Species-Specific Growth Responses to Climate Variations in Understory Trees of a Central African Rain Forest

Species-Specific Growth Responses to Climate Variations in Understory Trees of a Central African Rain Forest
of 9
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  Species-specific Growth Responses to Climate Variations in Understory Treesof a Central African Rain Forest CamilleCouralet 1 , 2 , 4 ,FrankJ. Sterck  3 ,Ute Sass-Klaassen 3 , JorisVanAcker 2 , and Hans Beeckman 11 Laboratory for Wood Biology and Xylarium, Royal Museum for Central Africa, Leuvensesteenweg 13, 3080 Tervuren, Belgium 2 Laboratory for Wood Technology, Gent University, Coupure Links 653, 9000 Gent, Belgium 3 Forest Ecology and Forest Management Group, Wageningen University, PO Box 47, 6700 AA Wageningen, The Netherlands ABSTRACT Basic knowledge of the relationships between tree growth and environmental variables is crucial for understanding forest dynamics and predicting vegetation responsesto climate variations. Trees growing in tropical areas with a clear seasonality in rainfall often form annual growth rings. In the understory, however, tree growth issupposed tobe mainly affected by interferencefor access tolight and otherresources.In the semi-deciduous Mayombeforest of the Democratic Republicof Congo,theevergreen species  Aidia ochroleuca ,  Corynanthe paniculata  and  Xylopia wilwerthii   dominate the understory. We studied their wood to determine whether they formannualgrowthringsin responsetochangingclimateconditions.Distinctgrowthrings wereprovedtobeannualandtriggeredby acommonexternalfactorforthe threespecies. Species-specific site chronologies were thus constructed from the cross-dated individual growth-ring series. Correlation analysis with climatic variables revealedthat annual radial stem growth is positively related to precipitation during the rainy season but at different months. The growth was found to associate with pre-cipitation during the early rainy season for  Aidia  but at the end of the rainy season for  Corynanthe   and  Xylopia . Our results suggest that a dendrochronological approachallows the understanding of climate–growth relationships in tropical forests, not only for canopy trees but also for evergreen understory species and thus arguably forthe whole tree community. Global climate change influences climatic seasonality in tropical forest areas, which is likely to result in differential responses across specieswith a possible effect on forest composition over time. Abstract in French is available at Key words  :  Aidia ochroleuca ; climate–growth relationships;  Corynanthe paniculata ; Democratic Republic of Congo; tropical dendrochronology;  Xylopia wilwerthii  . T ROPICAL FORESTS ARE UNDER COMBINED PRESSURE  of global macro-climate changes and deforestation that rapidly modify local climaticconditions (Bonan 2008). In Africa, the average temperature is ex-pected to rise by 3–4 1 C during the 21st century, which is 1.5 timesmore rapid than the foreseen global temperature change (Boko  et al  .2007). Future climate scenarios also predict a 5–15 percent de-crease in precipitation during the rainy season and a decline of 3–4percent in annual rainfall per decade in the African tropics (Malhi& Wright 2004, Boisvenue & Running 2006). In addition, directhuman pressure is likely to increase the vulnerability of forests tothese warmer anddrier conditions (Koenig 2008). Recentmodeling studies showed that deforestation enforces the effects of warming and drought events (Malhi  et al  . 2008). It is, however, unknownhow different forest types will develop under these constraints, sincelong-term records on responses of tropical tree species to changes inenvironmental factors are lacking (Clark 2004, Phillips  et al  . 2009).Depending on climate region, forest type or canopy position, treespecies can differ in their tolerance to drought and shade (Condit et al  . 1995, Sterck   et al  . 2006, Engelbrecht  et al  . 2007). The highdiversity of species and their variable ecological preferences poten-tially allow a high diversity of reactions within the tree communi-ties. Yet, testing this hypothesis has so far proved difficult, due tothe limited research on species-specific responses to environmentalfactors in different tropical forest types. Field studies using perma-nent plots in the tropics provide growth data of at most a few de-cades and rarely with an annual resolution (Clark   et al  . 2003).Dendrochronologycanprovidelong-termrecordsoftreegrowthwith annual resolution, but is most successful where climate showsstrong and regular seasonality triggering annual ring formation(Jacoby 1989, Schweingruber 1996). Rain forest trees have long beenthought notto form annual rings because of weakseasonality andlow variability in temperature and day length in the tropics (Whitmore1998). This persistent assumption caused many dendrochronologiststofocustheireffortsontemperatetreesratherthantropicaltrees.Suchpreferenceis alsoexplainedbytheextensively complex wood anatomy in tropical trees (De´tienne 1989). Nevertheless, across the varioustropical biomes and for most trees, alternating environmental condi-tions induce periods of reduced or suspended cambial activity leading to the formation of identifiable growth layers (De´tienne 1989, Ma-riaux 1995, Worbes 2002). In lowland tropical rain forests, regulargrowth periodicity allowed the construction of species-specific tree-ring chronologies for several species (Devall  et al  . 1995, Mariaux 1995,De´tienne etal  .1998,Sch¨ongart etal  .2002,Worbes etal  .2003,Brienen & Zuidema 2005, Brookhouse 2006, Pumijumnong & Wanyaphet 2006, Sch¨ongart  et al  . 2006, Buckley   et al  . 2007, Lisi2008,Brienen et al  .2009).Suchlong-termgrowthseriescanbecross-compared with climate records to unravel the species-specific sensitiv-ity to past climatic conditions. This is a prerequisite to evaluate theresponse of vegetation formations to future environmental changes(Condit  et al  . 1995, Pumijumnong & Park 1999, Chidumayo 2005,Trouet  et al  . 2006). Received 25 June 2009; revision accepted 13 October 2009. 4 Corresponding author; e-mail: BIOTROPICA 42(4): 503–511 2010 10.1111/j.1744-7429.2009.00613.x r 2010 The Author(s) 503Journal compilation r 2010 by The Association for Tropical Biology and Conservation  From this perspective, the floristic and structural complexity of tropical rain forests encourages consideration of the large rangeof species they harbor. Much knowledge has been generated onupper-canopy trees that are, at the adult stage, not light-limitedand strongly exposed to water stress. On the other hand, the growthof lower-canopy or understory species spending their lifetime underother trees is assumed to be more limited by local site factors,such as light, than by water (Phipps 1982). Moreover, small-staturetrees with ambiguous seasonal variability in leaf fall preferentially use deeper sources of soil water than larger or deciduous trees(Meinzer  et al  . 1999). Ring-width series of understory trees arethus supposed to show a much more variable ontogenetic growthtrend, less directly related to external environmental factors andto water stress in particular. Studies of radial growth responses toclimate in understory trees of temperate forests support this hy-pothesis (Liu 1993, Orwig 1997, Rasmussen 2007, Mart´ın-Benito et al  . 2008). Nevertheless, growth response to climate is hardly documented for evergreen tree species. While in deciduous treesleaf shedding is an obvious sign of cambial dormancy and suggests a clear link between environmental factors and phenology, for ever-green trees this relation is thought to be erratic (Jacoby 1989, Worbes 1999). We conducted a study on three common, evergreen understo-ry tree species from the Mayombe forest, a tropical semi-evergreenrain forest west of the Democratic Republic of Congo (DRC). Cli-mate changes are expected to affect the region significantly in the21st century but studies of the influence of climate variations ontree growth in African forests are scarce (De´tienne  et al  . 1998, Worbes  et al  . 2003, Couralet  et al  . 2005, Verheyden  et al  . 2004,Sch¨ongart  et al  . 2006, Trouet  et al  . 2006, Sass-Klaassen  et al  . 2008)and to our knowledge still lacking for lower-canopy tree species. We aimed to assess whether dendrochronology is applicable tothese understory tree species and, if so, whether andhow their radialgrowth dynamics is related to inter- and intra-annual changes inclimate (temperature, precipitation and solar radiation).In this study, after demonstrating the annual nature of treerings in three understory species, we address the following questions: (1) Do trees within and across species synchronize stemgrowth in response to the same environmental factors? (2) How dotrees of different species respond in radial stem growth to year-to-year and within-year variations in climate? We expect to revealdiversity ingrowth patterns across different understory tree speciesand given the seasonality in rainfall, we predict positive growthresponses to precipitation. METHODS S TUDY SITE .—The Luki forest reserve is located in the southwesternDemocraticRepublicofCongo(DRC)(5 1 28 0 –42 0 N,13 1 4 0 –18 0 E),30km north from the port city of Boma (Fig. 1A). It is the south-ernmost remnant of the Mayombe forest, stretching along the Atlan-tic Ocean from the central coast of Gabon and renowned for its highfloristic diversity and the presence of large timber trees (Monteiro1962). The protected forest area covers 32,700ha of hilly landscape(150–500masl) on heterogeneous soils, generally ferrallitic and withpoor chemical content (Monimeau 1990).Climate records (1959–2006) were available from the Lukimeteorological station for precipitation, air temperature, relativehumidity and solar irradiance (Fig. 1B). The climate is character-ized by a mean annual temperature of 24.6 1 C with limited yearly variation and a mean annual rainfall of 1180mm/yr. A distinct dry period lasts from June to September. Such low annual rainfall and3–4mo with  o 50mm monthly precipitation generally do not fa-vor the presence of a dense humid forest; however, the strong oce-anic influence, the landscape of the region and the self-regulating effect of the vegetation create favorable conditions for the establish-ment of dense humid forest (Se´ne´chal  et al  . 1989, Pendje & Baya ki1992, Lubini 1997). Mists are present all yearlong and during thedriest months a thick, low-level but non-precipitating cloud layerblocks solar irradiance and causes temperature to drop. The rainfall FIGURE1. (A) Location of the study site with simplified vegetation cover of the area from Global Land Cover 2000 (Mayaux   et al  . 2004): darker zones are a mosaicof evergreen, deciduous and mixed forests with a minimum of 15 percent tree cover. (B) Climate diagram of the Luki meteorological station, Democratic Republicof Congo: monthly means of rainfall (  SD), temperature, air humidity (1959–2006) and solar irradiance (1959–1994). 504 Couralet  et al  .  shortage is thus partially compensated and the relative air humidity remains constantly high, always  4 80 percent. Consequently, de-spite a clear seasonal rainfall pattern plants may not suffer extremewater stress during the dry season.The forest of Luki can be generally classified as a tropical semi-evergreen rain forest of the Guineo-Congolean forest domain(Lebrun & Gilbert 1954, Lubini 1997, Whitmore 1998) butoccurs in a mosaic landscape with patches of agricultural fieldsand settlements. It consists of a mixture of deciduous and evergreentree species in the upper-stratum and mostly evergreen species inthe understory.S TUDY SPECIES .—  Aidia ochroleuca  (K. Schum.) Bullock ex E.M.A.Petit,  Corynanthe paniculata  Welw. (both Rubiaceae) and  Xylopiawilwerthii   Wild. & T. Durand var. cuneata De Wild. (Annonaceae)are abundant and commonly co-occurring species in the Guineo-Congolean rain forests, from Gabon to the eastern DRC (Petit1962, Schmitz 1988, Se´ne´chal  et al  . 1989, Lubini 1997). In thesecondary forest of Luki they can contribute to 4 60 percent of thetotal basal area (Donis & Maudoux 1951; C. Couralet, These evergreen, medium-sized trees of up to 20m in heightand 60cm diameter at breast height (dbh) are confined to thelower-canopy level and understory (Aubre´ville 1961, Petit 1961,Lubini 1997, Lebrun & Stork 2003).  Corynanthe paniculata  and  X. wilwerthii   have a straight trunk whereas  A. ochroleuca  is oftenramified from the base. All three species are characterized by a very hard and fine-textured wood used by local communities to producesolid tools and high-quality charcoal (Fouarge & Ge´rard 1964).S  AMPLING AND SAMPLE PREPARATION .—Stem discs of five trees perspecies were collected in 2005. We selected representative trees of each species, with crowns under the closed upper canopy layer anddiameters within the predominantly observed range (15–40cm).Trees were cut at 0.5–1m aboveground level such that a maximumnumber of growth rings were visible. The 4–6-cm-thick sectionswere deep-frozen for 2wk to prevent insect or fungal attacks andsubsequently air-dried. Their cross-sectional surface was planed andsanded up to a grid size of 1200.D ETECTION OF CONCENTRIC GROWTH LAYERS IN THE WOOD .—Growth-ring structure was analyzed macroscopically and microscopically following the International Association of Wood Anatomists hard-wood feature list (Wheeler  et al  . 1989). Ring boundaries of tropicaltree species are characterized by a variety of features such as marginalparenchyma bands, alternation of fiber and parenchyma tissues, vari-ation in the vessel size and distribution, variation in the fiber wallthickness, or a combination of all these features (De´tienne 1989, Worbes 1995). The wood of the three species is diffuse-porous withdifferent rhythmic variations in wood structure visible to the nakedeye (Fig. 2); however, microscopic observations were required to re-liably detect anatomical features that mark growth–ring boundariesand distinguish them from,  e.g  ., intra-annual density variations. Themain difficulties for the three study species were their low growthlevels (average ring width of 1.13–1.65mm) leading to frequently wedging or absent rings.Ring widths were measured to the nearest 0.01mm under a ste-reo-microscope coupled with a Lintab measuring device and Tsap- Win software (Rinn 2003), on two or three (for irregularly shapedsamples) radii for each stem disc. Becausethe supposed growth seasonspans two calendar years (rainy season typically from October n  1  toMay  n ), the year in which annual cambial activity is expected to stop(year  n )wasnominated to define the growth year and the correspond-ing ring-width value.D ENDROCHRONOLOGICAL ANALYSES .—Tree-ring analyses were ap-plied to all samples to show whether stem growth rate was synchro-nized over time for all trees. Cross-dating,  i.e  ., the matching of ring-width series, allows for the detection of missing rings (if one tree inthe sample set did not form a ring in a particular year) or false rings(variations in the wood anatomy mistaken with ring boundaries; Wils et al  . 2009). Successful cross-dating of ring-width series denotes con-sistent and synchronous patterns of variation (Cook & Kairiukstis1992) and indicates that a common external factor controls ring for-mation in different trees (Pilcher & Gray 1982, Worbes 1995, Cher-ubini  et al  . 1998). Cross-dating was performed visually incombination with a correlation analysis using Cofecha software(Tree-Ring Laboratory, Columbia University, New York, U.S.A.;Holmes 1983). Starting with radii from the same tree, alignment of the ring-width series allowed the identification of anomalies that werethen corrected after investigation on the stem discs. The series werethen averaged per tree and the cross-dating process was repeatedbetween tree averages. The detection of pointer years,  i.e  ., extremeyears common within a site, served as an extra-check during the cross-dating process (Schweingruber 1996).Tree growth is affected by climatic fluctuations and by a widearray of non-climatic factors (Brookhouse 2008), e  .  g  ., canopy dynam-ics or individual size-related trends (Pilcher & Gray 1982). Tree-ring width series thus reflect a complex set of variations. To amplify theclimate-induced signal, the series were standardized using Arstan soft-ware (Tree-Ring Laboratory; Cook 1985) to remove low-frequency,most likely non-climatic trends. A smoothing spline (wavelength of 32yr) was first fitted to each raw series (Cook & Kairiukstis 1992),then each measure was divided by the corresponding value of thefunction to transform the srcinal curves into stationary time-series(mean=1 and homogenous variance). A ring-width index series wascreated for each radius then tree. Autocorrelation (AC, year-to-yeardependence of ring-width values due to the influence of a growth sea-son on the next) was removed by applying an autoregressive model of the adequate order for eachseries. Eventually, the indexed curves wereaveraged to produce the final species chronologies.Statistical information on the tree-ring measurements is summa-rized in Table 1. The mean sensitivity (MS) indicates the level of be-tween-ring variability in the measured series and reflects the sensitivity of radial tree growth to a common external signal (Fritts 1976, Schwe-ingruber 1996). Inter-series Pearson’s correlations ( P  o 0.05) werecalculatedbetweenradiiofeachtree,betweentreesandbetweenspeciesto express the level of affinity of the measured curves and evaluate thequalityofthecomputedmeanseries.Student’s t  -valuesandcoefficients Climate–Growth Relationships of Understory Trees 505  FIGURE2. Tree-ring structure and growth-ring boundaries of three understory species of the Luki forest reserve, Democratic Republic of Congo. (A)  Aidiaochroleuca : vessel frequency steadily decreases toward the ring boundary, marked by a vessel-free fiber band. (B)  Corynanthe paniculata : ring boundary is marked by oneor two lines of radially flattened fibers, but can be confused with the numerous fiber-wall density variations visible on the wood. Black arrows point at these lines on a thin section that allows seeing the microscopic wood anatomical structure. (C)  Xylopia wilwerthii  : ring boundary is marked by two to three lines of very thick-walledlatewood fibers, making the wood appear darker. Notice the ladder-like pattern of axial parenchyma cells (scalariform parenchyma), typical for Annonaceae. Black crosses indicate ring boundaries. Radial growth direction is from bottom to top. Black scale bars=1 cm, white scale bar=1mm.TABLE1 . Descriptive statistics and quality control of the ring-width series contributing to the chronologies for three understory species of the Luki forest reserve, Democratic Republic of Congo:  Aidia ochroleuca  ,  Corynanthe paniculata   and   Xylopia wilwerthii . SD, standard deviation, AC, autocorrelation (1-year lag); MS, meansensitivity; Glk, Gleich¨ aufigkeitskoeffizient. Average of Pearson’s correlations,  t -value (  P o 0.05) and Glk or coefficient of parallel variation. Species  Aidia ochroleuca Corynanthe paniculata Xylopia wilwerthii  Number of samples 5 5 5Mean diameter (cm) (range) 21 (18–25) 24 (17–40) 16 (14–17)Mean age (yr) (range) 55 (39–85) 97 (69–112) 39 (26–49)Time span (yr) 1922–2006 1895–2006 1958–2006Ring width (mean  SD, mm) a  1.65  1.14 1.13  0.70 1.62  0.98 AC a  0.61 0.54 0.19MS a  0.42 0.42 0.50 Within-tree correlation b 0.70 0.60 0.72Between-treeCorrelation b 0.36 0.24 0.27 t  -value (  SD) b 5.6  0.7 5.9  2.0 4.4  1.7Glk (  SD, %) b 72.3  7.2 67.0  4.9 73.2  17.3Values based on: a  Raw series. b Detrended data. 506 Couralet  et al  .  of parallel variation ( Gleich¨ aufigkeitskoeffizient   [Glk]) were calculatedbetween trees of each species to assess the quality of the final chronol-ogies developed for climate-growth analysis (Baillie & Pilcher 1973). A  NALYSIS OF CLIMATIC EFFECTS .—To investigate the link between theradial stem growth of understory trees and climate, we compared eachspecies chronology with annual records of total rainfall, average temper-ature and average solar radiation from the Luki meteorological station.Moreover, to test whether the within-year timing of climate variationplays an important and differential role for the three species studied,Pearson’s correlations ( P  o 0.05) were computed between the specieschronologies and contemporary time-series of climate variables on a monthly basis and for groups of months (Fritts 1976). According toresults from earlier studies on tree growth in a seasonal tropicalclimate (Couralet  et al  . 2005, Trouet  et al  . 2006) and high-resolutionmeasurements of radial growth the growing period was assumed to co-incide approximately with the rainy season lasting from October toMay.We calculated correlationsfor a12-moperiod fromSeptember year n  1  to August  year  n , thus amply covering the rainy season.In addition, the dbh of 10 trees per species was measured every month with graduated tape during 15mo (April 2006–June 2007) tostudy seasonal growth dynamics of   Aidia ,  Corynanthe   and  Xylopia  inresponse to contemporary weather conditions. Because correlationswith growth were found only with precipitation, the measurementswere visually compared with contemporary monthly precipitation re-cords to explore intra-annual climate–growth relationships and refinethe results obtained from the ring-width chronologies. RESULTS S TATISTICAL CHARACTERISTICS OF SINGLE RING -  WIDTH SERIES ANDCHRONOLOGIES .—Mean ring width ranged from 1.13mm for  C. pan-iculata  to1.62 and 1.65mm for  X. wilwerthii   and  A. ochroleuca ,respec-tively. The large standard deviations for this trait indicated a highvariationinaverageringwidthbetweentreesofasamespecies(Table1). Within-tree correlations were high and allowed to merge thegrowth series into mean curves for each individual tree. Trees of thesame species were also characterized by high values for mean seriesinter-correlations,  t  -values and Glk, thus providing robust chronol-ogies per species. The rather high values of MS indicated that ring width varied widely between years. The higher AC values in stemgrowth between successive years for  Aidia  and  Corynanthe   probably resulted in slightly lower values for MS of these species. Overall,these results exhibit that trees of the same species synchronized theirradial stem growth to the same external environmental factor. Onthe other hand, the correlations between species chronologies werelow or very low (  Aidia–Xylopia : 0.20,  Corynanthe–Xylopia : 0.15,  Aidia–Corynanthe  : 0.03), suggesting that the three species differ inresponse to climatic factors. This is visible when looking at the threechronologies that show almost no common variation (Fig. 3).C LIMATE – GROWTH RELATIONSHIPS .—There was no significant corre-lation between the three species chronologies and annual total rain-fall, average temperature or average solar radiation, indicating thatradial growth of the study species was not controlled by long-termannual variation of these climate factors. Single-month temperatureand solar radiation were also not associated with the annual varia-tion in radial growth of the studied species; however, significantpositive correlations appeared between ring width and single-month precipitation values for the three species. Remarkably, themonths for which significant values were found differed betweenspecies (Fig. 3). For  Aidia  the correlation between radial growthand rainfall was significant at the onset of the rainy season, inNovember, whereas for  Corynanthe   and  Xylopia  the correlationwas significant at the end of the rainy season, in March and April,respectively (Figs. 3 and 4).These trends were supported by correlations between radialgrowth and monthly values of precipitation over the studied period,most clearly during the rainy season (Fig. 4).  Aidia  showed an over-all positive growth response to rainfall in the early rainy season(September–February), culminating in November, and no specifictrend in the late rainy season. Inversely, the radial growth of   Cory-nanthe   was negatively correlated to rainfall in the early rainy season(September–January) and positively correlated to rainfall in the laterainy season with a peak in March. For  Xylopia  there was no suchclear pattern but the correlation between radial growth and rainfallwas mostly positive over the whole rainy season and was strongest inthe end, in April. Both types of analysis (Figs. 3 and 4) thus suggestthat the amount of precipitation was critical for radial stem growthduring the transition months from dry to wet season for  Aidia , andfrom wet to dry season for  Corynanthe   and  Xylopia . FIGURE3. Illustration of the correlation ( P  o 0.05) between the ring-widthindex (RWI) of three understory tree species and single-month climatic recordsin the Luki forest reserve, Democratic Republic of Congo (1959–2006):  Aidiaochroleuca  and November rainfall ( r  =0.27),  Corynanthe paniculata  and Marchrainfall ( r  =0.29),  Xylopia wilwerthii   and April rainfall ( r  =0.34). Climate–Growth Relationships of Understory Trees 507
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!