1991), and is not associated with increased growth. Resp iration associated with growth is only indirectly affected by warming; it increases with warming only when growth itself is stimulated by rising temperature (Franz et al ., 2004). In tropical species, however, warming generally results in a decrease in grow th (see meta analyses by Lin et al ., 2010; Way & Oren, 2010; but see Cheesman & Winter, 2013). This suggests that unless respiration can sufficiently acclimate to warmer temperatures, carbon available for growth, and the potential of tropical forests to st ore carbon may diminish under future climate scenarios. Acclimation of respiration to elevated temperature is characterized by a decreased rate at the new temperature compared to non acclimated plants (Atkin & Tjoelker, 2003; Atkin et al ., 2005. See Fig. 4 1). A reduction in respiration may also occur when warming leads to more rapid depletion of respiratory substrate (primarily simple sugars and starch), but substrate limitation does not constitute acclimation. Net photosynthesis of tropical trees and lian as has been shown to decrease with warming a few degrees above current ambient temperatures (Doughty, 2011), suggesting limited capacity for physiological acclimation of tropical species. Whether
69 respiration in tropical tree and liana species can acclimate to elevated nighttime temperatures is currently unknown, but observations of reduced diameter increment of trees in a tropical lowland forest during years with above average nighttime temperatures (Clark et al ., 2003, 2010) suggest that acclimation may be incomplete at the very best. The main objective of this study was to determine whether dark respiration of fully expanded leaves of tropical canopy trees and lianas can acclimate to elevated nighttime temperature. We further asked whether along with chan ges in respiration, warming results in changes in other leaf functional traits. Such parallel changes could provide indications of the mechanism s underlying the acclimation process. For example, several studies on temperate and boreal tree species have sug gested an important role for leaf nitrogen and carbohydrates in thermal acclimation of respiration (Lee et al ., 2005; Tjoelker et al ., 2008, 2009). Furthermore, interspecific variation in changes of leaf traits with warming could potentially become valuabl e correlates or predictors of thermal acclimation of respiration Finally, we aimed to calculate the effect of acclimation on whole canopy leaf respiratory carbon release under elevated nighttime temperature regimes. We addressed these objectives by exper imentally warming branch segments of trees and lianas in the canopy of a semi deciduous tropical forest in Panama. The result of the warming treatment varied from leaf to leaf, such that nighttime leaf temperatures ranged from ambient to ca 8C above ambi ent. We related average nighttime leaf temperature to the rate of leaf dark respiration at 25 C, and to the short term temperature response of respiration. We chose 25C as a set temperature because it is widely used for comparison of respiration rates of plants from different biomes (e.g., Wright et al ., 2006) and it is close to the current mean nighttime
70 temperature at the study site Studies with plants from arctic, boreal, temperate and alpine climates have shown that respiration of many species accli mates to changes in ambient temperature (Billings et al ., 1971; Larigauderie & Krner, 1995; Collier, 1996; Arnone & Krner, 1997; Tjoelker et al ., 1999a,b, 2008,2009; Atkin et al ., 2000; Bolstad et al ., 2003; Lee et al ., 2005; Xu & Griffin, 2006; Xu et al ., 2007; Bruhn et al ., 2007; Ow et al ., 2008a,b, 2010). Tropical forests experience minimal seasonal temperature changes (Wright et al ., 2009), and because species have spent the past 2.6 million years of the Quaternary in conditions that were cooler than current tropical forests experience, it has been speculated that heat tolerance and the capacity to acclimate to elevated temperature may have been lost (Corlett, 2011). Our working hypothesis therefore is that respiration of these fully expanded mature le aves will not acclimation to nighttime warming and that warming consequently will lead to increased respiratory carbon release. Materials and Methods Study S ite and Species selection The study was conducted in Parque Natural Metropolitano, (PNM, 859'N, 7 933'W, 100 m a.s.l.) a seasonal tropical forest near the Pacific coast of the Republic of Panama, near Panama City. Annual rainfall at the site averages 1740 mm, most of which falls during the rainy season from May through December. Annual mean nighttime temperature at PNM between 1995 and 2012 was 24.5C (range 23.3 26.1C). This 256 hectare natural reserve consists of 80 150 y ea r old secondary forest with tree heights up to 40 m. A 42 m tall construction crane with a 51 m long jib enables access to canop y leaves. We selected three tree species ( Anacardium excelsum (Bertero & Balb. ex Kunth) Skeels; Luehea seemannii Triana & Planch.; Castilla elastica var.costaricana (Liebm.) C C. Berg) and two liana species ( Bonamia trichantha Hallier f;
71 Stigmaphyllon lin denianum A. Juss.) from the upper forest canopy. Henceforth the species will be referred to by their genus name only. Together these species contribute > 25% of the total canopy area (Avalos & Mulkey, 1999). In S itu W arming P rotocol Nighttime warming of te rminal shoots (or individual leaves on them ) was achieved by infrared reflective frames fitted with flexible heat rope (Big Apple Herpetological, Inc., New York, USA), positioned 5 10 cm below the target leaves (Fig. 4 2 ) The design was adjusted to the ar chitecture of the species; it was flat for most species with horizontal leaf display, but cone shaped for Anacardium to account on vertically oriented terminal branches Identical frames without heat rope were fit on control shoots The heat rope temperature was controlled by a thermostat, which triggered warming when air temperature dropped below 25C. This method resulted in warming of leaf temperature only during night by an average of 2 4C compared to the cont rol leaves. In total 154 leaves were included in the experiment, of which 67 were successfully warmed (average warming > 1C, and no warming > 10C at any time) and 87 were used as controls. To assess the repeatability and consistency of the results we set up warming and control frames twice in Anacardium and Luehea. These repeated experiments were done on different branches and several weeks apart. Temperatures of warmed and control leaves were monitored with type T copper constantan thermocouple wires att ached to the abaxial side of the leaf, and recorded at 5 minute intervals with a Campbell 21X datalogger (Campbell Scientific, Logan, Utah, USA). Dark R espiration M easurements After 6 8 days of treatment, t wigs were collected pre dawn at ca 6 a.m., immed iately re cut under water, and brought back to the laboratory in darkness for
72 measurements. Dark respiration was measured on whole leaves at 2 5 (mode = 3) temperature points between 20 and 32C with a Walz gas exchange cuvette ( GWK 3M Walz Mess und Rege ltechnik, Eiffeltrich, Germany) connected to a LI 6252 infrared gas analyzer (Licor, Lincoln, Nebraska, USA). Petioles were cut under water and sealed in a 5 ml glass vial with P arafilm to protect the leaves against dehydration during measurement. For each leaf that was measured at 3 or more temperatures a least square regression line was fit to the log 10 transformed leaf respiration rate ( R ) versus leaf temperature (T Leaf ) data according to: log 10 ( R ) = a + bT leaf ( 4 1 ) where a and b, respectively the intercept and the slope of the response curve, are leaf specific constants. Q 10 values were calculated from these equations as: Q 10 = 10 10b ( 4 2 ) When R was measured at only two temperatures, Q 10 was calculated as: ( 4 3 ) where T1 and T2 are the lower and higher measurement temperatures respectively. Subsequently respiration rate at 25 C ( R 25 ) of each leaf was calculated as: ( 4 4 ) where R Ts et is R Leaf measured at the set cuvette temperate ( T Set ). We averaged R 25 over the 2 5 set cuvette temperatures to get a leaf level R 25 to use in our analyses. All measurements were made in the wet season, between June and early October, when all selected species had mature but non senescing leaves. Functional T rait D ata Leaf area was measured with a LI 3000 leaf area meter (Licor) and leaf mass per unit area (LMA) was calculated as leaf mass (excluding petiole) after drying for >
73 96 hours at 60C divided by leaf area Tissue N concentration was determined with an elemental analyzer (Costech Analytical, Los Angeles, California, USA). Concentrations of non structural carbohydrates ( soluble sugars and starch) were determined following Dubois (1956) with mod ifications. Briefly, simple sugars (monosaccharides) were extracted in 80% (v/v) ethanol from 10 15 mg ground sample by shaking followed by 2 two hour incubations. The supernatant from each sample was collected in a volumetric flask and brought up to 10 ml Starch was hydrolyzed to glucose from the pellet in 1.1% hydrochloric acid at 100C. Glucose concentrations were determined colorimetrically with the phenol sulfuric acid method. We measured photosynthetic capacity (A max ) after 7 nights of warming in o ne warming experiment of Anacardium (16 leaves) and one Luehea experiment (12 leaves). A max was measured at ambient temperature (range 29 31C) and saturating 2 s 1 with an LI 6400 (Licor). The CO 2 concentration during mea surement s was maintained at 400 ppm using the built in CO 2 regulator and relative humidity was kept between 65 and 85%. All photosynthesis measurements were taken before 9.00 a.m. to avoid mid day stomatal closure, which can occur as early as 10 a.m. ( Zot z et al ., 1995; M. Slot, pers. obs. ). Data A nalysis To assess acclimation we calculated acclimation ratios using the set temperature method (Loveys et al 2003) as: ( 4 5 )
74 If acclimation has taken pla ce Acclim SetTemp values are > 1.0. To determine to what extent acclimation approached complete homeosta sis of respiration across temperatures we used the homeostasis method (Loveys et al 2003): ( 4 6 ) where T Warm and T Control are the mean nighttime leaf temperatures experienced by warmed and control leaves respectively. When acclimation is completely homeostatic, Acclim Homeo is 1.0; values less than 1.0 indicate that acclimation of the warmed leaf is not completely homeostatic. Acclimation ratios are particularly useful in controlled experiments where temperature variation within a treatment is small. However, a simple dichotomous comparison of warmed leaves and controlled leaves wou ld not account for the considerable variation in temperature within both warmed and control leaves in our experiments We therefore also assessed acclimation by regressing R 25 against the average nighttime leaf temperature over the week preceding the R 25 m greater acclimation. Acclimation slopes and ratios were calculated by species, and by experiment within species for the two Anacardium and Luehea experiments. To calculate the average % decline in R 25 per degree of nighttime leaf warming across species R 25 of every leaf was divided by R 25 at the average T C ontrol for each species Treatment and species effects were analyzed with analysis of variance; temperature effects on leaf traits were calculate d as least square regression. All statistical analyses were performed in R version 2.14.1 (R Development Core Team, 2011). Estimating S tand level R espiration F luxes To assess the effect of acclimation of respiration to nighttime temperature on the annual n octurnal leaf respiratory carbon (C) flux at the stand level, we compared
75 different scenarios of warming and thermal acclimation with the leaf respiratory C flux under current climate. First we estimate the current flux. We have previously collected R 25 an d Q 10 data for 28 species that constitute the canopy at our study site ( Chapter 3 ). We combined this dataset with a 1 7 year temperature record from PNM ( http://biogeodb.stri.si .edu/physical_monitoring/research/metpark ) to calculate the leaf respiratory C flux under current temperature conditions In short, we calculated respiration for each species based on 15 or 60 minute mean temperature and integrated that over 12 hour nigh ts assuming a 4 month dry season during which deciduous species are leafless and semi deciduous species are leafless for 2 months. We let shade leaves respire at 50% the rate of sun leaves and assumed that s pecies with a leaf area index (LAI: leaf area pe r unit ground area in m 2 m 2 ) > 1.0 had LAI minus 1 l ayers of shade leaves. Where available, we used species specific estimates of LAI (Kitajima et al ., 2005) to calculate total flux per unit ground area. For the remaining species we used mean LAI per grow th form from Clark et al (2008). Respiration was scaled to the stand level by determining the relative abundance of the canopy trees from their basal area in 2010 census data [Census data from PNM are collected by Smithsonian Tropical Research Institute F orest Dynamics project (Hubbell et al ., 2005; Condit 1998; Condit et al ., 2013)]. We re ran the calculations for elevated nighttime temperature (current + 4C) to estimate stand level respiratory C efflux without acclimation. 4C was chosen to represent e nd of century predictions for tropical South America (predicted rise of 2 .5 4.7C between 2000 and 2100 ; Cramer et al 2004) Next we accounted for acclimation by reducing R 25 by 10% according to findings in the current study, and estimated the effect on C fluxes. Finally, we let R 25 acclimate to the running average of nighttime temperature of the preceding week, and again total C efflux was
76 calculated. We l et al l canopy species acclimate equally, as our results (below) did not suggest acclimation differe nces among species and growth forms studied. Results Warming E ffect on R espiration The warming frames increased leaf temperature at night by an average of 2 4C (Fig. 4 2 ). Average nighttime leaf temperature and species identity both had a significant eff ect on R 25 (AN C OVA, P < 0.01), but their interaction was not significant. Leaves that had experienced warmer nights had lower respiration rates at 25C ( R 25 ). Th e qualitative pattern was consistent across species, and across repeated experiments within spe cies (Fig. 4 3 ). The intra specific variation in R 25 that was not explained by leaf temperature was considerable, and when analyzing the effect of nighttime leaf temperature at the species level, it was only significant for the two experiments in Anacardiu m Some of the leaf to leaf variation in R 25 was attributable to differences in leaf nitrogen content. Consequently, the temperature dependence of R 25 per unit leaf N ( R 25 /N) contained less intra specific scatter (Fig. 4 3 ). All experiments showed a signif icant decline of R 25 /N with leaf temperature ( P < 0.05) except for the liana Stigmaphyllon ( P = 0.07) and one of the Luehea experiments ( P = 0.09). Across species R 25 decreased by an average of 2.9% per degree of warming (Fig. 4 4 ), and R 25 /N decreased by 4.2% per degree of warming. Acclim SetTemp values for R 25 were consistently > 1.0 and very similar among species, with values ranging from 1.0 4 to 1.14 (Table 4 1). Acclim SetTemp values for R 25 /N were a little higher, with an overall mean of 1.1 7 Acclimat ion did not result in complete homeosta s i s of respiration; respiration of warmed leaves measured at their average nighttime temperature was higher than respiration of control leaves measured at the average nighttime temperature of control leaves ( Acclim Hom eo ratios between 0.7 3 and 0.9 4 ; Table 4 1). Even for Castilla which exhibited the steepest
77 decline in R 25 with nighttime temperature (Fig 4 3 ), respiration rates at their average nighttime T Leaf of 28.8C were 10% higher than respiration rates of contro l leaves at 26.8C (Fig. 4 1 ). Q 10 values rang ed from 2.5 to 3.0, and did not differ systematically between trees and lianas (Table 4 1). Q 10 was not affected by nighttime temperature. Warming E ffects on O ther L eaf T raits Nighttime leaf warming resulted i n a significant increase in nitrogen per unit area (N Area ; +1.9% per C relative to the mean of the control leaves; P = 0.002), sugar per unit area (Sugar Area ; +2.0% per C; P = 0.019), and s ugar to s tarch ratio (SS r atio; 2.9% per C; P = 0.008). The incr eases in area based N (N Area ) and simple s ugars (Sugar Area ) were caused by marginally significant increases in both LMA and concentrations of N and Sugars per unit mass The increase in N Area with temperature contributed to the increased significance of th e temperature response of R 25 when expressed on a per unit N basis. Photosynthetic capacity was not affected by nighttime warming in either Anacardium or Luehea (t test of control versus warming P > 0.05; no correlation with nighttime temperature at the le af level). Across species C orrelates of A cclimation The effect of nighttime leaf temperature on R 25 was consistent across species with little difference in the degree of acclimation among species. We explored whether these small differences among species correlated with leaf traits that are associated with plant metabolism, using trait means of control leaves to characterize the species. The acclimation slopes of the regression of R 25 on the mean nighttime leaf temperature over the week preceding the meas urements correlated strongly with species means of SS r atio ( P = 0.050; R 2 = 0.77; Fig. 4 5). Small differences among species in the average leaf temperature during the week prior to respiration measurements also had a (marginally significant) effect on th e extent of acclimation,
78 with Castilla which had the highest mean nighttime temperature showing the strongest acclimation, and Stigmaphyllon experiencing the coolest nights showing the smallest acclimation after Bonamia (Fig. 4 5). Acclimation slopes did not correlate significantly with leaf N content, LMA or species means of R 25 (data not shown). Consequence of A cclimation for S tand level R espiration F luxes We estimate d the leaf respiratory nighttime carbon efflux to be 4.1 Mg C ha 1 yr 1 for this site ( Fig. 4 6). Increasing nighttime temperature by 4C, not accounting for acclimation, increase d the flux to 5.8 Mg C ha 1 yr 1 an increase of 41%. Acclimation in which R 25 was dependent on the average nighttime leaf temperature of the preceding week (based on data shown in Fig. 4 4) resulted in an annual nighttime C flux of 4.9 Mg C ha 1 This represents a 16% reduction compared to 4C warming without acclimation The flux of warm acclimated leaves under 4C nighttime warming is still 19 % higher than the cur rent flux highlighting the non homeostatic nature of respiration in this system. Discussion Consistent A cclimation of R espiration to E levated N ighttime T emperature Respiration acclimated to higher nighttime temperature, as illustrated by the Acclim SetTemp values greater than 1.0 and by the consistent negative correlation between nighttime temperature and R 25 Respiration expressed per unit leaf nitrogen acclimated even more strongly. We warm ed leaves in multiple common tree and liana species to assess the generality of the acclimation response and to obtain information useful for scaling carbon fluxes associated with leaf respiration up to the canopy level. The patterns we found were indeed similar among the experiments done in trees and lianas of different species. We can therefore interpret our observations as indicative of thermal acclimation occurring in trees and lianas of the common species in the upper forest canopy at PNM.
79 These results confirm the observations on arctic, boreal and temperate species that thermal acclimation of respiration of pre existing leaves can be rapid (Billings et al 1971; Atkin et al 2000; Bolstad et al 2003; Lee et al 2005). Despite the thermal stability of the biome (Wright et al 2009), tropical species have the ca pacity to thermally acclimate. I n tropical tree species acclimation of photosynthesis to elevated daytime temperature has been found to be limited (Cunningham & Read 2002, 2003 a ; Doughty 2011) The fact that respiration does acclimate highlights the impo rtance of maintaining optimal respiratory functioning under changing temperatures, while minimizing carbon loss from maintenance respiration. This is also illustrated by frequent observations in other biomes that respiration acclimates better to temperatur e than photosynthesis (e.g. Campbell et al ., 2007; Ow et al ., 2008a b, 2010). Correlates of A cclimation Carbohydrate concentration was not lower in warmed leaves than in control leaves, nor did it decrease with average nighttime leaf temperature. Thus, o ur results are not an artifact of measuring respiration of warmed leaves that have become substrate limited. In fact, expressed per unit leaf area, the concentration simple sugars, the primary substrate of respiration, actually increased with temperature, as did the SS ratio. Nighttime warming can stimulate photosynthesis (Turnbull et al ., 2002, 2004), which could in turn increase leaf sugar content, but photosynthetic capacity was not affected by warming. Warming can also restrict translocation of sugars w hen sieve plate pores get blocked by callose (McNairn & Currier, 1968). However, translocation blockage would also increase Sugar Mass but this was not significant.
80 The absence of correlation between R 25 and Sugar suggest that R 25 of these leaves is not s ubstrate limited. Nor does it seem likely that there were treatment differences in demand for respiratory products through changes in sink strength away from the source leaves, as warmed and control leaves always came from the same plant. Most likely the d ecrease in R 25 with temperature resulted from a change in respiratory capacity, associated with changes in concentrations or relative amounts of mitochondrial enzymes (Atkin et al ., 2005). The change in N Area with temperature may be associated with shifts in relative amounts of enzymes that differ in N content but it is unclear what the functional significance of increased Sugar and SS ratio is with regards to thermal acclimation of respiration. Leaf nitrogen and carbohydrates have repeatedly been associat ed with respiratory acclimation (Lee et al ., 2005; Tjoelker et al ., 2008, 2009). Interestingly, the species in which carbohydrates and nitrogen are associated with acclimation are mid to high latitude tree species that routinely experience low temperature s. Acclimation to low temperature results in elevated respiration rates compared to non acclimated plants (Atkin & Tjoelker, 2003). At the same time acclimation to low temperature tends to involve an increase in leaf sugar concentrations associated with cr yoprotection (Kozlowski, 1992; Strand, 2003). It seems therefore unclear whether the correlation of carbohydrates with respiration across acclimation temperatures is related to respiratory acclimation per se or represents the composite of multiple acclimat ion processes including cold hardening. Leaf nitrogen concentration is also higher in cold grown plants (Weih & Karlsson, 2001; Tjoelker et al ., 1999b, 2008,2009; Lee et al ., 2005), possibly in conjunction with elevated protein content to compensate for in hibition of photosynthesis at low temperature (Pyl et al ., 2012). Interestingly, expressed per unit area, nitrogen content increased in warmed leaves
81 in our study. As a result, R 25 /N decreased more strongly with temperature than R 25 per unit leaf area did. The underlying mechanism and the functional importance of this are not clear. Acclimation to elevated temperature is likely to involve different processes than acclimation to low temperatures, and consequently the traits that correlate with the observed a cclimation are likely to differ as well. Acclimation in both T rees and L ianas Acclimation was consistent among species, and there was no evidence that trees and lianas differed in their accl i matory capacity. Doughty (2011) found that A max of canopy tree le aves was more negatively affected by in situ warming than A max of liana leaves, suggesting lianas had a greater capacity for acclimation, but a mechanistic explanation for this observation was not identified. Boreal evergreen tree species acclimate more co mpletely than broadleaved species (Tjoelker et al 1999a,b), which fits the general assumption that species that experience large intra annual temperature variations have greater acclimation capacity than species that do not. Studies using temperate speci es have not found systematic differences in acclimation capacity among different plant functional types (Campbell et al 2007) or between inherently slow and fast growing species (Loveys et al 2003). A more complete assessment of thermal acclimation of tropical trees and lianas will be necessary to verify the consistency of our results of comparable acclimation capacity in these growth forms. Across species acclimation correlated positively with the sugar to starch ratio of leaves, but not with nitrogen or total non structural carbohydrate (TNC) concentrations. Similarly, Loveys et al (2003) found no relationship between concentrations of TNC and nitrogen and the capacity of species to thermally acclimate. The liana leaves in our experiment had significa ntly lower concentrations
82 of simple sugars and higher concentrations of starch, and consequently much lower sugar to starch ratios. The correlation between sugar to starch ratio and the acclimation slopes of leaves from different species was, however, not solely driven by the lianas (Fig. 4 5). Consequences of A cclimation for P r edicted R espiratory C arbon F luxes from T ropical F orests We let R 25 of all canopy species acclimate to the same extent according to our observations from a series of 6 8 day warming experiments. The estimated effect of acclimation under 4C nighttime warming was considerable at our study site; a reduction of nocturnal leaf respiratory carbon release of 0.9 Mg ha 1 yr 1 (16%) compared to a no acclimation scenario Because acclimation d id not result in maintenance of homeosta tic respiration rates the calculated flux was still 0.8 Mg C ha 1 yr 1 (19%) larger than the flux at current temperatures. However, acclimation of leaves developed at a new temperature tends to be greater than accli mation of pre existing leaves (Loveys et al 2003; Armstrong et al 2006). As nights gradually warm over the coming decades, respiration is likely to acclimate to a greater degree than what we observed for pre existing leaves, which would further mitigat e the increase in carbon release from nighttime respiration. Significance for M odeling It has long been recognized that acclimation to temperature should be considered in global vegetation models, as it can reduce the magnitude of the positive feedback be tween climate and the carbon cycle in a warming world (King et al 2006). Many DGVMs and ecosystem models still do not address acclimation, however, and many that do, use a temperature dependence of Q 10 to represent acclimation (see for reviews of such mo dels Wythers et al 2005; Smith & Dukes 201 3 ). The use of a temperature dependent Q 10 is based on the observation that Q 10
83 values decrease with increasing temperature of measurement (Tjoelker et al 2001; Atkin & Tjoelker 2003). A temperature dependent Q 10 can contribute to success simulati on of carbon fluxes in highly seasonal biomes (Chen & Zhuang 2013; Wythers et al 2013). However, implementing a temperature dependent Q 10 in a global model may result in low Q 10 values being assigned to vegetation in tropical climates. In the current study Q 10 values ranged from 2.5 to 3.0 for the temperature interval 22 32C (midpoint 26C) and Q 10 did not decrease with warming. These values are much higher than what would be predicted for 26C from the temperature dependent Q 10 model (Q 10 = 3.09 0.043T (Atkin & Tjoelker 2003). Q 10 = 1.97 at T = 26C), so for calculation of daily respiratory carbon fluxes the assumption of a declining Q 10 would result in erroneous estimates. The use of a temperature dependent Q 10 in global models require s careful consideration of the timescale over which the Q 10 responds to temperature. Considering acclimation of R 25 to the temperature of the previous days or nights offers a more realistic approach, and is now supported by empiric al data from tropical forests (this study) as well as from other biomes. Concluding R emarks Here we showed that mature leaves of tropical trees and lianas can acclimate to elevated nighttime temperatures in 6 8 days by down regulating R 25 independent of a vailability of respiratory substrate. Despite the short duration of the warming experiments the nature of the acclimation response suggested that the respiratory capacity of leaves was adjusted. Implementation of these acclimation responses int o a simulati on of stand level nocturnal leaf respiration under 4C warming reveals the potential to compensate simulated nighttime warming, a level of warming predicted for the end of the 21 st century. Global models could be improved by accounting for
84 temperature accl imation of leaf respiration. Assuming respiration at a set temperature acclimates to the temperature of the previous 6 8 nights is preferable to assuming a temperature dependent Q 10
85 Table 4 1. The tree and liana species used in this study; their mean res piration rates at 25C ( R 25 ), R 25 per unit nitrogen ( R 25 /N) and Q 10 Acclim SetTemp Acclim Homeo ), and the acclimation slopes for R 25 and R 25 /N, where the ac climation slope represents the change in R 25 or R 25 /N per C ( flux C 1 ; more negative indicates stronger acclimation). For Anacardium and Luehea results from two experiments are shown. Different letters indicated significant differences between control and treatment at P < 0.05. Significant acclimation slopes ( P < 0.05) are indicated with an Species Growth form Treatment R 25 R 25 /N Q 10 Acclim SetTemp Acclim Homeo 1 ) mol m 2 s 1 mol (mg N) 1 s 1 R 25 R 25 /N R 25 R 25 /N R 2 5 R 25 /N Anacardium excelsum Tree 1 Control 0.91 0.044 2.3 1.13 1.09 0.83 0.80 0.047 0.0020 Warmed 0.81 0.040 2.6 2 Control 0.75 0.040 2.6 1.12 1.08 0.84 0.82 0.027* 0.0011* Warmed 0.67 0.037 2.4 Castilla elastica Tree Control 1.06 a 0.048 a 2.9 1.14 1.15 0.94 0.95 0.062 0.0032 Warmed 0.94 b 0.042 b 3.0 Luehea seemannii Tree 1 Control 0.94 0.050 a 2.7 1.04 1.14 0.84 0.92 0.029 0.0021 Warmed 0.90 0.044 b 2.6 2 Control 0.99 0.0 45 2.5 1.14 1.24 0.82 0.89 0.015 0.0018* Warmed 0.88 0.036 2.7 Bonamia trichantra Liana Control 0.73 0.059 a 3.0 1.05 1.21 0.73 0.84 0.008 0.0027 Warmed 0.69 0.048 b 2.9 Stigmaphyllon lindenianum Lian a Control 1.03 a 0.053 a 2.8 1.14 1.27 0.91 1.00 0.021 0.0026 Warmed 0.91 b 0.042 b 3.0
86 Figure 4 1. Example of acclimation, using the actual leaf temperature data and R 25 and Q 10 values for warmed and control leaves of Castilla elastica Warme d leaves exhibit down regulation of respiration at a set temperature (A 1 > B 1 ), but acclimated leaves respire more at their average nighttime temperature (28.8C) than control leaves experiencing their average temperature (26.6C) (A 1 < B 2 ), so acclimation is not completely homeostatic.
87 Figure 4 2 Example of experimental warming set up. A) W arming frame fit to the termina l shoot of the liana Bonamia B) A verage temperatures of control and warm ed leaves of 5 days of the 6 d ay experiment with this species Natural leaf angles were maintained as much as possible, but leaves were prevented from touching the heating rope by fitting rubber coated twist tie frames between the leaf and the heating rope.
88 Figure 4 3 Leaf dark respiration at 25C ( R 25 ) in rela tion to the average nighttime temperature these leaves experienced in the preceding 6 8 days for the tree species Anacardium Castilla and Luehea and the liana species Bonamia and Stigmaphyllon A E ). R 25 per unit leaf area. F J) R 25 per unit leaf nitrogen For Anacardium and Luehea two experiments were performed. Trend lines for the different experiments are shown in different colors. Dashed lines are not significant at P = 0.05.
89 Figure 4 4 Respiration at 25C ( R 25 ) in relation to the average nighttim e leaf temperature during the experiment, standardized by the species mean R 25 of control leaves (blue) at their average nighttime temperature (= 100%). Horizontal box plots indicate the mean, median and spread of warmed (red) and control (blue) leaves acr oss species. Vertical box plots show the mean, median and spread of standardized R 25 values of warmed and control leaves. Notice that some control leaves are warmer than warmed leaves because of variation within species, among branches, leaves and experi ments; some experimental periods had overall higher temperatures (see also Fig. 4 3 ), control leaves in these experiments experienced were warmer nighttime temperature than warmed leaves in other experiments.
90 Figure 4 5. Correlations between species le vel acclimation other leaf traits where acclimation is represented by the slope of R 25 versus the mean nighttime leaf temperature (more negative indicates stronger acclimation) A) Acclimation vs. leaf sugar to starch ratio B) Acclimation vs. a ve rage nigh ttime leaf temperature. The s olid line in A indicate s a significant correlation at P < 0.05 ; the dashed line is non significant ( P = 0.1).
9 1 Figure 4 6. The estimated annual nighttime carbon efflux from leaf respiration at Parque Natural Metropolitano un 2011), and under 4C warming with different acclimation scenarios ; no acclimation, acclimation results in 10% reduction of R 25 or R 25 acclimated to the average nighttime leaf temperature (T Leaf ) during the preceding week I n black numbers the percentage increase in fluxes relative to current are shown. Red numbers indicated the percentage decline in fluxes relative to the scenario of no acclimation. Error bars represent the standard deviation of flux estimates for the 1 7 yea rs.
92 CHAPTER 5 GENERAL PATTERNS OF THERMAL ACCLIMATION OF LEAF DARK RESPIRATION ACROSS BIOMES AND PLANT TYPES Background Climate warming is predicted to increase the release of carbon dioxide from the terrestrial biosphere into the atmosphere, thus trigger ing a positive climate terrestrial carbon feedback that accelerates warming (Cox et al ., 2000; Luo, 2007). However, plant respiration (non photorespiratory mitochondrial CO 2 release) may acclimate to warming and this acclimation may reduce the potential de cline in net primary productivity (NPP) (King et al ., 2006; Smith & Dukes, 2013). This chapter is an attempt to synthesize results from empirical studies on thermal acclimation of leaf dark respiration from across the globe. First, we briefly review the cu rrent understanding of thermal acclimation of respiration. We then discuss the aspects of tropical forests that lead to the supposition that tropical vegetation may respond differently to climate warming than cooler climate vegetation. Finally, we analyze published data on thermal acclimation of leaf dark respiration and discuss the results of the meta analysis in the context of climate warming as anticipated for the current century. Thermal A cclimation of Leaf Dark Respiration Respiration increases exponen tially with short term temperature increment. Th is sensitivity of respiration to changes in temperature is primary driven by an increase in the demand for cellular maintenance, associated with increased protein turnover and membrane leakage at higher tempe ratures (Amthor, 1984; Ryan, 1991). Respiration is essential for growth and survival of plants as it provides energy and carbon skeletons for biosynthesis (Penning de Vries, 1975), but respiration associated with growth only increases with warming when gro wth itself is stimulated by rising temperature (Franz et
93 al ., 2004). Thermal acclimation of respiration thus primarily involves changes in respiration associated with maintenance processes, rather than with growth. Thermal acclimation of respiration is a p hysiological, structural, or biochemical adjustment by an individual plant in response to a change in temperature, that is manifested as an alteration in the short term response to temperature (Smith & Dukes 2013 ) (Fig. 5 1). Acclimation of respiration to a higher temperature regime results in a decreased rate at the new, elevated temperature compared to non acclimated plants measured at that temperature (Atkin & Tjoelker, 2003; Atkin et al ., 2005). T hermal acclimation of respiration functions to maintain optimal supply of ATP and carbon skeletons while minimizing carbon loss from respiration associated with maintenance processes. This may be achieved by changes in mitochondrial membrane composition to minimize ion leakage under warmer conditions (Raison et al ., 1980), or by a reduction in the overall protein turnover rate, e.g., by a change in mitochondrial protein composition (Atkin et al ., 2005). When respiration under warmed conditions equals the respiration rate exhibited by leaves under control conditi ons, homeostasis of respiration is maintained ( Fig. 5 1). Acclimation may occur within a few days of a temperature change (Rook, 1969; Billings et al ., 1971; Atkin et al ., 2000; Bolstad et al ., 2003; Lee et al ., 2005), but longer exposure to a new tempera ture may result in a greater degree of homeostasis (Smith & Hadley, 1974). As climate warming is a gradual process, the potential for acclimation needs to be assessed not just over a few days, but also following extended periods of warming. Furthermore, le aves developed under an experimental temperature are often more completely acclimated than fully formed leaves transferred to that temperature
94 (e.g., Campbell et al ., 2007). Campbell et al (2007) found no systematic differences in acclimation among differ ent growth forms, but Tjoelker et al (1999) found boreal evergreen tree species to acclimate better to experimentally imposed temperature differences than deciduous tree species. Given that most dynamic global vegetation models (DGVMs) use plant functiona l types to characterize vegetation, it would be valuable to identify systematic differences in acclimation potential among plant functional types if such differences were to exist. Two types of acclimation of respiration have been identified (Atkin & Tjoel ker, 2003) ( Fig 5 1). The type of acclimation indicates the mechanism underlying the acclimation process. Type I acclimation involves a decrease in the slope of the respiration temperature response curve (i.e., lower short term temperature sensitivity (Q 1 0 ) in warm acclimated leaves), typically under influence of regulatory changes of existing respiratory enzymes (Atkin et al ., 2005). Type II acclimation, a decrease in the elevation of the temperature response curve of respiration (i.e., lower respiration across the temperature range, without a change in Q 10 ), typically involves a change in overall respiratory capacity. The respiratory capacity may change under the influence of a change in the relative amounts of individual respiratory enzymes, or in the co ncentration of mitochondrial proteins (Atkin et al ., 2005). Type II acclimation is expected to be more common for leaves developed at elevated temperature whereas Type I acclimation, associated with changes in existing enzymes, is thought to be more common in leaves that existed prior to the change in temperature (Atkin & Tjoelker, 2003; Atkin et al ,. 2005). Ultimately, both types result in a reduction in respiration at warm conditions compared to non acclimated leaves. It will nevertheless be valuable to i dentify patterns
95 in acclimation type to aid in predictions of changes in respiratory fluxes, as the short term sensitivity (Q 10 ) changes in Type I but not in Type II. Potential Differences in Warming Response of Tropical and Cool Climate Vegetation Based p rimarily on mid and high latitude acclimation studies, the current consensus is that most plant species can, in principle, acclimate to changes in temperature (Atkin & Tjoelker, 2003; Atkin et al ., 2005), although acclimation responses are often species s pecific (e.g., Larigauderie & Krner, 1995). There are, however, three important differences between lowland tropical forests and higher latitude ecosystems in their responses to climate warming. First, tropical forests are close to their thermal optimum temperature (Doughty & Goulden, 2008) and further warming of tropical ecosystems will push the majority of tropical forests into a climate envelope currently not occupied by closed canopy forest (Wright et al ., 2009). Respiration at ambient temperature inc reases exponentially with mean annual temperature, and is thus already high in tropical forests (Wright et al ., 2006). The absolute increase in respiration per degree of warming above ambient will be greater in the tropics than in high latitudes, as warmin g will occur along the steeper end of the exponential temperature response curve (Dillon et al ., 2010). Compared to warming of cool grown plants, a much greater down regulation of respiration is required for acclimation to maintain homeostasis of respirati on per degree of increase of temperature. This may constitute a considerably greater challenge than down regulation of the relatively small change in absolute metabolic rates with warming on the cooler end of the tem perature spectrum.
96 Second, tropical for ests experience minimal seasonal temperature fluctuations (Wright et al ., 2009), and the thermally stable environment of the tropics may not have favored evolution of the capacity to acclimate to temperature changes (Janzen, 1967; Cunningham & Read, 2003a; Ghalambor et al ., 2006). Comparative studies of thermal acclimation of photosynthesis of temperate and tropical rainforest species indeed found that tropical species do not acclimate as completely as temperate species (Cunningham & Re a d, 2002, 2003b). Thi rd, for the past 2.6 million years of the Quaternary tropical regions have experienced conditions that were relatively cool compared to current and near future temperatures, and natural selection would not have favored heat protective genes and traits (Cor lett, 2011, 2012). The proximity of tropical vegetation to experiencing supra optimal leaf temperatures makes the issue of high temperature stress particularly pressing in the tropics. Heat stress can lead to protein denaturing (Vierling, 1991) and increas ed membrane fluidity (Quinn, 1989), factors that increase the respiratory demand for maintenance, and that are as such conflicting with an acclimatory decrease in respiration. Very few studies on thermal acclimation have been done in the tropics and the p redictions about respiratory acclimation in tropical forests are therefore difficult to make. By synthesizing data from around the globe in this meta analysis we sought to determine the effects on thermal acclimation of respiration of 1) the biome of origi n of the study species, 2) the duration of exposure to warming, 3) the developmental status of the leaf (pre existing when temperature change was imposed, or newly developed at the experimental temperature), 4) the growth form under investigation, and 5) t he
97 degree of warming or the temperature difference across contrasting temperatures. General patterns in thermal acclimation across the globe may inform us about the likely response of tropical forests to climate warming. Methods Data Selection We analyzed the results of studies where leaf dark respiration was measured for plants grown under different temperatures; measured repeatedly under changing ambient temperature conditions; were grown in common gardens at different ambient temperature regimes; or were exposed to experimental warming above ambient temperature. We searched Google Scholar and the Institute for Scientific Information (ISI) Web of Science for studies that 1) used non cultivated plant species, 2) that exposed plants to at least two growth/ac climation temperatures, and 3) that measured respiration at the respective ambient temperatures, or at the same temperature for both groups, or both (e.g., by measuring full temperature response curves of respiration of control and warmed leaves). Most stu dies used growth cabinets to assess the effect of growth temperature or short term temperature changes. Research on physiological cold acclimation commonly uses the same set up, but these are not included in this study. In total 30 studies were included (T able B 1), reporting on 237 temperature contrasts of 87 species. While the motivation for this study was to synthesize information that could inform us about plant responses to climate warming, only a small number of the available studies compared respirat ion at ambient temperature with respiration of leaves warmed to above ambient temperatures (18 out of a total of 237 species by temperature contrasts; 14 of 87 species). These studies report on species from alpine, a rctic and Antarctic, boreal, temperate, and tropical climates and include
98 forbs, graminoids (sedges and grasses), and evergreen and deciduous shrubs, trees and lianas (Table 5 1). Studies that warmed plants or leaves during the night only were also included, as respiration does not necessarily a cclimate to mean daily temperature (e.g., Atkin et al ., 2000), and has been shown to acclimate to nighttime temperature instead (Bruhn et al ., 2007). Data were extracted from tables, (enlarged) figures or from online supplementary information. Data A nalys i s To assess acclimation responses quantitatively we extracted information from temperature method and the homeostasis method ( Fig 5 2). For studies that used growth cabinets to expose plants to two or more different temperatures, the lower of the two was considered as the control temperature. Similarly, when temperature changes associated with seasons, changes in weather systems, or geographical range of common gardens were us ed, the lower temperature regime was considered to represent the control. To calculate Acclim SetTemp the control temperature was used as the set temperature whenever possible; sometimes a temperature intermediate to the control and warming temperature was used instead. For 56 temperature contrasts (41 species) data were available to calculate both acclimation ratios, but more commonly only Acclim SetTemp (139 contrasts, 39 species) or Acclim Homeo (43 contrasts, 36 species) could be calculated. Because Accli m LTR10 ( Fig 5 2) requires information on both the initial Q 10 and the long term acclimation ratio LTR 10 this metric could not be calculated for most studies and was not included in the following analyses. It was not always possible to determine the uncer tainty associated with the values used to calculate the acclimation ratios; variances were not always given; in some cases values were
99 extracted from fitted curves for which no confidence intervals were presented; or tables did not specify whether standard deviations of standard errors of the mean were presented. When sample size and standard errors were available, we tested whether respiration of control and warmed leaves were significantly different by assuming a normal distribution of the data and calcul ating 95% confidence intervals. Acclimation Type To properly assess which type of acclimation has occurred one would ideally have temperature response curves that go down to the low temperature respiration In case of Type I acclimation the contro l and acclimation curve would intersect. Generally, however, respiration is not measured at low enough temperatures identify where the temperature response curve would int ersect. It is more common to compare the slopes of log transformed temperature response curves to test difference in slopes. When the slope of the warm acclimated leaf is lower, Type I acclimation is supposed. If the slopes are not significantly different, but have lower intercept for warm acclimated leaves then Type II acclimation is implicated. Accordingly, when data was available to compare slopes of control and acclimated leaves, results were classified as Type I or Type II, or no acclimation. To explo re whether certain environmental factors may influence either Acclim SetTemp or Acclim Homeo general linear models were tested for the following explanatory variables: biome of origin, growth form, temperature difference, duration of exposure, maximum nightt ime temperature, method (warming above ambient, or not), and leaf developmental status (whether the measured leaves existed prior to exposure to the warmer temperature, or that they developed at the warm temperature). Leaf habit
100 (evergreen v ersus deciduous ), was determined for woody species only, so in addition to the above models, we analyzed Acclim SetTemp and Acclim Homeo of the subset of woody species with models that included leaf habit. We also determined the most parsimonious significant linear regres sion model of Acclim SetTemp and Acclim Homeo that revealed minimal pattern in the residuals using the same candidate predictors as above. All statistical analyses were performed in R version 2.14.1 (R Development Core Team 2011). Results Temperature contras ts over which acclimation was determined ranged from 0.3C in an in situ infrared heating experiment of eucalypt seedlings (Bruhn et al ., 2007), to 21C in a growth cabinet study on temperate tree, forb and graminoid species grown in hydroponics (Campbell et al ., 2007). Respiratory response to change in temperature ranged from acclimation leading to complete homeostasis of respiration across the study temperatures, to no change in the instantaneous temperature response curve whatsoever. Acclimation to warme r temperatures resulted in a down regulation of respiration at a set temperature in 179 out of 195 contrasts for which information on respiration at a set temperature was available (Fig. 5 3 ). This down regulation was significant in 75% of the cases for wh ich data was available to determine significance. Complete homeostasis (Acclim Homeo = 1.0), or acclimation leading to over compensation (Acclim Homeo > 1.0) was, however, rare and the mean Acclim Homeo value (0.76 0.29; mean SD) was significantly smaller than 1.0 (t test, t = 7.9, df = 97, P < 0.0001). The tendency for down regulation of respiration in warm acclimated leaves was consistent across biomes and growth forms (Table 5 1).
101 Treatment and Species E ffects on A cclimation of R espiration Summary by b iome appears to suggest lower Acclim SetTemp in the tropics than in cooler ecosystems (Fig. 5 4 ). However, this apparent pattern in Acclim SetTemp across biomes could be explained by the average degree of warming used in the experiments; in cooler ecosystems studies warmed plants by a greater degree than studies from the tropics (Table 5 1). When the effect of the degree of warming was accounted for, the biome of origin was no longer a significant predictor of either of the acclimation ratios. The mean maximu m nighttime temperature that treatment plants were exposed to did not affect the acclimation potential of respiration either; studies at high temperature ( warmest 33%: mean T Night Max = 25.6C) did not result in systematically lower acclimation ratios than studies done at low temperatures ( coldest 33%: mean T Night Max = 18.0C) (Fig. 5 4 C,H ). The duration of warming had no effect on Acclim SetTemp (Fig. 5 4 ), and the mean Acclim SetTemp of studies exposing leaves to an experimental temperature for less than 25 days was not significantly different from the categories of warming 25 50 days and warming for more than 50 days. All three categories had mean Acclim SetTemp values that were signifi cantly greater than 1.0 (t test P < 0.001 for all). Duration of warmi ng had a positive effect on Acclim Homeo when only pre existing leaves were considered; the longer pre existing leaves were warmed, the more res piration approached homeostasis Acclim Homeo of leaves developed at the experimental temperature was independent of duration of the experiment. The developmental status of the leaves did not affect Acclim SetTemp but it did affect Acclim Homeo ; newly developed leaves exhibited greater homeostatic acclimation of respiration than pre existing leaves (Fig. 5 5 ). When ot her factors, such as duration and
102 the degree of warming were accounted for, leaf development status affected Ac clim Homeo only in woody species ( P = 0.04 Table 5 2). There were no significant differences in Acclim SetTemp and Acclim Homeo among growth forms, with large variation existing within each (except for lianas, for which only two contrasts were included) (Fig. 5 4 ). In a simple one factorial comparison evergreen and deciduous leaves did not differ significantly in either Acclim SetTemp or Acclim Homeo In the full model, however, leaf habit was significant at P < 0.05 for both acclimation ratios, with evergreen leaves exhibiting lower Acclim SetTemp values, but higher Acclim Homeo than deciduous leaves. The degree of warming (temperature interval of the co ntrasts) had a significant effect on both Acclim SetTemp and Acclim Homeo However, whereas Acclim SetTemp increased with the degree of warming, Acclim Homeo decreased with the degree of warming (Fig. 5 4 ). There was a significant interaction between the durat ion and the degree of warming ( P < 0.05; Table 5 2) that affect ed Acclim SetTemp : Acclim SetTemp increased more strongly with the degree of warming in leaves that were warmed longer than in leaves warmed for a shorter duration (Fig. 5 6 ). Leaf habit (ever green v ersus deciduous) was significant for both Acclim SetTemp and Acclim Homeo in the subset of data for which information on leaf habit was available. Leaf devel opmental status (pre existing versus newly developed leaves) was also significant in the Accli m Homeo model of the woody taxa (Table 5 2). For both acclimation ratios the most parsimonious significant model only included the degree of warming. In Situ W arming The eight studies that elevated temperature above ambient temperature in the field warmed leaves by an average of 2.0C (median 2.2C). This contrasts with a mean
103 temperature difference of 11.1C (median 10.0C) in the other studies. Median duration of the experimental warming was the same for the two groups at 30 days, whereas the mean duratio n was much longer for in situ studies because of two studies that warmed Scots pine trees in Finland for several years (Wang et al 1995; Zha et al 2002). Mean Acclim SetTemp with in situ warming was 1.05, which was significantly lower than the mean of t he remaining studies (t test t = 9.8, P < 0.0001) (Fig. 5 5 ). Acclim Homeo of in situ warmed leaves was 0.96, marginally higher than in other studies (t test t = 3.4, P = 0.056). When the degree of warming was taken into consideration, results from the i n situ warming studies no longer differed from the other studies. Acclimation T ype The type of acclimation exhibited by warmed leaves could be determined for 54 temperature contrasts (11 on pre existing leaves, 43 on leaves developed under the experimental temperature). Both pre existing and newly developed leaves exhibited Type II acclimation more often than Type I acclimation (Fig. 5 7 A ). Evergreens showed an even split between Type I and Type II acclimation, but deciduous species exhibited Type II acclim ation in almost 75% of the cases. Interestingly, 4 out of 6 contrasts of pre existing evergreen leaves showed Type I acclimation, whereas all 5 contrasts of pre existing deciduous leaves exhibited Type II acclimation. Newly developed leaves of both deciduo us and evergreen species exhibited Type II acclimation in > 60% of the cases (Fig. 5 7 C ). Leaves that were warmed by less than 5C exhibited Type I acclimation in the majority of the cases, whereas Type II acclimation was more common in leaves tha t were wa rmed by 5 10C or by > 10C (Fig. 5 7 B ). Type II acclimation was associated with a greater down regulation of respiration (higher Acclim SetTemp ) and more homeostatic acclimation than Type I acclimation (Fig. 5 8 ).
104 Discussion Several trends in respiratory a cclimation could be identified from the current analysis of thermal acclimation data from multiple studies. There is an overwhelming tendency for reduction in respiration in warm acclimated leaves when compared with control leaves at a set temperature, and there is no indication th at different growth forms and plants from different biomes differ systematically in this respect. However, the two acclimation indices that we used gave conflicting results, which highlights a problem with quantitative assessment of acclimation. Considerations for Q uantifying A cclimation When acclimation is simply defined as a change of the short term temperature response curve of respiration, any significant deviation in elevation or slope indicates that acclimation has occurred. However, assessment of the degree of acclimatory changes requires careful consideration of the most relevant metric. Here we used two metrics of acclimation both of which give higher values when more acclimation has occurred. The higher the Acclim SetTemp v alue, the greater the down regulation of respiration following warming, and thus the greater the degree of acclimation. Similarly, the higher Acclim Homeo the smaller the increase in respiration compared to leaves at control temperature, and the greater th e degree of acclimation. However, the degree of warming, the strongest single predictor of both ratios, increases Acclim SetTemp while decreasing Acclim Homeo Fig ure 5 1 illustrates three acclimation scenarios for Type I and acclimation, in which respiration of warm and hot grown leaves is down regulated, but respiration is not homeostatic across temperatures. A greater degree of warming (i.e., n a greater Acclim SetTemp (A 2 /C 2 > A 2 /B 2 ), but in a smaller
105 Acclim Homeo value (A 1 /B 2 > A 1 /C 3 ). Partial acclimation occurred in 67 of the 99 cases for which Acclim Homeo could be calculated (Acclim Homeo 0.95; in 35 cases R Control at T Control was significan tly higher than R Warm at T Warmed ). Under the scenario of complete homeostasis, Acclim SetTemp increases with warming as before, while Acclim Homeo stays the same (A 1 /B 2 = A 1 /C 3 (Acclim Homeo value be tween 0.95 and 1.05, and no significant difference between R Control at T Control and R Warm at T Warmed ). Only in the scenario of over compensation does Acclim Homeo increase with the degree of warming. This requires an enormous degree of down regulation of re spiration, corresponding with a very large decrease in Q 10 (in case of Type I acclimation) or a considerable down regulation of the respiratory capacity (in case of Type II acclimation). Indeed, over compensation appears to be an uncommon phenomenon (10 ca ses in this meta analysis with Acclim Homeo >1.05, but only in one case was R Warm at T Warmed significantly smaller than R Control at T Control ). Clearly, complete homeostasis requires a considerable alteration of the short term temperature response, and acros s wide temperature ranges, complete acclimation is often not achieved. What, then, is the better indicator of the degree of acclimation? When Type I acclimation occurs Acclim SetTemp is dependent on the temperature at which it is determined (A 1 /C 1 > A 2 /C 2 i n Fig 5 1). The question when determining acclimation according to the set temperature method then is, what is the ecologically relevant temperature at which to determine respiration of plants that are grown at contrasting temperature regimes. The choice of reference temperature to determine respiration is often arbitrary and not necessarily of ecological relevance (Bruhn et al ., 2007), and furthermore difficult to standardize across climate regions with contrasting temperature
106 regimes. In contrast, Acclim Homeo is not inherently dependent on the measurement temperature regardless of the type of acclimation, and it only considers environmentally relevant temperatures. Acclim Homeo thus appears to be the more useful indicator of acclimation. Estimates of Accli m SetTemp may, however, still contribute to improving global carbon flux estimates by implement ing them in DGVMs. In such models respiration at a given temperature is generally calculated by adjusting a base rate of respiration at an (arbitrary) reference t emperature to current temperature by multiplying this rate by a temperature sensitivity parameter, e.g., based on a Q 10 value. With information on Acclim SetTemp the base rate of respiration itself can be made dependent on the acclimation temperature (e.g. nighttime temperature in the past week See Chapter 4 ). While Acclim Homeo is biologically more relevant, Acclim SetTemp may thus still have its value in quantifying acclimation of respiration in dynamic models. B iome dependent A cclimation Potential? Two recent meta analyses found that warming of tropical plants decreases biomass accumulation whereas warming stimulates biomass accumulation in all other ecosystems and climate regions (Lin et al ., 2010; Way & Oren, 2010). Furthermore, in a lowland tropical f orest in Costa Rica tree diameter increment is reduced in years with above average nighttime temperatures (Clark et al ., 2003, 2010). These observations suggest that tropical species may be close to their thermal optimum and have limited capacity for therm al acclimation. After accounting for factors such as the degree of warming and the duration of the experiment, the biome of species origin had no effect on either Acclim SetTemp or Acclim Homeo in the current analysis. The relatively low values of Acclim SetT emp of the seven tropical species were explained by the small degree of warming the tropical species in this analysis were exposed to and do not preclude the
107 possibility for a greater degree of acclimation when warmed more. However, no studies to date have looked at the effect of warming tropical species by 5C or more above their current ambient temperature under otherwise natural conditions to simulate predicted end of century temperatures. As an exception, in a growth cabinet study Cheesman & Winter (201 3) exposed well watered seedlings of the tropical pioneer tree species Ficus insipida Willd. and Ochroma piramidale (Cav. ex Lam.) Urb. to nighttime temperatures of up to 31C while maintaining daytime temperature at 33C. Krause et al (2013) grew F. insi pida at ever warmer conditions at 39C, with nighttime temperature of 32C. Mean Acclim Homeo values were 0.56 and 0.51 for F. insipida ( R Warmed at T Warm > R Control at T Control P < 0.05) and 0.80 for O. piramidale This suggests that warming to temperature s that are predicted for the end of the current century may have different consequences for the respiratory carbon flux of different species. Interestingly however, in both these studies seedlings accumulated significantly more biomass in the high nighttim e temperature treatment, despite the fact that photosynthesis did not increase with warming. These results show that increase nighttime carbon loss does not necessarily impede tree growth in tropical forests. Species differences in thermal acclimation may have consequences for species composition, vegetation dynamics, and ecosystem functioning in a warmer world, so it will be important to verify the generality of this pattern with later successional tree species and to identify causes of interspecific varia tion in the capacity for thermal acclimation. Near homeostatic Acclimation to Development Temperature Exposure to warmer temperature during leaf development leads to more homeostatic acclimation. Adaptive plasticity should optimize performance at new envir
108 transferred to warmer temperatures achieved a smaller degree of homeostasis, but the longer pre existing leaves are exposed to the new temperature, the more they approach h omeostasis. Type I and Type II A cclimation in P re existing and N ewly developed L eaves In contrast to previous reports (Atkin et al ., 2005), pre existing leaves exhibited Type II acclimation more frequently than Type I acclimation, similar to newly develo ped leaves. In some of the species in which pre existing leaves exhibited Type II acclimation nitrogen concentrations decreased with increasing temperature (Lee et al ., 2005), which suggests down regulation of the metabolic capacity. In other species, howe ver, nitrogen concentration did not decrease with warming ( see also Chapter 4 ). Clearly, pre existing leaves can down regulate the respiration capacity at higher temperatures, but the mechanism employed to do so is not well understood. Acclimation and Clim ate Warming As climate continues to warm, plants experience temperature changes from one year to the next that are relatively small compared to some of the temperature differences included in the current study. Small temperature differences are more likely to result in homeostatic rates of respiration than large temperature differences. Furthermore, newly developed leaves maintain a greater degree of homeostasis than pre existing leaves. Gradual warming is unlikely to expose leaves to dramatically higher me an annual or mean nighttime temperatures than those they experienced when they developed, especially in conditions where intra annual temperature variation s are small, such as in tropical forests. This suggests that most species are indeed likely to acclim ate to a certain degree to warming. Warming is, however, not gradual, even if the rise in mean annual temperature change is. Heat waves will occur more frequently, and
109 with increasing intensity over the current century (Meehl & Tebaldi, 2004). Under heat w ave conditions pre existing leaves will be exposed to temperatures considerably higher than their development temperature, and acclimation will at best result in limited homeostasis. We did not find evidence for biome differences in the capacity for acclim ation, nor were leaves exposed to high maximum nighttime temperatures less likely to exhibit acclimation. However, whether the response of pre existing leaves to extreme warming during a heat wave event differs by biome remains unknown. Recent tropical se edling studies (Cheesman & Winter, 2013; Krause et al ., 2013) show that even among pioneer species, large differences in thermal acclimation of respiration exist. Clearly more tropical data are needed to be able to make generalizations about the potential for thermal acclimation in the tropics. The current study included 13 temperature contrasts of seven tropical species. In recent global meta analyses of temperature effects on biomass accumulation Way & Oren (2010) included ten data points for tropical and subtropical species out of 434 total data points, and in Lin et al (2010) only nine out of 537 data points came from conditions of a mean annual temperature of 20C or higher Given the angiosperm diversity of tropical regions, the amount of empirical da ta from tropical forests is decidedly small and more data are needed to predict global patterns in the acclimation response of respiration, and to enable more reliable predictions of respiratory carbon fluxes by informing global circulation models (Malhi e t al ., 2009; Reed et al ., 2012).
110 Table 5 1. Summary of studies of thermal acclimation of respiration in the lab (growth cabinets) and field. Mean Acclim SetTemp and Acclim Homeo ( SD ) are shown by biome and growth form. Mean degree (range) of warming ( ), the number of temperature contrasts (n c on ) and species (n sp ) are provided. See Appendix B for r eference s Biome Lab Field Total Ref. Growth form n c on n s p Acclim SetTemp Acclim Homeo n c on n s p. Acclim SetTem p Acclim Homeo n c o n n s p Alp ine 19 19 Forbs 15 15 10.4 (10.0 11.0) 1.65 0.32 0.62 0.26 15 15 3,7,12 Graminoids 4 4 10.3 (10.0 11.0) 1.86 0.59 0.19 4 4 7,12 Arctic & Antarctic 7 4 Forbs 3 2 7.7 (5.0 10.0) 1.52 0.40 0.94 0.12 3 2 1,3,28 Graminoids 2 1 6.5 (5.0 8.0) 1.46 0.22 1.04 0.01 2 1 28 Shrubs Evergreen 2 1 15.0 1.42 0.02 0.54 0.07 2 1 20 Boreal 17 7 Tr ees Deciduous 6 3 9.0 (6.0 12.0) 1.38 0.35 0.68 0.16 4 2 4.3 (3.4 5.3) 1.27 0.49 nd 10 4 10,22 Evergreen 4 2 9.0 (6.0 12.0) 1.45 0.24 0.78 0.17 3 2 4.0 (1.5 8.0) 1.06 0.19 nd 7 3 22 24,26,29 Temperate 189 46 Forbs 68 17 12.1 (5.0 21.0) 1.56 0.55 0.88 0.42 3 3 1.1 0.94 0.06 nd 71 20 1,12 Graminoids 30 7 13.0 (5.0 21.0) 1.76 0.57 0.73 0.35 2 2 1.6 (1.1 2.1) 1.17 0.34 nd 32 9 6,9,12,14,30 Shrubs Evergreen 16 3 12.8 (6.0 21.0) 1.66 0.60 0.71 0.16 16 3 6,25 Trees Deciduous 17 5 11.8 (5.0 21.0) 1.93 0.90 0.43 0.18 10 5 5.3 (1.0 10) 1.33 0.34 nd 27 7 4,6,10,13,15,16 Evergreen 36 7 12.8 (5. 0 21.0) 1.55 0.41 0.76 0.18 7 3 1.8 (0.3 3.9) 1.06 0.45 1.12 0.30 43 7 2,5,6,14,16 18,21,27 Tropical 13 7 Lianas Evergreen 2 2 2.9 (2.5 3.3) 1.10 0.06 0.83 0.05 2 2 19 Trees Deciduous 7 2 6.6 (3.0 10.0) 1.16 0.75 0.17 3 3 2.8 (2.4 3.5) 1.08 0.06 0.88 0.04 10 4 8,11, 19 Evergreen 1 1 3.2 1.14 0.87 1 1 19
111 Table 5 2. P values of model s of the dependence of Acclim SetTemp and Acc lim Homeo on species traits and experimental conditions. Biome, biome of species origin (Arctic/Antarctic, Alpine, Boreal, Temperate and Tropical); Growth Form, (Forbs, Graminoids, Shrubs, Trees, Lianas); Leaf habit evergreen or deciduous ( only a factor fo ); Pre e xisting, leaves existing prior to experiencing the experimental temperature vs. leaves developed at the experimental temperature; Max T Night highest nighttime temperature in the experiment ( e.g., the target nighttime t emperature in a warming treatment) ; Method, the method used in the study ( in situ warming above ambient or all else) ; Duration, duration of exposure to experimental temperature; Degree of Warming, mean temperature difference between control and warmed lea ves. The most parsimonious significant model only included Degree of Warming for both acclimation ratios. Acclim SetTemp Acclim Homeo Full model Parsimonious model Full model Parsimonious model Biome 0.44 0.20 0.54 0.23 Growth Form 0.62 0.82 0.75 0.09 Leaf habit 0.01 0.02 Pre e xisting 0.75 0.28 0.19 0.04 Max T Night 0.80 0.94 0.63 0.98 Method 0.52 0.75 0.28 0.09 Duration 0.97 0.88 0.57 0.09 Degree of Warming <0.0001 <0.0001 <0.0001 <0.01 <0.0001 <0.01 Durat ion Degree of Warming 0.04 0.05 0.64 0.18 Duration Pre e xisting 0.78 0.71 0.41 0.52 Model R 2 0.36 0.30 0.46 0.34 0.20 0.65 P <0.0001 <0.0001 <0.0001 <0.01 <0.0001 <0.0001
112 Figure 5 1. Acclimation of respiration illustrated
113 Figure 5 2. Ex planation of methods of q uantif ication of acclimation.
114 Figure 5 3 Relative change in respir ation rate at a set temperature following acclimation of 64 plant species, based on 195 sets of respiration measurements at contr asting acclimation temperatures. A) F orbs B) G raminoids C) S hrubs D ) Trees and lianas. Respiration at the control temperature is set to 100% and respiration of warmed leaves was either taken directly from the published source, or calculated from the Acclim SetTemp value. All contrasts a re shown in this graph, also if the change in respiration at a set temperature was not significant.
115 Figure 5 4 Acclimation parameters Acclim SetTemp ( A E ) and Acclim Homeo ( F J ) in relation to plant traits and experimental conditions. A,F ) Effect of b iom e of orig in (Alp., Alpine; Arct., Arctic/ Antarctic; Bor., Boreal; Temp., Temperate; Trop., Tropical) B,G) Effect of growth form (Gram., Graminoids; Lian., Lianas; Shru., Shrubs) C,H) Effect of maximum nighttime temperature warmed plants were exposed to D,I) Effect of degree of warming, or temperature difference across contrasts E,J) Effect of duration of exposure to the experimental temperature. The continuous data in C E and H J were also binned in three groups of similar sample size and boxplots are s hown at the median x axis value of the bins. Two data points from multi year warming experiments were omitted from the scatter plot in E but were included in the box plot. Different letters indicate significant difference s among groups at P < 0.05 (one wa y ANOVA). The black lines in D and I represent significant linear regressions at P < 0.05
117 Figure 5 5 Acclimation parameters Acclim SetTemp ( A C ) and Acclim Homeo ( D F ) in relation to experimental conditions and leaf traits. A,D) Acclimation in relation to leaf developmental age (pre existing at the time of warming vs. newly developed under warmed conditions). B,E) Acclimation in relation to leaf type (Evergreen vs. Deciduous) C,F) Effect of method of warming on acclimation ( in situ warming above ambien t, vs. other methods. Different letters indicate significant difference between groups at P < 0.05 (one way ANOVA). The number of observations per category is shown below each boxplot.
118 Figure 5 6 Interaction effect of duration of warming and degree of warming on Acclim SetTemp illustrated by the stronger temperature effect on leaves warmed for more than 30 days than for leaves warmed for up to 30 days. Data points are slightly jittered to reduce overlap and increase visibility.
119 Figure 5 7 Frequency of Type I and Type II acclimation of respiration A ) Acclimation type for leaves present prior to warming (Pre existing) and leaves developed under warmed conditions (Newly developed) B ) Acclimation type in relation to th e degree of warming experienced C ) Acclimation type of Pre existing (PE) and Newly developed (ND) leaves of evergreen and deciduous woody species In total 54 sets of observations on 41 species were included in A and B and 34 sets of observations on 23 species in C
120 Figure 5 8 Accli mation parameters Acclim SetTemp and Acclim Homeo under Type I and Type II acclimation of respiration. A) Acclim SetTemp B) Acclim Homeo Different letters indicate significant difference between groups at P < 0.05 (one way ANOVA). The number of observations per category is shown below each boxplot.
121 CHAPTER 6 CONCLUSIONS There is uncertainty about the effect s of climate change on tropical forests. While increased CO 2 in the atmosphere may, in theory, result in an increase in net primary productivity ( NPP ) by s timulating photosynthesis (Lloyd & Farquhar, 2008), rising temperatures may reduce NPP by increasing plant respiration more than gross photosynthetic rates. A decline in NPP can reduce the carbon sequestration service provided by tropical forests, possibly transforming them from net carbon sinks to carbon sources to the atmosphere which would have important implications for the global carbon cycle. Yet, little is known about the temperature responses of leaf respiration in tropical forest. In my dissertati on, I assessed the temperature response of leaf dark respiration of trees and lianas in the upper canopy of a tropical forest at different time scales. I determined the instantaneous physiological responses at the species level in the field ( C hapter 2 ) and at the leaf level in the laboratory (Chapter 3), and I assessed thermal acclimation response to elevated nighttime temperatures in the longer term (6 8 days) ( C hapter 4) In Chapter 2, I report the results of in situ measurements of leaf dark respiration of 461 upper canopy leaves of 26 species of tree and liana in a tropical forest in Panama using a new protocol that ensured thermal equilibrium of the measured leaves with the rest of the plant. Leaves were darkened with foil overnight until they were mea sured sequentially during ambient temperature rise in the morning. From these measurements temperature response curves of respiration were constructed for each of the 26 species, and values of Q 10 ( the proportional increase of respiration per 10C warming) were calculated. These measurements revealed to respiration rates at 25 C
122 ( R 25 ) were on the high end of what has previously been published for tropical forest trees. The high dark respiration rates may reflect the relatively high proportion of early and mid successional tree species in this secondary forest or the relatively high fertility of the site. However, respiration in early and mid successional species was not significantly higher than in late successional species within the site, and respiratio n rates were still high compared to published data when expressed per unit leaf nitrogen. Another possible reason for the comparatively high respiration rates was that I was better able to access fully sun exposed leaves with the canopy crane t han authors of previous studies, who used towers or sling shots to access canopy leaves. The Q 10 values estimated from the temperature response curves of respiration were on average higher than what is commonly assumed in global models, which means that published glo bal models may underestimate leaf respiratory carbon fluxes from tropical forests under warm conditions. Interspecific differences in respiration rate s correlated positively with species differences in photosynthetic capacity, leaf nitrogen content, and le af mass per unit leaf area, and negatively with median leaf lifespan. These trait correlations are in line with predictions from the leaf economic spectrum ( I. J. Wright et al ., 2004), which constrains leaf traits of plant species across the globe to a sin gle axis of variation based on the return on investments. The results of the correlations among in situ respiration rates and leaf traits reported in C hapter 2 support the notion that even within the upper canopy of a single tropical forest, species differ ences can be expressed as suites of traits associated with slow versus rapid metabolism. The advantage of the ecological relevance of in situ measurements reported in Chapter 2 came at the cost of the inability to document temperature sensitivity for
123 indiv idual leaves. Hence, the intra specific variation in Q 10 could not be quantified in Chapter 2. Chapter 3 addresses this concern by determining temperature sensitivity of leaf dark respiration on detached leaves in the laboratory, where leaf temperature co uld be controlled precisely and temperature response curves of respiration could be determined at the leaf level. I determined Q 10 values for a total of 123 upper canopy leaves from 28 species. Decomposing the variance in Q 10 data revealed that > 40% of th e total variance was explained by Q 10 differences among leaves within species, with most of the remaining variance occurring at the level of species within plant functional types. The total variance in Q 10 was, however, small compared to variance of R 25 I n agreement with the results of in situ measurements reported in Chapter 2, the average Q 10 was significantly greater than 2.0 (range of species means 2.0 2.9), and Q 10 did not differ between growth forms or among plant functional types based on tree succe ssional status. Q 10 values were not related to leaf chemical and structural traits, and interspecific variation in Q 10 thus appears to be independent of the leaf economic spectrum. Q 10 could, however, be explained by a multiple regression model containing total non structural carbohydrates and growth form (lianas versus trees), but only 26% of the variance could be explained. Among these 28 co occurring species significant pair wise correlations existed between R 25 and other leaf traits similar to the resul ts of the in situ measurements from Chapter 2. However, the best multivariate models of respiration did not include leaf nitrogen content, which is widely regarded as a key trait in the leaf economics spectrum, and is incorporated as a predictor of respira tion in some coupled climate vegetation models (e.g., Friend et al ., 1997; Bonan et al ., 2003). Instead, leaf phosphorus content was the strongest single predictor and the only
124 variable that was part of all significant multiple regressions of R 25 The best model contained leaf phosphorus content, photosynthetic capacity and leaf mass per unit area and explained 64% of the variance in R 25 Using these multiple regressions, I calculated the annual leaf respiratory carbon (C) flux from this forest to be 7.4 Mg C ha 1 yr 1 Th is estimate is comparable to those from Amazonian tropical forests that maintain greater leaf area index and have a shorter dry season than my study site. The instantaneous temperature sensitivity of respiration is valuable for calculation of respiration at different nighttime temperatures, but predictions of the long term warming effects on respiration require information on temperature response of respiration at a different time scale. Acclimation of leaf respiration to elevated temperatu re regimes i.e., down regulation of respiration under warming, has been described in mid to high latitude plant species, and increasingly, attempts are made to incorporate algorithms for thermal acclimation of respiration into carbon flux models for regi onal (Chen & Zhuang, 2013; Wythers et al ., 2013) and global scales (King et al ., 2006; Atkin et al ., 2008). In these models acclimation is uniformly applied to all growth forms and biomes, even though data are largely lacking from tropical forests. It has been suggested, however, that the temperature stability of tropical forests both intra annually and over geological time scale, might render the species less capable to acclimate to changing temperatures. Observations on high latitude tree species suggest that leaves from different plant functional types may differ in their capacity to acclimate to temperature (Tjoelker et al ., 1999b), but no systematic differences in thermal acclimation were found among different growth forms (Campbell et al ., 2007). Whil e incorporation of acclimation algorithms into flux models has the potential to more
125 realistically predict future respiratory carbon fluxes (King et al ., 2006 ), the current attempts to do so are based on relatively poorly tested assumptions. In an effort t o reconcile the need to address thermal acclimation of respiration in carbon flux models with the lack of consensus on the underlying assumptions, I conducted two studies reported in Chapters 4 and 5. First, using an in situ leaf warming experiment, I test ed whether leaves of tropical forest trees and lianas can acclimate to nighttime warming by several degrees Celsius Tropical forests experience little seasonal temperature fluctuations (Wright et al ., 2009) and the capacity of organisms to respond to dyna mic changes in temperature is expected to be limited in environments that commonly experience minimal temperature variation (Janzen, 1967; Cunningham & Read, 2002; Ghalambor et al ., 2006). I therefore hypothesized that respiration of tropical leaves would not show active acclimation to nighttime warming. I examined changes in the temperature response of leaf respiration in terms of R 25 and Q 10 determined for individual leaves. While Q 10 did not change with warming, a significant pattern of decline of R 25 wi th nighttime leaf temperature during the previous week was found across leaves of three tree species and two liana species. Across species R 25 decreased by an average of 2.9% per degree Celsius of leaf warming above the current ambient temperature There w as no evidence that this down regulation of respiration was due to depletion of respiratory substrate, and thus the results of this experimental warming study suggest that these tropical canopy leaves actively acclimate to warmer nighttime temperature s De spite the consistent pattern of down regulation of respiration, acclimation did not result in complete homeostasis of respiration across temperatures; i.e., respiration rates at the
126 elevated experimental temperature w ere higher than the rate s at the ambien t (control) temperature. Nevertheless, the capacity of tropical trees and lianas to acclimate to warmer nights has the potential to reduce the magnitude of the positive feedback between climate and the carbon cycle in a warming world. This is the first rep ort on thermal acclimation of respiration of tropical trees and lianas, and will be valuable for ecophysiologists and carbon flux modelers alike. In Chapter 5, I report a meta analysis of thermal acclimation of respiration based on data reported in publish ed studies as well as my new data reported in Chapter 4. I compiled 23 7 temperature contrasts of 87 plant species, and tested for differences in acclimation among different biomes, growth forms and experimental duration and degree of warming. Considerable variation existed among species, and within growth forms and biomes, but several patterns could be identified. 1) The larger the temperature difference plants were exposed to, the less complete did respiration acclimate (i.e., less complete homeostasis of respiration) 2) The longer leaves experienced a new temperature the greater the degree of homeostasis of respiration that was achieved 3) When new leaves develop under warmed conditions, the degree of homeostasis is greater than for leaves developed und er one cooler control conditions and transferred to warmer conditions. The available data do not suggest that plants of different growth forms and biomes differ significantly in acclimation of leaf dark respiration. This suggests that with gradual warming the capacity of respiration to acclimate will reduce the potential positive feedback of warming stimulated respiratory CO 2 release across biomes.
127 In conclusion, I found evidence that leaf respiration of tropical forest canopy trees and lianas exhibit s a h igher sensitivity to short term temperature changes (i.e., higher Q 10 ) than often assumed, but that longer term exposure to elevated nighttime temperature results in acclimatory down regulation of respiration similar to what has been shown for plant specie s from thermally more variable regions than the tropics Furthermore, the development of new leaves under elevated temperatures is likely to increase the degree of respiratory homeostasis that acclimation results in. The results presented in this dissertat ion thus suggest that long term warming will not necessarily cause dramatic increases in leaf respiration loads in tropical forest trees and that tropical forests may be more resilient to climate change that previously thought (Huntingford et al ., 2013).
128 APPENDIX A SUPPLEMENTARY TABLES AND FIGURES TO CHAPTER 2 Table A 1. Results of standardized major axis regression analyses that are shown in Fig. 2 3. R 25Area (in mol m 2 s 1 ) is regressed against A max Area (in mol m 2 s 1 ), leaf mass per area (LMA; g m 2 ), nitrogen content per unit area (N Area ; g m 2 ), phosphorus per unit area (P Area ; mg m 2 ) and leaf lifespan (in days). R 25 Mass (in nmol g 1 s 1 ) is regressed against A max Mass (in nmol g 1 s 1 ), LMA, N Mass (mg g 1 ), P Mass (mg g 1 ) and leaf lifespan. Respiration Trait Regression Trait Elevation Slope R 2 P R 25 Area A max Area 1.143 1.012 0.337 0.002 LMA 1.497 0.676 0.205 0.020 N Area 0.261 0.859 0.237 0.012 P Area 1.582 0.762 0.419 0.001 Lifespan 1.434 0.657 0.057 0.273 R 25 Mass A max Mass 1.330 1.090 0.474 <0.001 LMA 2.949 0.956 0.448 <0.001 N Mass 0.658 1.075 0.300 0.004 P Mass 1.978 1.076 0.315 0.005 Lifespan 2.813 0.780 0.195 0.035
129 Table A 2. Study species and families, plant functional types (PFT); early successional (ES) mid successional (MS), late successional (LS) tree species; lianas (L). Q 10 based on R M ass temperature response curves and its 95% confidence intervals; photosynthetic capacity on a leaf mass basis ( A max M ass ), nitrogen (N) and phosphorus (P) content pe r unit leaf area, R per unit N, and the ratio of R at 25C and photosynthetic capacity ( R 25 / A max ) are shown. Species Family PFT Q 10 Mass (95% CI) A max Mass N Area P Area R 25 /N R 25 /A max nmol g 1 s 1 g m 2 g m 2 mol (g N) 1 s 1 ) Albizia guachapele (Kunth) Harms Fabaceae ES 3.47 (1.27 7.45) 142 2.9 0.09 0.26 0.06 Annona spraguei Saff. Annonaceae ES 2.94 (2.28 3.83) 152 2.5 0.15 0.38 0.07 Cecropia peltata L. Urticaceae ES 3.16 (1.23 8.15) 199 2.8 0.17 0.45 0.06 Pittoniotis trichantha Griseb. Rubiaceae ES 2.60 (1.56 4.35) 199 1.6 0.10 0.76 0.09 Astronium graveolens Jacq Anacardiaceae MS 1.87 (1.30 2.67) 145 2.6 0.18 0.46 0.10 Castilla elastica var.costaricana, (Liebm.) C.C. Berg Moraceae MS 2.89 (2.18 3.84) 187 2.9 0.18 0.39 0.06 F icus insipida Willd. Moraceae MS 1.82 (1.11 2.98) 160 3.4 0.25 0.43 0.06 Luehea seemannii Triana & Planch. Tiliaceae MS 1.78 (1.36 2.37) 190 2.5 0.16 0.45 0.06 Pseudobombax septenatum (Jacq.) Dugand Malvaceae MS 1.70 (1.28 2.31) 159 2.3 nd 0.50 0.07 S pondias mombin L. Anacardiaceae MS 1.91 (1.51 2.23) 162 2.8 0.13 0.54 0.09 Zuelania guidonia (Sw.) Britt. & Millsp. Flacourtiaceae MS 2.57 (1.49 4.33) 140 2.5 nd 0.55 0.08 Anacardium excelsum (Bertero & Balb.ex Kunth) Skeels Anacardiaceae LS 1.93 (1.68 2 .20) 126 1.5 0.15 0.80 0.09 Chrysophyllum cainito L. Sapotaceae LS 2.04 (1.72 2.42) 121 2.1 0.12 0.39 0.06 Amphilophium paniculatum (L. ) kunth Bignoniaceae L 2.21 (1.67 2.93) 166 1.8 0.11 0.45 0.18 Aristolochia tonduzii O.C. Schmidt Aristolochiaceae L 2 .45 (1.71 3.51) 164 2.5 0.12 0.41 0.08 Bonamia trichantha Hallier f Convolvulaceae L 1.88 (1.04 3.51) 151 2.2 0.12 0.36 0.08 Cissus erosa Rich Vitaceae L 1.96 (0.81 3.40) 301 1.8 nd 0.71 0.08 Combretum fruticosum (Loefl.) Stuntz Combretaceae L 3.33 ( 1.99 5.59) 235 1.7 0.12 0.63 0.06 Forsteronia myriantha Donn. Sm. Apocynaceae L 1.65 (0.66 4.21) 220 1.9 0.11 0.51 0.07 Gouania lupuloides (L.) Urb. Rhamnaceae L 1.35 (0.75 2.42) 266 1.9 0.09 0.47 0.07 Mikania leiostachya Benth. Asteraceae L 2.55 (1.05 6.03) 167 1.4 0.08 0.51 0.07 Phryganocydia corymbosa (Vent.) Bureau ex K. Schum Bignoniaceae L 1.89 (1.36 2.61) 125 3.6 0.15 0.49 0.15 Serjania mexicana (L ) willd. Sapindaceae L 2.25 (1.31 3.84) 139 2.2 0.14 0.45 0.08 Stigmaphyllon lindenianum A. Ju ss. Malpighiaceae L 1.60 (1.24 2.14) 204 2.2 0.11 0.49 0.06 Trichostigma octandrum (L.) H. Walt. Bignoniaceae L 1.99 (1.30 3.66) 263 2.3 0.18 0.49 0.08 Vitis tillifolia Humb. & Bonpl. ex Roem. & Schult. Vitaceae L 3.07 (1.26 7.47) 197 1.4 0.09 0.68 0.07
130 Figure A 1. Correlations between Q 10 values, determined from species level temperature response curves of area normalized respiration data ( A F ) and mass normalized respiration data ( G L ), and other leaf traits. A,G) Respiration at 25C ( R 25 ) B,H ) P hotosynthetic capacity ( A max ) C,I) L eaf mass per unit area (LMA) D,J) L eaf nitrogen content (N) E,K) L eaf phosphorus content (P) F,L) L eaf lifespan. Dashed lines indicate non significant correlations.
131 APPENDIX B REFERENCES USED IN THE META ANALYSIS IN CHAPTER 5 Studies used in the meta analysis of studies of thermal acclimation of leaf dark respiration reported in C hapter 5 The numbers correspond to the reference numbers in Table 5 1. 1. Arnone JA, Krner C. 1997. Temperature adaptation and acclimati on potential of leaf dark respiration in two species of Ranunculus from warm and cold habitats. Arctic and Alpine Research 29 : 122 125. 2. Atkin OK, Holly C, Ball MC. 2000. Acclimation of snow gum ( Eucalyptus pauciflora ) leaf respiration to season and diurna l variations in temperature: the importance of changes in the capacity and temperature sensitivity of respiration. Plant, Cell & Environment 23 : 15 26. 3. Billings WD, Godfrey PJ, Chabot BF, Bourque DP. 1971. Metabolic acclimation to temperature in arctic an d alpine ecotypes of Oxyria digyna Arctic and Alpine Research 3 : 277 289. 4. Bolstad PV, Reich P, Lee T. 2003. Rapid temperature acclimation of leaf respiration rates in Quercus alba and Quercus rubra Tree Physiology 23 : 969 976. 5. Bruhn D, Egerton JJG, Lov eys BR, Ball MC. 2007. Evergreen leaf respiration acclimates to long term nocturnal warming under field conditions. Global Change Biology 13 : 1216 1223. 6. Campbell C, Atkinson L, Zaragoza Castells J, Lundmark M, Atkin O, Hurry V. 2007. Acclimation of photos ynthesis and respiration is asynchronous in response to changes in temperature regardless of plant functional group. New Phytologist 176 : 375 389. 7. Chabot B, Billings W. 1972. Origin and ecology of the Sierran alpine vegetation. Ecological Monographs 42 : 163 199. 8. Cheesman AW, Winter K. 2013. Elevated night time temperatures increase growth in seedlings of two tropical pioneer tree species. New Phytologist 197 : 1185 1192. 9. Chi Y, Xu M, Shen R, Yang Q, Huang B, Wan S. 2013. Acclimation of foliar respirati on and photosynthesis in response to experimental warming in a temperate steppe in northern China. PLoS ONE 8 : e56482. 10. Dillaway DN, Kruger EL. 2011. Leaf respiratory acclimation to climate: comparisons among boreal and temperate tree species along a lati tudinal transect. Tree Physiology 31 : 1114 1127. 11. Krause GH, Cheesman AW, Winter K, Krause B, Virgo A. 2013. Thermal tolerance, net CO 2 exchange and growth of a tropical tree species, Ficus insipida cultivated at elevated daytime and nighttime temperature s. Journal of Plant Physiology 170 : 822 827.
132 12. Larigauderie A, Krner C. 1995. Acclimation of leaf dark respiration to temperature in alpine and lowland plant species. Annals of Botany 76 : 245 252. 13. Lee TD, Reich PB, Bolstad PV. 2005. Acclimation of leaf re spiration to temperature is rapid and related to specific leaf area, soluble sugars and leaf nitrogen across three temperate deciduous tree species. Functional Ecology 19 : 640 647. 14. Loveys BR, Atkinson LJ, Sherlock DJ, Roberts RL, Fitter AH, Atkin OK. 2003 Thermal acclimation of leaf and root respiration: an investigation comparing inherently fast and slow growing plant species. Global Change Biology 9 : 895 910. 15. Ow LF, Griffin KL, Whitehead D, Walcroft AS, Turnbull MH. 2008a. Thermal acclimation of leaf respiration but not photosynthesis in Populus deltoides nigra New Phytologist 178 : 123 134. 16. Ow LF, Whitehead D, Walcroft AS, Turnbull MH. 2008b. Thermal acclimation of respiration but not photosynthesis in Pinus radiata Functional Plant Biology 35 : 44 8 461. 17. Ow LF, Griffin KL, Whitehead D, Walcroft AS, Turnbull MH. 2010. Seasonal variation in foliar carbon exchange in Pinus radiata and Populus deltoides : respiration acclimates fully to changes in temperature but photosynthesis does not. Global Change B iology 16 : 288 302. 18. Rook DA. 1969. The influence of growing temperature on photosynthesis and respiration of Pinus radiata seedlings. New Zealand Journal of Botany 7 : 43 55. 19. Slot M, Rey Snchez AC, Winter K, Kitajima K Thermal acclimation of leaf resp iration in tropical canopy trees and lianas. (Manuscript in preparation) 20. Smith EM, Hadley EB. 1974. Photosynthetic and respiratory acclimation to temperature in Ledum groenlandicum populations. Arctic and Alpine Research 6 : 13 27. 21. Teskey RO, Will RE. 199 9. Acclimation of loblolly pine ( Pinus taeda ) seedlings to high temperatures. Tree Physiology 19 : 519 525. 22. Tjoelker, MG, Oleksyn J, Reich PB. 1999. Acclimation of respiration to temperature and CO 2 in seedlings of boreal tree species in relation to plant size and relative growth rate. Global Change Biology 5 : 679 691. 23. Tjoelker, MG, Oleksyn J, Reich PB, ytkowiak R. 2008. Coupling of respiration, nitrogen, and sugars underlies convergent temperature acclimation in Pinus banksiana across wide ranging sites and populations. Global Change Biology 14 : 782 797. 24. Tjoelker, MG, Oleksyn J, Lorenc Plucinska G, Reich PB. 2009. Acclimation of respiratory temperature responses in northern and southern populations of Pinus banksiana New Phytologist 181 : 218 229. 25. Tobi essen PL. 1976. Thermal acclimation of dark respiration in coastal and desert populations of Isomeris arborea American Midland Naturalist 96 : 462 567.
133 26. Wang K, Kellomki S, Laitinen K. 1995 Effects of needle age, long term temperature and CO 2 treatments on the photosynthesis of Scots pine. Tree Physiology 15 : 211 218. 27. Wertin TM, McGuire MA, van Iersel M, Ruter JM, Teskey RO. 2012. Effects of elevated temperature and [CO 2 ] on photosynthesis, leaf respiration, and biomass accumulation of Pinus taeda seedli Canadian Journal of Forest Research 42 : 943 957. 28. Xiong FS, Mueller EC, Day TA. 2000. Photosynthetic and respiratory acclimation and growth response of Antarctic vascular plants to contrasti ng temperature regimes. American Journal of Botany 87 : 700 710. 29. Zha TS, Wang KY, Ryypp A, Kellomki S. 2002. Impact of needle age on the response of respiration in Scots pine to long term elevation of carbon dioxide concentration and temperature. Tree Ph ysiology 22 : 1241 1248. 30. Zhou Z, Liu X, Wallace LL, Luo Y. 2007. Photosynthetic and respiratory acclimation to experimental warming for four species in a tallgrass prairie ecosystem. Journal of Integrative Plant Biology 49 : 270 281.
134 LIST OF REFERENCES Ahls trm A, Schurgers G, Arneth A, Smith B. 2012. Robustness and uncertainty in terrestrial ecosystem carbon response to CMIP5 climate change projections. Environmental Research Letters 7 : 044008. Alvarez Aorve MY, Quesada M, Snchez Azofeifa GA, Avila Caba dilla LD, Gamon JA. 2012. Functional regeneration and spectral reflectance of trees during succession in a highly diverse tropical dry forest ecosystem. American Journal of Botany 99 : 816 826. Alward RD, Detling JK, Milchunas DG. 1999. Grassland vegetatio n changes and nocturnal global warming. Science 283 : 229 231 Amthor JS. 1984. The role of maintenance respiration in plant growth. Plant, Cell & Environment 7 : 561 569. Armstrong AF, Logan DC, Tobin AK, O'Toole P, Atkin OK. 2006. Heterogeneity of plant mitochondrial responses underpinning respiratory acclimation to the cold in Arabidopsis thaliana leaves. Plant, Cell & Environment 29 : 940 949. Arneth A, Mercado L, Kattge J, Booth BB. 2012. Future challenges of representing land processes in studies on l and atmosphere interactions. Biogeoscience 9 : 3587 3599. Arnone JA, Krner C. 1997. Temperature adaptation and acclimation potential of leaf dark respiration in two species of Ranunculus from warm and cold habitats. Arctic and Alpine Research 29 : 122 125 Atkin OK, Atkinson LJ, Fisher RA, Campbell CD, Zaragoza Castells J, Pitchford JW, Woodward FI, Hurry V. 2008. Using temperature dependent changes in leaf scaling relationships to quantitatively account for thermal acclimation of respiration in a coupled global climate vegetation model. Global Change Biology 14 : 2709 2726. Atkin OK, Bruhn D, Tjoelker MG. 2005. Response of plant respiration to changes in temperature: mechanisms and consequences of variations in Q 10 values and acclimation. In: Lambers H, R ibas Carbo M, eds. Plant r espiration: f rom c ell to e cosystem Dordrecht, the Netherlands: Springer, 95 135. Atkin OK, Evans JR, Ball MC, Lambers H, Pons TL. 2000. Leaf respiration of snow gum in the light and dark; interactions between temperature and irr adiance. Plant Physiology 122 : 915 923. Atkin OK, Holly C, Ball MC. 2000. Acclimation of snow gum ( Eucalyptus pauciflora ) leaf respiration to season and diurnal variations in temperature: the importance of changes in the capacity and temperature sensitivi ty of respiration. Plant, Cell & Environment 23 : 15 26.
135 Atkin OK, Scheurwater I, Pons TL. 2007. Respiration as a percentage of daily photosynthesis in whole plants is homeostatic at moderate, but not high, growth temperatures. New Phytologist 174 : 367 380 Atkin OK, Tjoelker MG. 2003. Thermal acclimation and the dynamic response of plant respiration to temperature. Trends in Plant Science 8 : 343 351. Avalos G, Mulkey SS. 1999. Seasonal changes in liana cover in the upper canopy of a neotropical dry fores t. Biotropica 31 : 186 192. Azcn Bieto J, Osmond CB. 1983. Relationship between photosynthesis and respiration. The effect of carbohydrate status on the rate of CO 2 production by respiration in darkened and illuminated wheat leaves. Plant Physiology 71 : 5 74 581. Baker TR, Phillips OL, Malhi Y, Almeida S, Arroyo L, Di Fiore A, Erwin T, Higuchi N, Killeen TJ, Laurance SG, et al. 2004. Increasing biomass in Amazonian forest plots. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 359 : 353 365. Beer C, Reichstein M, Tomelleri E, Ciais P, Jung M, Carvalhais N, Rdenbeck C, Altaf Arain M, Baldocchi D, Bonan GB et al 2010. Terrestrial gross carbon dioxide uptake: global distribution and covariation with climate. Science 32 9 : 834 838. Billings WD, Godfrey PJ, Chabot BF, Bourque DP. 1971. Metabolic acclimation to temperature in arctic and alpine ecotypes of Oxyria digyna Arctic and Alpine Research 3 : 277 289. Bolstad PV, Mitchell K, Vose, JM. 1999. Foliar temperature respi ration response functions for broad leaved tree species in the southern Appalachians. Tree Physiology 19 : 871 878. Bolstad PV, Reich P, Lee T. 2003. Rapid temperature acclimation of leaf respiration rates in Quercus alba and Quercus rubra Tree Physiology 23 : 969 976. Bonan GB, Levis S, Sitch S, Vertenstein M, Oleson KW. 2003. A dynamic global vegetation model for use with climate models: concepts and description of simulated vegetation dynamics. Global Change Biology 9 : 1543 1566. Booth BBB, Jones CD, Collins M, Totterdell IJ, Cox PM, Sitch S, Huntingford C, Betts RA, Harris GR, Lloyd J. 2012. High sensitivity of future global warming to land carbon cycle processes. Environmental Research Letters 7 : 024002. Bruhn D, Egerton JJG, Loveys BR, Ball MC. 20 07. Evergreen leaf respiration acclimates to long term nocturnal warming under field conditions. Global Change Biology 13 : 1216 1223.
136 Campbell C, Atkinson L, Zaragoza Castells J, Lundmark M, Atkin O, Hurry V. 2007. Acclimation of photosynthesis and respir ation is asynchronous in response to changes in temperature regardless of plant functional group. New Phytologist 176 : 375 389. Cavaleri MA, Oberbauer SF, Ryan MG. 2008 Foliar and ecosystem respiration in an old growth tropical rain forest. Plant, Cell & Environment 31 : 473 483. Chambers JQ, Tribuzy ES, Toledo LC, Crispim BF, Higuchi N, Dos Santos J, Arajo AC, Kruijt B, Nobre AD, Trumbore SE. 2004. Respiration from a tropical forest ecosystem: partitioning of sources and low carbon use efficiency. Ecol ogical Applications 14 : S72 S88. Cheesman AW, Winter K. 2013. Elevated night time temperatures increase growth in seedlings of two tropical pioneer tree species. New Phytologist 197 : 1185 1192. Chen M, Zhuang, Q. 2013. Modeling temperature acclimation e ffects on carbon dynamics of forest ecosystems in the conterminous United States. Tellus B 65 : 19156. Clark DA, Piper SC, Keeling CD, Clark DB. 2003. Tropical rain forest tree growth and atmospheric carbon dynamics linked to interannual temperature variat ion during 1984 2000. Proceedings of the National Academy of Sciences USA 100 : 5852 5857. Clark DB, Clark DA, Oberbauer SF. 2010. Annual wood production in a tropical rain forest in NE Costa Rica linked to climatic variation but not to increasing CO 2 Glo bal Change Biology 16 : 747 759. Clark DB, Clark DA, Rich PM, Weiss S, Oberbauer SF. 1996. Landscape scale evaluation of understory light and canopy structure: methods and application in a neotropical lowland rain forest. Canadian Journal of Forest Researc h 13 : 747 757. Clark DB, Olivas PC, Oberbauer SF, Clark DA, Ryan MG. 2008 First direct landscape scale measurement of tropical rain forest Leaf Area Index, a key driver of global primary productivity. Ecology Letters 11 : 163 172. Collier DE. 1996. No di fference in leaf respiration rates among temperate, subarctic, and arctic species grown under controlled conditions. Canadian Journal of Botany 74 : 317 320. Condit R, Engelbrecht BMJ, Pino D, Prez R, Turner BL. 2013. Species distributions in response to individual soil nutrients and seasonal drought across a community of tropical trees. Proceedings of the National Academy of Sciences USA 110 : 5064 5068. Condit R. 1998. Tropical forest census plots Berlin, Germany: Springer Verlag & Georgetown, TX, USA: R. G. Landes Company.
137 Corlett RT. 2011. Impacts of warming on tropical lowland rainforests. Trends in Ecology and Evolution 27 : 145 150. Corlett RT. 2012. Climate change in the tropics: the end of the world as we know it? Biological Conservation 151 : 22 5. Cox PM, Betts RA, Collins M, Harris PP, Huntingford C, Jones CD. 2004. Amazonian forest dieback under climate carbon cycle projections for the 21st century. Theoretical and Applied Climatology 78 : 137 156. Cox PM, Betts RA, Jones CD, Spall SA, Totterd ell IJ. 2000. Acceleration of global warming due to carbon cycle feedbacks in a coupled climate model. Nature 408 : 184 187. Cox PM, Pearson D, Booth BB, Friedlingstein P, Huntingford C, Jones CD, Luke CM. 2013. Sensitivity of tropical carbon to climate ch ange constrained by carbon dioxide variability. Nature 494 : 341 344. Cramer W, Bondeau A, Schaphoff S, Lucht W, Smith B, Sitch S. 2004. Tropical forests and the global carbon cycle: impacts of atmospheric carbon dioxide, climate change and rate of defores tation. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences 359 : 331 343. Cramer W, Bondeau A, Woodward FI, Prentice IC, Betts RA, Brovkin V, Cox M, Fisher V, Foley JA, Friend AD et al 2001. Global response of terrest rial ecosystem structure and function to CO 2 and climate change: results from six dynamic global vegetation models. G lobal Change Biology 7 : 357 373. Cunningham SC, Read J. 2002. Comparison of temperate and tropical rainforest tree species: photosynthetic responses to growth temperature. Oecologia 133 : 112 119. Cunningham SC, Read J. 2003a. Do temperate rainforest trees have a greater ability to acclimate to changing temperatures than tropical rainforest trees? New Phytologist 157 : 55 64. Cunningham S C Read J. 2003b. Comparison of temperate and tropical rainforest tree species: growth responses to temperature. Journal of Biogeography 30 : 143 153 Diffenbaugh NS, Scherer M. 2011 Observational and model evidence of global emergence of permanent, unpreced ented heat in the 20th and 21st centuries. Climatic Change 107 : 615 624. Dillon ME, Wang G, Huey RB. 2010. Global metabolic impacts of recent climate warming. Nature 467 : 704 706.
138 Domingues TF, Berry JA, Martinelli LA, Ometto J, Ehleringer JR. 2005. Para meterization of canopy structure and leaf level gas exchange for an eastern Amazonian tropical rain forest (Tapajs National Forest, Para, Brazil). Earth Interactions 9 : 1 23. Domingues TF, Martinelli LA, Ehleringer JR. 2007. Ecophysiological traits of pl ant functional groups in forest and pasture ecosystems from eastern Amazonia, Brazil. Plant Ecology 193 : 101 112. Doughty CE, Goulden ML. 2008. Are tropical forests near a high temperature threshold? Journal of Geophysical Research Biogeosciences 113 : G 00B07. Doughty CE. 2011. An in situ leaf and branch warming experiment in the Amazon. Biotropica 43 : 658 665. Dubois M, Gilles K, Hamilton J, Rebers P, Smith F. 1956. Calorimetric method for determination of sugars and related substances. Analytical Chem istry 28 : 350 356. Easterling DR, Horton B, Jones PD, Peterson TC, Karl TR, Parker DE, Salinger MJ, Razuvayev V, Plummer N, Jamason P et al. 1997. Maximum and minimum temperature trends for the globe. Science 277 : 346 366. Feeley KJ, Wright SJ, Supardi M NN, Kassim AR, Davies SJ. 2007. Decelerating growth in tropical forest trees. Ecology Letters 10 : 461 469. Florez Sarasa I, Araujo WL, Wallstrom SV, Rasmusson AG, Fernie AR, Ribas Carb M. 2012. Light responsive metabolite and transcript levels are mainta ined following a dark adaptation period in leaves of Arabidopsis thaliana New Phytologist 195 : 136 148. Foley JA, Prentice CI, Ramankutty N, Levis S, Pollard D, Sitch S, Haxeltine A. 1996. An integrated biosphere model of land surface processes, terrestr ial carbon balance, and vegetation dynamics. Global Biogeochemical Cycles 10 : 603 628. Frantz JM, Cometti NN, Bugbee B. 2004. Night temperature has a minimal effect on respiration and growth in rapidly growing plants. Annals of Botany 94 : 155 166. Friend AD, Stevens AK, Knox RG, Cannell MGR. 1997. A process based, terrestrial biosphere model of ecosystem dynamics (Hybrid v3.0). Ecological Modelling 95 : 249 287. Galbraith D, Levy PE, Sitch S, Huntingford C, Cox P, Williams M, Meir P. 2010. Multiple mechan isms of Amazonian forest biomass losses in three dynamic global vegetation models (DGVMs) under climate change. New Phytologist 187 : 647 665.
139 Ghalambor CK, Huey RB, Martin PR, Tewksbury JJ, Wang G. 2006. Are mountain passes higher in the tropics? Janzen's hypothesis revisited Integrative and Comparative Biology 46 : 5 17. Goulden ML, Munger JW, Fan S M, Daube BC, Wofsy SC. 1996. Measurements of carbon sequestration by long term eddy covariance: methods and a critical evaluation of accuracy. Global Change Biology 2 : 169 182. Griffin KL, Turnbull MH, Murthy R, Lin GH, Adams J, Farnsworth B, Mahato T, Bazin G, Potasnak M. Berry JA. 2002a. Leaf respiration is differentially affected by leaf vs. stand level night time warming. Global Change Biolog y 8 : 479 485 Griffin KL, Turnbull MH, Murthy R. 2002b. Canopy position affects the temperature response of leaf respiration in Populus deltoides New Phytologist 154 : 609 619. Hansen J, Sato M, Ruedy R, Lo K, Lea DW, Medina Elizade M. 2006. Global temperature chang e. Proceedings of the National Academy of Sciences USA 103 : 14288 14293. Hubbell SP, Condit R, Foster RB. 2005. Barro Colorado forest census plot data. [WWW document] URL https://ctfs.ar narb.harvard.edu/webatlas/datasets/bci [accessed 22 April 2013]. Huete AR, Didan K, Shimabukuro YE, Ratana P, Saleska SR, Hutyra LR, Yang W, Nemani RR, Myneni R. 2006. Amazon rainforests green up with sunlight in dry season. Geophysical Research Letters 33 : L06405. Huntingford C, Zelazowski P, Galbraith D, Mercado LM, Sitch S, Fisher R, Lomas M, Walker AP, Jones CD, Booth BBB et al 2013. Simulated resilience of tropical rainforests to CO 2 induced climate change. Nature Geoscience 6 : 268 273. Huston MA, Wolverton S. 2009. The global distribution of net primary production: resolving the paradox. Ecological Monographs 79 : 343 77. Hve K, Bichele I, Rasulov B, Niinemets 2011. When it is too hot for photosynthesis: heat induced instability of photosynthe sis in relation to respiratory burst, cell permeability changes and H 2 O 2 formation. Plant, Cell & Environment 34 : 113 126. Ingwell LL, Wright SJ, Becklund KK, Hubbell SP, Schnitzer SA. 2010. The impact of lianas on 10 years of tree growth and mortality on Barro Colorado Island, Panama. Journal of Ecology 98 : 879 887. IPCC. 2007. Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL, eds. Climate change 2007: the physical science basis. Contribution of Working Group I to the fourth assessment report of the Intergovernmental Panel on Climate Change. Cambridge, UK & New York, NY, USA: Cambridge University Press.
140 Janzen DH. 1967. Why mountain passes are higher in the tropics. American Naturalist 101 : 233 249. Kalacska M, Bohlman S, S anchez Azofeifa GA, Castro Esau K, Caelli T. 2007. Hyperspectral discrimination of tropical dry forest lianas and trees: c omparative data reduction approaches at the leaf and canopy levels. Remote Sensing of Environment 109 : 406 415. King AW, Gunderson CA Post WM, Weston DJ Wullschleger SD. 2006. Plant respiration in a warmer world. Science 312 : 536 537. Kitajima K, Mulkey SS, Wright SJ. 2005. Variation in crown light utilization characteristics among tropical canopy trees. Annals of Botany 95 : 535 547. Kitajima K, Poorter L. 2008. Functional basis for resource niche partitioning by tropical tree s. In: Carson WP, Schnitzer SA, eds. Tropical forest community ecology Oxford UK: Wiley Blackwell, 160 181. Kitajima K. 1994. Relative importance of photosynt hetic traits and allocation patterns as correlates of seedling shade tolerance of 13 tropical trees. Oecologia 98 : 419 428. Kozlowski TT. 1992. Carbohydrate sources and sinks in woody plants. Botanical Review 58 : 107 222. Krause GH, Cheesman AW, Winter K Krause B, Virgo A. 2013. Thermal tolerance, net CO 2 exchange and growth of a tropical tree species, Ficus insipida cultivated at elevated daytime and nighttime temperatures. Journal of Plant Physiology 170 : 822 827. Krause GH, Winter K, Krause B, Jahns P, Garca M, Aranda J, Virgo A. 2010. High temperature tolerance of a tropical tree, Ficus insipida : methodological reassessment and climate change considerations. Functional Plant Biology 37 : 890 900. Krmer S. 1995. Respiration during photosynthesis. A nnual Review of Plant Physiology and Plant Molecular Biology 46 : 45 70. Kukla G, Karl TR. 1993. Nighttime warming and the greenhouse effect. Environmental Science & Technology 27 : 1468 1474. Larigauderie A, Krner C. 1995. Acclimation of leaf dark respir ation to temperature in alpine and lowland plant species. Annals of Botany 76 : 245 252. Lavigne MB, Ryan MG, Anderson DE, Baldocchi DD, Crill PM, Fitzjarrald DR, Goulden ML, Gowe r T, Massheder JM, McCaughey JH et al 1997. Comparing nocturnal eddy covaria nce measurements to estimates of ecosystem respiration made
141 by scaling chamber measurements at six coniferous boreal sites. Journal of Geophysical Research Atmospheres 102 : 28977 28985. Lee TD, Reich PB, Bolstad PV. 2005. Acclimation of leaf respiration t o temperature is rapid and related to specific leaf area, soluble sugars and leaf nitrogen across three temperate deciduous tree species. Functional Ecology 19 : 640 647. Lefsky MA, Cohen WB, Parker GG, Harding DJ. 2002. Lidar remote sensing for ecosystem studies. BioScience 52 : 19 30. Levy PE, Cannell MGR, Friend AD. 2004. Modelling the impact of future changes in climate, CO 2 concentration and land use on natural ecosystems and the terrestrial carbon sink. Global Environmental Change 14 : 21 30. Lewis SL Lopez Gonzalez G, Sonk B, Affum Baffoe K, Baker TR, Ojo LO, Phillips OL, R eitsma JM, White L, Comiskey JA et al 2009. Increasing carbon storage in intact African tropical forests. Nature 457 : 1003 1006. Lin D, Xia J, Wan S. 2010. Climate warming and b iomass accumulation of terrestrial plants: a meta analysis. New Phytologist 188 : 187 198. Lloyd J, Farquhar GD. 2008. Effects of rising temperatures and [CO 2 ] on the physiology of tropical forest trees. Philosophical Transactions of the Royal Society B B iological Sciences 363 : 1811 1817. Loescher HW, Oberbauer SF, Gholz HL, Clark DB. 2003. Environmental controls on net ecosystem level carbon exchange and productivity in a Central American tropical wet forest. Global Change Biology 9 : 396 412. Loveys BR, Atkinson LJ, Sherlock DJ, Roberts RL, Fitter AH, Atkin OK. 2003. Thermal acclimation of leaf and root respiration: an investigation comparing inherently fast and slow growing plant species. Global Change Biology 9 : 895 910. Luo YQ. 2007. Terrestrial car bon cycle feedback to climate warming. Annual Review of Ecology, Evolution, and Systematics 38 : 683 712. Malhi PL Grace J 2000. Tropical forests and atmospheric carbon dioxide Trends in Ecology and Evolution 15 : 332 337 Malhi Y, C Arago LEO, Metcalf e DB, Paiva R, Quesada CA, Almeida S, Anderson L, Brando P, Chambers JQ, da Costa ACL et al 2009. Comprehensive assessment of carbon productivity allocation and storage in three Amazonian forests. Global Change Biology 15 : 1255 1274.
142 Malhi Y, Doughty C, Galbraith D. 2011. The allocation of ecosystem net primary productivity in tropical forests. Philosophical Transactions of the Royal Society B 366 : 3225 3245. Malhi Y. 2012. The productivity, metabolism and carbon cycle of tropical forest vegetation. Jour nal of Ecology 100 : 65 75. Mallows CL. 1973. Some comments on Cp. Technometrics 15 : 661 675. McNairn RB, Currier HB. 1968. Translocation blockage by sieve plate callose. Planta 82 : 369 380. Meehl GA, Tebaldi C. 2004. More intense, more frequent, and lon ger lasting heat waves in the 21st Century. Science 305 : 994 997. Meir P, Grace J, Miranda AC. 2001. Leaf respiration in two tropical rainforests: constraints on physiology by phosphorus, nitrogen and temperature. Functional Ecology 15 : 378 387. Messier J, McGill BJ, Lechowicz MJ. 2010 How do traits vary across ecological scales? A case for trait based ecology. Ecology Letters 13 : 838 848. Metcalfe DB, Lobo do Vale R, Chaves MM, Maroco JP, C Arago LEO, Malhi Y, Da Costa AL, Braga AP, Gonalves PL, De A thaydes J et al 2010. Impacts of experimentally imposed drought on leaf respiration and morphology in an Amazon rain forest. Functional Ecology 24 : 524 533. Metcalfe DB, Meir P, Arago LEOC, Lobo do Vale R, Galbraith D, Fisher RA, Chaves MM, Maroco JP, d a Costa ACL, de Almeida SS et al 2010. Shifts in plant respiration and carbon use efficiency at a large scale drought experiment in the eastern Amazon. New Phytologist 187 : 608 621. Miller A. 2002. Subset selection in regression New York, USA: Chapman & Hall. Miranda EJ, Vourlitis GL, Filho NP, Priante PC, Campelo Jr JH, SanchesSuli G, Fritzen CL, De Almeida Lobo F, Shiraiwa S. 2005. Seasonal variation in the leaf gas exchange of tropical forest trees in the rain forest savanna transition of the souther n Amazon basin. Journal of Tropical Ecology 21 : 451 460. Noguchi K, Terashima I. 1997 Different regulation of leaf respiration between Spinacia oleracea a sun species, and Alocasia odora a shade species. Physiologia Plantarum 101 : 1 7. Noguchi K. 2005 Effects of light intensity and carbohydrate status on leaf and root respiration. In: Lambers H, Ribas Carb M, eds. Plant r espiration: f rom cell to ecosystem Dordrecht, the Netherlands: Springer, 63 83.
143 Ow LF, Griffin KL, Whitehead D, Walcroft AS, Turn bull MH. 2008a. Thermal acclimation of leaf respiration but not photosynthesis in Populus deltoides nigra New Phytologist 178 : 123 134. Ow LF, Whitehead D, Walcroft AS, Turnbull MH. 2008b. Thermal acclimation of respiration but not photosynthesis in Pi nus radiata Functional Plant Biology 35 : 448 461. Ow LF, Griffin KL, Whitehead D, Walcroft AS, Turnbull MH. 2010. Seasonal variation in foliar carbon exchange in Pinus radiata and Populus deltoides : respiration acclimates fully to changes in temperature but photosynthesis does not. Global Change Biology 16 : 288 302. Parker GG, Smith AP, Hogan KP. 1992. Access to the upper forest canopy with a large tower crane. Bioscience 42 : 664 670. Pearcy RW. 1987. Photosynthetic gas exchange responses of Australian tropical forest trees in canopy, gap and understory micro environments. Functional Ecology 1 : 169 178. Penning de Vries FWT. 1975. Cost of maintenance processes in plant cells. Annals of Botany 39 : 77 92. Prez R, Condit R. Tree a tlas of Panama. [WWW doc ument] URL http://ctfs.arnarb. harvard.edu/webatlas/maintreeatlas.ph p [accessed 12 Sept ember 2012 ] Phillips OL, Malhi Y, Higuchi N, Laurance WF, Nuez VP, Vasquez MR, Laurance SG Ferreira LV, Stern M, Brown S et al 1998. Changes in the carbon balance of tropical forests: evidence from long term plots. Science 282 : 439 442. Phillips OL, Vsquez Martnez R, Arroyo L, Baker TR, Killeen T, Lewis SL, Malhi Y, Mendoza AM, Neill D, Varg as PN et al 2002. Increasing dominance of large lianas in Amazonian forests. Nature 418 : 770 774. Pons TL, Welschen RAM. 2003. Midday depression of net photosyntheis in the tropical rainforest tree Eperua grandiflora : contributions of stomatal and intern al conductances, respiration and Rubisco functioning. Tree Physiology 23 : 937 947. Pyl E T, Piques M, Ivakov A, Schulze W, Ishihara H, Stitt M Sulpice R. 2012. Metabolism and growth in Arabidopsis depend on the daytime temperature but are temperature comp ensated against cool nights. The Plant Cell 24 : 2443 2469. Quinn PJ. 1989. The phase behaviour of membrane lipids and the organisation of the photosynthetic membrane. In: Biacs PA, Gruiz K, Kremmer T, eds. Biological role of plant lipids NY US: Springer Verlag, 209 213.
144 R Development Core Team. 2011. R: a language and environment for statistical computing. R f oundation for s tatistical c omputing, Vienna, Austria. Raison JK, Berry JA, Armond PA, Pike CS, Turner NC, Kramer PJ. 1980. Membrane properties in relation to the adaptation of plants to temperature stress. In: Turner NC, Kramer PJ, eds. Adaptation of plants to water and high temperature stress NY USA: John Wiley & Sons, 261 273. Reed SC, Wood TE, Cavaleri MA. 2012. Tropical forests in a warming world. New Phytologist 193 : 27 29. Reich PB, Ellsworth DS, Uhl C. 1995. Leaf carbon and nutrient assimilation and conservation in species of differing successional status in an oligotrophic Amazonian forest. Functional Ecology 9 : 65 76. Reich PB, Walters MB, Ellsworth DS, Vose JS, Volin JC, Gresham C, Bowman WD. 1998. Relationships of leaf dark respiration to leaf nitrogen, specific leaf area and leaf life span: a test across biomes and functional groups. Oecologia 114 : 471 482. Ryan MG. 1991. Effects of climate change on plant respiration. Ecological Applications 1 : 157 167. Saugier B, Roy J, Mooney HA. 2001. Estimations of global terrestrial productivity: converging toward a single number? In: Roy J, Saugier B, Mooney HA, eds. Terrestrial global produc tivity. NY USA: Academic Press 543 557. Schnitzer SA, Bongers F. 2011. Increasing liana abundance and biomass in tropical forests: emerging patterns and putative mechanisms. Ecology Letters 14 : 397 406. Sharp RE, Matthews MA, Boyer JS. 1984. Kok effect and the quantum yield of photosynthesis: light partially inhibits dark respiration. Plant Physiology 75 : 95 101. Sitch S, Smith B, Prentice IC, Arneth A, Bondeau A, Cramer W, Kaplan JO, Levis S, Lucht W, Sykes MT et al 2003. Evaluation of ecosystem dyna mics, plant geography and terrestrial carbon cycling in the LPJ dynamic global vegetation model. Global Change Biology 9 : 161 185. Slot M, Wright SJ, Kitajima K. 2013. Foliar respiration and its temperature sensitivity in trees and lianas: in situ measure ments in the upper canopy of a tropical forest. Tree Physiology 33 : 505 515. Slot M, Zaragoza Castells J, Atkin OK. 2008. Transient shade and drought have divergent impacts on the temperature sensitivity of dark respiration in leaves of Geum urbanum Func tional Plant Biology 35 : 1135 1146.
145 Smith NG, Dukes JS. 2013. Plant respiration and photosynthesis in global scale models: Incorporating acclimation to temperature and CO 2 Global Change Biology 19 : 45 63. Stahl C, Burban B, Wagner F, Goret J Y, Bompy F, Bonal D. 2013 I nfluence of seasonal variations in soil water availability on gas exchange of tropical canopy trees. Biotropica 45 : 155 164. Stocker O. 1935. Assimilation und Atmung westjavanischer Tropenbume. Planta 24 : 402 445. Strand Foyer CH, Gu stafsson P, Gardestrm P, Hurry V. 2003. Altering flux through the sucrose biosynthesis pathway in transgenic Arabidopsis thaliana modifies photosynthetic acclimation at low temperatures and the development of freezing tolerance. Plant, Cell & Environment 26 : 523 535 Thornton PE, Law BE, Gholz HL, Clark KL, Falge E, Ellsworth DS, Goldstein AH, Monson RK, Hollinger D, Falk M et al. 2002. Modeling and measuring the effects of disturbance history and climate on carbon and water budgets in evergreen needlelea f forests. Agricultural and Forest Meteorology 113 : 185 222. Tjoelker, MG, Oleksyn J, Lorenc Plucinska G, Reich PB. 2009. Acclimation of respiratory temperature responses in northern and southern populations of Pinus banksiana New Phytologist 181 : 218 22 9. Tjoelker, MG, Oleksyn J, Reich PB, ytkowiak R. 2008. Coupling of respiration, nitrogen, and sugars underlies convergent temperature acclimation in Pinus banksiana across wide ranging sites and populations. Global Change Biology 14 : 782 797. Tjoelker, MG, Oleksyn J, Reich PB. 1999a. Acclim ation of respiration to temperature and CO 2 in seedlings of boreal tree species in relation to plant size and relative growth rate. Global Change Biology 5 : 679 691 Tjoelker, MG, Oleksyn J, Reich PB. 1999b. Changes in leaf nitrogen and carbohydrates unde rlie temperature and CO 2 acclimation of dark respiration in five boreal tree species. Plant, Cell & Environment 22 : 767 778 Tjoelker, MG, Oleksyn J, Reich PB. 2001. Modelling respiration of vegetation: evidence for a general temperature dependent Q 10 Gl obal Change Biology 7 : 223 230. Turnbull MH, Murthy R, Griffin KL. 2002. The relative impacts of daytime and night time warming on photosynthetic capacity in Populus deltoides Plant, Cell & Environment 25 : 1729 1737.
146 Turnbull MH, Tissue DT, Griffin KL, Richardson SJ, Peltzer DA, Whitehead D. 2005. Respiration characteristics in temperate rainforest tree species differ along a long term soil development chronosequence. Oecologia 143 : 271 279. Turnbull MH, Tissue DT, Murthy R, Wang X, Sparrow AD, Griffin KL. 2004. Nocturnal warming increases photosynthesis at elevated CO 2 partial pressure in Populus deltoides New Phytologist 161 : 819 826. Turnbull MH, Whitehead D, Tissue DT, Schuster WSF, Brown KJ, Griffin KL. 2003. Scaling foliar respiration in two co ntrasting forest canopies. Functional Ecology 17 : 101 114. Van Bodegom PM, Douma JC, Witte JPM, Ordoez JC, Bartholomeus RP, Aerts R. 2012. Going beyond limitations of plant functional types when predicting global ecosystem atmosphere fluxes: exploring th e merits of traits based approaches. Global Ecology and Biogeography 21 : 625 636. Van de Weg MJ, Meir P, Grace J, Damian Ramos G. 2012. Variation of photosynthetic capacity, leaf dark respiration and leaf traits in the canopy of a tropical montane cloud f orest in Peru. Oecologia 168 : 23 34. Vierling E. 1991. The roles of heat shock proteins in plants. Annual Review of Plant Physiology and Plant Molecular Biology 42 : 579 620 Wang K, Kellomki S, Laitinen K. 1995 Effects of needle age, long term temperat ure and CO 2 treatments on the photosynthesis of Scots pine. Tree Physiology 15 : 211 218. Wang W, Dungan J, Hashimoto H, Michaelis AR, Milesi C, Ichii K, Nemani RR. 2011. Diagnosing and assessing uncertainties of terrestrial ecosystem models in a multimode l ensemble experiment: 1. Primary production. Global Change Biology 17 : 1350 1366. Wang W, Ichii K, Hashimoto H, Michaelis AR, Thornton PE, Law BE, Nemani RR. 2009. A hierarchical analysis of terrestrial ecosystem model Biome BGC: e quilibrium analysis an d model calibration. Ecological Modelling 220 : 2009 2023. Warton D, Duursma R, Falster D, Taskinen S. 2012. SMATR: (Standardized) major axis estimation and testing routines. R package version 3.2.0. h ttp://cran.r project. org/package=smatr ). Way DA, Oren R. 2010. Differential responses to changes in growth temperature between trees from different functional groups and biomes: a review and synthesis of data. Tree Physiology 30 : 669 688.
147 Weih M, Karls son PS. 2001. Growth response of Mountain birch to air and soil temperature: is increasing leaf nitrogen content an acclimation to lower air temperature? New Phytologist 150 : 147 155. Wright IJ, Reich PB, Atkin OK, Lusk CH, Tjoelker MG, Westoby M. 2006. I rradiance, temperature and rainfall influence leaf dark respiration in woody plants: evidence from comparisons across 20 sites. New Phytologist 169 : 309 319. Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender Bares J, Chapin FS, Cor nelissen JHC, Diemer M et al 2004. The worldwide leaf economics spectrum. Nature 428 : 821 827. Wright SJ, Calderon O, Hernandez A, Paton S. 2004. Are lianas increasing in importance in tropical forests? A 17 year record from Panama. Ecology 85 : 484 489. Wright SJ, Muller Landau HC, Schipper J. 2009. The future of tropical species on a warmer planet. Conservation Biology 23 : 1418 1426. Wythers KR, Reich PB, Bradford JB. 2013. Incorporating temperature sensitive Q 10 and foliar respiration acclimation algorithms modifies modeled ecosystem responses to global change. Journal of Geophysical Research Biogeosciences 118 : 1 14. Wythers KR, Reich PB, Tjoelker MG, Bolstad PB. 2005. Foliar respiration acclimation to temper ature and temperature variable Q 10 alter ecosystem carbon balance. Global Change Biology 11 : 435 449. Xu C Y, Griffin KL. 2006. Seasonal variation in the temperature response of leaf respiration in Quercus rubra : foliage respiration and leaf properties. F unctional Ecology 20 : 778 789. Xu C Y, Schuster WSF, Griffin KL. 2007. Seasonal variation of temperature response of respiration in invasive Berberis thunbergii (Japanese barberry) and two co occurring native understory shrubs in a northeastern US deciduo us forest. Oecologia 153 : 809 819. Zha TS, Wang KY, Ryypp A, Kellomki S. 2002. Impact of needle age on the response of respiration in Scots pine to long term elevation of carbon dioxide concentration and temperature. Tree Physiology 22 : 1241 1248. Zhao MS, Running SW. 2010. Drought induced reduction in global terrestrial net primary production from 2000 through 2009. Science 329 : 940 943 Ziehn T, Kattge J, Knorr W, Scholze M. 2011. Improving the predictability of global CO 2 assimilation rates under cl imate change. Geophysical Research Letters 38 : L10404.
148 Zotz G, Harris G, Kniger M, Winter K. 1995. High rates of photosynthesis in a tropical pioneer tree, Ficus insipida Flora 190 : 265 272.
149 BIOGRAPHICAL SKETCH Martijn grew up in a small town in the north of The Netherlands, where he spent a comfortable youth not thinking much about trees other than as climbing objects and producers of pulp for the newspapers that he delivered every morning. Nonetheless, in 1997 he enrolled in Wageningen University to study forestry, specializing in forest ecology. When not studying, he read books, favoring the smoky comfort of Russian classics. Consequently, when the opportunity to go to Russia arose, he happily accepted, and in early 2002 he moved into a log cabin i n central Siberia. There he got his first exposure to plant ecophysiology, working on cold induced photoinhibition of Scots pine in the Siberian taiga as an intern at the Max Planck Institute for Biogeochemistry. After he earned his degree as a forestry en gineer in Wageningen he worked in the tropical south of China for a few m onths, before starting a Master of Research in e cology at the University of York. While at York he continued to work on plant ecophysiology and tropical forest ecology. After some ye ars working as a field assistant, lab technician, intern, supermarket clerk, and garbage man, Martijn came to the University of Florida in 2007 to join the PhD program in the Department of Botany. Here, his interests in tropical forest ecology and plant ec ophysiology were combined in his PhD project that focused on temperature response of dark respiration in tropical forest trees and lianas in Panama. In summer 2013 Martijn will return to the tropics to start as a post doctoral research associate at the Smi thsonian Tropical Research Institute in Panama. He plans to read many more books and hopes to go back to climb ing trees in old age