Meteorological seasonality affecting individual tree growth in forest plantations in Brazil

Forest Ecology and Management 380 (2016) 149–160

Contents lists available at ScienceDirect

Forest Ecology and Management journal homepage: www.elsevier.com/locate/foreco

Meteorological seasonality affecting individual tree growth in forest plantations in Brazil Otávio C. Campoe a,⇑, Juliana S.B. Munhoz b, Clayton A. Alvares c, Rafaela L. Carneiro d, Eduardo M. de Mattos b, Ana Paula C. Ferez e, José Luiz Stape c a

Federal University of Santa Catarina – UFSC, 89.520-000 Curitibanos, SC, Brazil Department of Forest Sciences, University of São Paulo – ESALQ, 13416-000 Piracicaba, SP, Brazil Suzano Papel e Celulose Brasil, Av. Dr. José Lembo, 1010, Itapetininga, SP 18207-780, Brazil d Forestry Science and Research Institute – IPEF, 13415-000 Piracicaba, SP, Brazil e Instituto Centro de Vida, Cuiabá, MT 78043-055, Brazil b c

a r t i c l e

i n f o

Article history: Received 22 May 2016 Received in revised form 29 August 2016 Accepted 30 August 2016

Keywords: Dendrometer Meteorological seasonality Eucalyptus Pinus Multivariate analysis

a b s t r a c t Seasonal meteorological variability within and among years has significant impact on forest productivity, thus understanding its detailed effects on tree growth contributes to the knowledge of the processes controlling forest productivity. This study used high frequency measurements of dendrometer bands (every 2–4 weeks over 1–2 years) to assess tree growth of four different planted forest types (Brazilian native tree species, Eucalyptus grandis, Pinus caribaea var. hondurensis, and Pinus taeda) in response to meteorological variables in distinct regions of Brazil. Multivariate linear regression analysis was applied to relate growth as function of meteorological variables. All species responded to multiple meteorological variables related to air temperature and water availability (maximum air temperature (p < 0.05), vapor pressure deficit (p < 0.01) and actual evapotranspiration (p < 0.01)). Dominant trees were more responsive to meteorological seasonality, compared to suppressed trees. Understanding the relations between forest growth and meteorological variables is useful and has practical applications, including the optimization of species zoning, modeling and the evaluation of climate change impacts on planted tree species for restoration or wood production purposes. Ó 2016 Elsevier B.V. All rights reserved.

1. Introduction Meteorological variability among and within years has direct and indirect effects on resources availability (light, water and nutrients) to the trees, with significant impacts on forest growth and development (Binkley et al., 2004). Studies performed on natural and planted forest ecosystems have shown that tree growth responds to meteorological variation at different spatial and temporal scales (King et al., 2013; Cristiano et al., 2014; UrrutiaJalabert et al., 2015). Commercial Eucalyptus plantations can significantly increase wood productivity due to increasing rainfall across climatic gradients (Stape et al., 2010). They also show a high temporal resolution response to water supply, growing within the same week after the increase in soil water content (Drew et al., 2008). Pine plantations follow similar patterns of responsiveness to water availability and temperature variations within and among years (Alvarez et al., 2012). ⇑ Corresponding author. E-mail address: [email protected] (O.C. Campoe). http://dx.doi.org/10.1016/j.foreco.2016.08.048 0378-1127/Ó 2016 Elsevier B.V. All rights reserved.

Depending on the climatic region, tree growth is affected by specific meteorological variables. In general, trees growing on dry climates, experiencing limited water availability across periods or seasons, are more affected by the amount and distribution of rainfall and vapor pressure deficit (Callado et al., 2013; Rowland et al., 2014), while trees growing on high latitudes or altitudes and no dry seasons are more affected by variation in air temperature (Spanner et al., 1990; Boisvenue and Running, 2006). Forest plantations in Brazil, for restoration and wood production, are very productive mainly due to favorable climatic conditions (Stape et al., 2010; Gonçalves et al., 2013; Ferez et al., 2015; Venegas-González et al., 2016), therefore, understand the role of meteorological seasonality on tree growth is essential to maintain current levels of productivity and to predict potential impacts of future climatic scenarios. Meteorological variability within years is especially important to identify specific variables, and how their ranges affect forest growth on specific periods of the year (Rawal et al., 2014; Watt et al., 2014). Previous studies have successfully assessed the influence of meteorological variables and silvicultural treatments on growth

150

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

rates of planted (Sette et al., 2012; Battie-Laclau et al., 2014) and native tree species (Callado et al., 2004; Cardoso et al., 2012; Grogan and Schulze, 2012), however, studies relating temporally detailed forest growth as a function of meteorological seasonality are still limited for planted forests for both wood production and restoration purposes. Simulations of climatic scenarios for the next century in South America suggest significant increase in air temperature and changes in distribution and reduction in precipitation (Chou et al., 2014). These changes will certainly affect forest productivity at different scales of space and time. Therefore, a detailed understanding of the interactions between tree growth and meteorological seasonality is important for practical uses, such as species zoning aiming the optimization of wood production on commercial planted tree, modeling production and development of specific species to distinct climatic regions, and to estimate potential impacts of climate change on growth of different tree species (Booth, 2013; Janowiak et al., 2014). Forest productivity is usually assessed by annual surveys, preventing a detailed evaluation of intra-annual meteorological effects on tree growth. The only method to capture meteorological seasonality affecting individual tree growth is by increasing the frequency of measurements over short periods of time, such as months, weeks or days (Wunder et al., 2013). Thus, this study used dendrometers aiming to precisely evaluate the influence of intraannual meteorological seasonality on bole cross-sectional area growth of different planted tree species in distinct climatic regions of Brazil. We studied different types of tree plantations (native tree species for restoration purposes, and Eucalyptus grandis, Pinus caribaea var. hondurensis, and Pinus taeda for wood production purposes), hypothesizing that regardless the species, individual tree growth variation will match the seasonality of specific meteorological variables, depending on the location in the country. 2. Material and methods 2.1. Experimental sites, tree selection and measurements The experimental sites were distributed across a wide range of meteorological conditions and soil types in Center-southern Brazil, comprising the states of Minas Gerais (MG), São Paulo (SP) and Paraná (PR) (Fig. 1, Table 1). The tree species presented encompass the two major genus planted in Brazil (Eucalyptus and Pinus, Gonçalves et al., 2013) and a mix of native tree species most commonly used in forest restoration plantations, with different ages and levels of productivity (Table 2). 2.1.1. Anhembi-SP The experiment was installed in riparian areas of Tietê River at Anhembi Forest Research Station (AFRS), University of São Paulo (Table 1), located within the Atlantic Forest Biome, dominated by a semi-deciduous seasonal forest (Morellato and Haddad, 2000). In March 2004, 20 native tree species were planted on an abandoned and degraded pasture (Campoe et al., 2010, Table 2). Before planting, the site was dominated by the African signal grass (Brachiaria decumbens) which was eliminated by applying 5 L ha
1 of glyphosate. Leaf-cutting ants (Atta spp. and Acromyrmex spp.) were controlled with the systematically placement of sulfluramidbased baits throughout the experimental area. The experimental design is a complete 2  2  2 factorial with two levels of each study factor: (i) proportion of pioneer and non-pioneer species: 50:50 and 67:33; (ii) planting spacing: 3 m  1 m (3333 plants ha
1) and 3 m  2 m (1667 plants ha
1) and (iii) silvicultural technology: traditional and intensive management (Campoe et al., 2010). For the current study, we selected

trees under intensive silviculture at 3  2 m planting spacing. This species planting spacing is widely used on restoration plantations on Atlantic Forest Biome (Rodrigues et al., 2009). The intensive silviculture provided complete control of weeds and alleviated any nutrient limitations to the planted trees. Weed control was carried out chemically by the application of 5 L ha
1 of glyphosate across the entire plot, every three months for the first two years, followed by spot applications when necessary. Fertilization was performed annually since planting time, totaling 81 kg ha
1 of N, 62 kg ha
1 of P, 33 kg ha
1 of K, 452 kg ha
1 of Ca and 180 kg ha
1 of Mg. Forty pioneers and 17 non-pioneer trees from 10 species were selected across the plots of the selected treatment (described above), with average sizes of cross-sectional area of 153 cm2, height of 7.5 m, and aboveground biomass of 22 kg for nonpioneer, and cross-sectional area of 303 cm2, height of 9.1 m, and aboveground biomass of 54 kg to for pioneers. We decided to present growth results separately for pioneers and non-pioneer species to show their differential response to the meteorological seasonality. After the installation of the 57 dendrometers 30 cm above ground level, and 30-day for band adjustments on trees boles, the measurements were performed from 30 to 60 days intervals. 2.1.2. Itatinga-SP The study was conducted within the Eucflux project (www.ipef. br/eucflux), located in the Southeastern Brazil, in São Paulo State, in a typical commercial Eucalyptus plantation belonging to the Duratex company (Campoe et al., 2012, Table 1). The area was planted in December 2002 with Eucalyptus grandis (W. Hill ex Maiden) seedlings from a 5th generation seed orchard (Australian provenance of the Coff’s Harbour) following minimum cultivation techniques of site preparation (Gonçalves et al., 2014), in a 3.75 m  1.60 m spacing (1667 trees per hectare, Table 2). Glyphosate was used (4 L ha
1) to eliminate competing vegetation prior to site preparation until canopy closure stage (approximately 18 months after planting). Leaf cutting ants (Atta spp. and Acromyrmex spp.) were controlled using sulfluramid-based baits whenever necessary. Fertilization was split according to common practice in commercial forest plantations in Brazil. Fertilizer was applied at planting time, 6 months, 1 and 2 years after planting. The total amount of nutrients applied was 62 kg ha
1 of N, 26 kg ha
1 of P, 97 kg ha
1 of K, 300 kg ha
1 of Ca, 144 kg ha
1 of Mg, 11 kg ha
1 of S, 2.4 kg ha
1 of B, 1.6 kg ha
1 of Zn and 1.3 kg ha
1 of Cu. Three hundred trees were selected across the 90 ha of the experimental area, comprising 50 suppressed (diameter at breast height, at 1.30 m above ground level, DBH = 10 cm, height = 19.3 m, aboveground biomass = 39.3 kg), 200 average (DBH = 17.5 cm, height = 26.4 m, aboveground biomass = 145.6 kg), and 50 dominant trees (DBH = 22.5 cm, height = 30.3 m, aboveground biomass = 277 kg). Selecting trees from different social status (by size for Itatinga, Nova Ponte and Telêmaco Borba experimental sites) provides the opportunity to show their different ability to respond to meteorological seasonality. After the installation of the 300 dendrometers at DBH position and 30-day buffer for band adjustments on trees boles, the measurements were performed at 14-day intervals. 2.1.3. Nova Ponte-MG The study was conducted within the project Potential Productivity of Pine in Brazil (www.ipef.br/pppib) located in Southeastern Brazil, Minas Gerais state. The area was planted in January 2008 with Pinus caribaea var. hondurensis using seedlings from seed orchard located in Agudos-SP, belonging to the Duratex Company. The weed competition was controlled by applying glyphosate

151

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

Fig. 1. Location of the four experimental sites across Center-southern Brazil. Site descriptions in Tables 1 and 2. Köppen climate classification map modified from Alvares et al. (2013a).

Table 1 Location, meteorological, and soil characteristics of the experimental sites. Site

Specie

Latitude (degrees)

Longitude (degrees)

Alt (m)

MAT (°C)

AR (mm)

RH (%)

CT

Soil type

Claya (%)

Sanda (%)

OMb (%)

pHc

Pd (mg kg
1)

CEC (mmolc kg
1)

Anhembi

Brazilian native tree species Eucalyptus grandis

22.72 S

48.17 W

460

20.9

1290

73

Cfa

18.0

70.0

34.4

4.5

8.0

69.0

22.97 S

48.73 W

760

19.3

1350

74

Cfa

16.8

79.8

23.1

4.5

2.6

53.7

19.27 S

47.77 W

960

19.8

1220

73

Cwb

Yellow-Red Oxisol Yellow-Red Oxisol Red Oxisol

80.0

13.0

29.6

4.5

5.5

75.5

24.30 S

50.38 W

850

18.4

1410

77

Cfb

Red Oxisol

43.0

51.0

31.7

4.1

2.2

123.0

Itatinga Nova Ponte Telêmaco Borba

Pinus caribaea var. hondurensis Pinus taeda

Alt: altitude above sea level; MAT: mean annual air temperature; AR: annual rainfall; RH: relative humidity; CT: climate type; Climate types from Alvares et al. (2013b): Cfa = Humid subtropical of oceanic climate, without dry season and with hot summer; Cfb = Humid subtropical of oceanic climate, without dry season and with temperate summer; Cwb = Humid subtropical with dry winter and temperate summer.OM: organic matter; CEC: cationic exchange capacity. a Pipette method (Donagemma et al., 1997). b Determined by potassium dichromate and sulfuric acid extraction. c CaCl2 0.01 mol L
1 soil to solution 1:2.5 ratio. d Extracted by ion exchange resin (Van Raij et al., 2001).

before planting and when necessary. Leaf-cutting ants (Atta spp. and Acromyrmex spp.) were controlled systematically with sulfluramid-based baits throughout the experimental area.

The experimental design is a complete 2  2  2 factorial with two management (thinned and unthinned), two fertilization regimes (not fertilized and fertilized), two levels of water availability (rainfed

152

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

Table 2 Planted tree species and productivity descriptions of the experimental sites. Experimental site

Species

Planting date

Age (year)

Planting spacing (m)

Stocking (tree ha
1)

Basal areaa (m2 ha
1)

Dominant height (m)

MAI (m3 ha
1 y
1)

Anhembi Itatinga Nova Ponte Telêmaco Borba

Brazilian native tree species Eucalyptus grandis Pinus caribaea var. hondurensis Pinus taeda

March-2004 December-2002 January-2008 February-2007

4.7 5.8 5.2 6.3

3.0  2.0 3.75  1.6 3.0  2.0 3.0  2.0

1667 1667 1667 1667

20.0 26.5 37.7 26.9

13.2 31.5 13.7 9.9

13.7 42.7 38.5 16.2

MAI: mean annual increment. a Basal area: at Itatinga, Nova Ponte, and Telêmaco Borba basal area was calculated based on measurements of bole diameter at 1.30 m above ground; at Anhembi was calculated based on bole diameter at 0.3 m above ground level.

and irrigated), with four replicates. For the current study, we selected trees under unthinned, not fertilized and rainfed treatments, which represents a typical commercial pine management in Brazil. Each plot has 12 rows  16 plants (1152 m2), with measurable plots of 8 rows  12 plants, with 3 m  2 m plant spacing. Twenty-four trees were select across these four plots (six trees per plot), comprising two suppressed (DBH = 12 cm, height = 9.7 m, aboveground biomass = 26.7 kg), two average (DBH = 15.8 cm, height = 10.1 m, aboveground biomass = 41.5 kg) and two dominant trees (DBH = 18 cm, height = 10.5 m, aboveground biomass = 44.2 kg). After the installation of the 24 dendrometer bands at DBH position and 30-day buffer for band adjustments on trees boles, the measurements were performed at 14-day intervals. 2.1.4. Telêmaco Borba-PR The study was conducted within of the same project Potential Productivity of Pine in Brazil (www.ipef.br/pppib), but located in the Southern Brazil, Paraná state, where the subtropical Pinus taeda is climatically suitable. The trial was planted in January 2007 with Pinus taeda seedlings from seed orchard located in Três Barras-SC, belonging to the Rigesa Company. Weed competition was controlled by applying glyphosate before planting and when necessary. Leaf-cutting ants (Atta spp. and Acromyrmex spp.) were controlled systematically with sulfluramid-based baits throughout the experimental area. The experimental design is a complete 2  2 factorial with two management (thinned and unthinned), two fertilization regimes (not fertilized and fertilized). For the current study, trees under unthinned and not fertilized treatments were selected, which represent a typical commercial pine management in Brazil. Each plot has 12 rows  16 plants (1152 m2), with measurable plots of 8 rows  12 plants, with 3 m  2 m plant spacing. Twenty-four trees were select across four plots (six trees per plot), comprising two suppressed (DBH = 12.6 cm, height = 7.7 m, aboveground biomass = 26.8 kg), two average (DBH = 14.3 cm, height = 8 m, aboveground biomass = 34.9 kg) and two dominant (DBH = 17 cm, height = 8.6 m, aboveground biomass = 43.2 kg). After the installation of the 24 dendrometers at DBH position and 30-day buffer for band adjustments on trees boles, the measurements were performed at 30-day intervals. 2.2. Meteorological data Meteorological data of the experimental sites was obtained from automatic weather stations from different sources: (i) at Anhembi, the data set were obtained from AFRS’ weather station; (ii) at Itatinga, the data set were obtained from flux tower of Eucflux project; (iii) at Nova Ponte and Telêmaco Borba, data sets were provided by the Brazilian National Institute of Meteorology, considering the nearest location for each site. Raw meteorological data were processed and consolidated from hourly to daily time scale. Daily data from each station were organized to match the period of dendrometric measurements and compiled in the

respective days of growth for the maximum air temperature (Tmx), minimum air temperature (Tmn), air relative humidity (RH) and rainfall (R). Vapor pressure deficit (VPD) was calculated from Tmx, Tmn and RH (Tetens, 1930; Allen et al., 1998). Potential evapotranspiration (PET) of each period and for each site was calculated based on method of Thornthwaite (1948). Water balance (WB) calculated by Thornthwaite and Mather (1955) method continues to be widely used (Rodríguez-García et al., 2015) for interpretation of the crop growth and productivity. We calculated the sequential WB (SWB) of each experimental site (Pereira, 2005; Rolim et al., 1998) and obtained the actual evapotranspiration (AET), surplus, and water deficit. Telêmaco Borba (the coldest site) was used to calculate basal minimum temperature (Tb, air temperature at which growth stops). We applied regression analysis using the same dataset from P. taeda growth and monthly mean minimum air temperature of each observation period. CSA change of the three size classes was combined in a single average growth for each measurement period. All measurements within periods with soil water deficit were excluded to specifically evaluate the effect of air temperature (Boyer, 1970). 2.3. Data analysis For all sampled trees from each experimental site, the bole cross-sectional area (CSA) was converted from bole circumference data provided by the dendrometer bands readings. Current CSA growth was calculated by subtracting the prior from the current measurement and cumulated CSA was calculated by summing all current growth values across the study period of each experimental site. We applied multivariate linear analysis to develop regressions to estimate the CSA growth rate, as function of meteorological variables, for each group of trees (size class or successional status) of the different species and locations (Eq. (1)). CSA and cumulative climate variables were converted to monthly basis.

CSA ¼ b0 þ b1 Tmx þ b2 Tmn þ b3 R þ b4 PET þ b5 AET þ b6 VPD 2


1

ð1Þ
1

where CSA = cross-sectional area increment (cm tree month ); Tmx = maximum air temperature (°C); Tmn = minimum air temperature (°C); R = rainfall (mm); PET = potential evapotranspiration (mm); AET = actual evapotranspiration (mm); VPD = vapor pressure deficit (kPa); b0 to b6 are coefficients of the multivariate regression equation. Regressions were developed using the software SigmaPlot 12 (Systat Software, San Jose, CA, USA) and the backward method using multivariate regression technique, considering a 5% probability. The motivation to eliminate independent variables is based on the biases and loss of predictability that are introduced when irrelevant variables are added to the regression. Additionally to the adjusted determination coefficient (R2adj), we used root mean square error (RMSE) and mean absolute percentage error (MAPE) to evaluate the growth regressions.


5.429**
11.855**
6.899**

0.94 0.93 0.90 0.95 6.37 8.58 7.16 5.55 2.69 2.44 5.09 2.02 <0.001 <0.001 <0.001 <0.001 0.71 0.64 0.88 0.82

0.95 0.94 0.90 0.88 6.90 7.49 11.23 14.12 5.15 6.87 14.61 14.34 <0.001 <0.001 <0.001 <0.001 0.81 0.81 0.83 0.85


0.096* 0.016** 0.01** 0.747** 0.57**

1.984** 1.954** 0.474**

Suppressed Average Dominant Grouped Size class Telêmaco Borba


27.339**
19.063**
15.243**
11.35**

0.663** 0.914**

0.538** 0.805**


4.293*
12.545*
3.457*
6.474** Size class Nova Ponte

Suppressed Average Dominant Grouped

Size class Itatinga

Suppressed Average Dominant Grouped

32.07* 0.132n.s 19.379* P NP Grouped Successional status1 Anhembi

1.262** 8.738** 10.582** 6.407**


1.031


0.212**


0.044**

0.01**

0.001**

* *


0.012


0.022*
1.737*


0.153**

0.015* 0.028* 0.036*

0.272** 0.016* 0.159**


2.69**
6.317**
4.593**
4.672**

0.94 0.96 0.94 0.96 10.60 6.19 9.61 7.81 0.28 1.76 5.22 1.88 <0.001 <0.001 <0.001 <0.001 0.54 0.71 0.70 0.64

0.93 0.87 0.92 0.91 0.45 0.89

0.001 0.01 0.001

1.23 0.82 0.93

RMSE (cm2 tree
1) p-value R2adj VPD (cm2 tree
1 kPa
1) PET R Tmn

(cm2 tree
1 °C
1)

Tmx

b0 (cm2 tree
1) Tree groups Site

Table 3 Coefficients of CSA (cross sectional area) growth equations and statistical indices by experimental site and tree group.

(cm2 tree
1 mm
1)

AET

3. Results Bole CSA growth was influenced by meteorological variables (p 6 0.01) at all experimental sites, forest types (eucalypts, pine, and Brazilian native species), and tree groups (successional status or tree size) (Table 3). All regressions showed optimum performance (Pi > 0.87). At Anhembi (Cfa climate), growth of the 10 Brazilian native tree species was affected by maximum air temperature, rainfall, and actual evapotranspiration for pioneer species, and only actual evapotranspiration for non-pioneer species (Table 3). Growth rates of pioneer species, compared to non-pioneer, were higher across the study period, reaching up to 4.5 times larger in February 2009, month with high rainfall (Fig. 2). However, during August 2009, after three months of soil water deficit, growth of pioneer and non-pioneer species were similar. At Itatinga experimental site (Cfa climate), eucalypt trees (all tree size classes grouped) were most responsive to VPD (Table 3). The largest trees showed a highly variable growth rate that most closely corresponded to VPD and rainfall. Small trees showed a little growth response despite changes in meteorological variables (Table 3). Large trees (CSA = 398 cm2 tree
1) showed the highest growth rate in January 2009, during a period of high rainfall and low VPD, reaching 5.7 cm2 tree
1 (14-day measurements interval). Growth almost stopped (<0.5 cm2 tree
1) for all tree size classes during a period in late 2008 with low rainfall and soil water deficit (Fig. 3). Across the period, large trees showed growth rates 8–60 times higher than small trees. At Nova Ponte (Cwb climate), pine trees (all tree size classes grouped) were responsive to minimum air temperature and VPD, and when separated into size classes, actual evapotranspiration was also significant (Table 3). Growth among tree size classes were similar and large trees did not always show the highest growth rate during the measurements period. The largest differences between small and large classes reached up to 2-fold. Large trees showed the highest growth rate (6.72 cm2 tree
1, 14-day measurements intervals) in February 2013 at low VPD and high soil water availability (Fig. 4). Growth stopped twice for all tree size classes during the study period, at August–September 2012 and July– September 2013. During these months, trees experienced high VPD and soil water deficit (Fig. 4). At Telêmaco Borba (Cfb climate), the growth of pines (all tree size classes grouped) were responsive to maximum air temperature and rainfall, however separating into tree size classes, different meteorological variables were significant (Table 3). For small trees, maximum and minimum air temperature, and potential evapotranspiration. For average trees, minimum air temperature

5.15 11.57 6.32

MAPE (%)

Pi

The performance of the growth regressions was also tested using the Performance index ‘‘Pi” (Alvares et al., 2013b), which is the product of the coefficient of correlation ‘‘r”, which defines regression precision, and refined agreement index ‘‘dr” (Willmott et al., 2012), which defines regression accuracy. The criteria for interpreting the ‘‘Pi” is: Pi P 0.75 – optimum performance; 0.6 6 Pi < 0.75 – very good performance; 0.45 6 Pi < 0.6 – good performance; 0.3 6 Pi < 0.45 – tolerable performance; 0.15 6 Pi < 0.3 – poor performance; 0 6 Pi < 0.15 – bad performance; and Pi < 0 – very bad performance (Alvares et al., 2013b). To test growth regressions and given that dendrometric data are scarce, we performed an approximate validation as follows: growth regressions were applied to each growth period to get the calculated cumulated CSA for each species and each tree group. Then, these results were plotted against the observed cumulated CSA. The comparison between calculated and observed cumulated CSA was also tested by the Performance index.

Multivariate equation: CSA ¼ b0 þ b1 Tmx þ b2 Tmn þ b3 R þ b4 PET þ b5 AET þ b6 VPD. P = pioneer species, NP = Non-pioneer species; CSA = cross-sectional area (cm2 tree
1 month
1); Tmx = maximum air temperature (°C); Tmn = minimum air temperature (°C); R = rainfall (mm); PET = potential evapotranspiration (mm); AET = actual evapotranspiration (mm); VPD = vapor pressure deficit (kPa); a0 to a6 are coefficients of the multivariate regression equation. n.s Not significant at the 0.05 level. * Significant at 0.05 level. ** Significant at 0.01 level. 1 The classification of pioneer and non-pioneer species is related to the concurrency of the species during the process of secondary succession of tropical forest ecosystems.

153

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

154

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

Fig. 2. Cumulated and current bole cross-sectional area (CSA) growth of native tree species, sequential water balance and climate (top to bottom) at Anhembi. Mean CSA: pioneers = 303 cm2 tree
1; non-pioneers = 153 cm2 tree
1.

and rainfall. For large trees, maximum temperature and rainfall. Similarly to Nova Ponte, Telêmaco Borba pine trees alternate in showing the highest growth rate over the measurement period. Growth almost stopped (<0.5 cm2 tree
1) in July 2012 and August 2013, during periods of low air temperature, however, they were followed by the highest growth rates (>10 cm2 tree
1, 30-day measurements intervals), occurred in September 2012 and October 2013 (Fig. 5). The rate of CSA increment and minimum air temperature showed a significant relationship at Telêmaco Borba (R2 = 0.90,

p < 0.0001, Fig. 6). During winter, growth was below 1 cm2 tree
1 month
1 and at the end of the season, growth almost stopped (Fig. 5). For zero increment, when y = 0, minimum air temperature was 10.4 °C. Minimum air temperature showed increasing standard deviation (Fig. 6), indicating that the winter at this region has high within-day variability. In July 2013 the average minimum air temperature was 10.7 °C, however, the absolute minimum air temperature was
2.1 °C, and average maximum air temperature reached 15.6 °C. During this period, growth was approximately 0.5 cm2 tree
1.

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

155

Fig. 3. Cumulated and current bole cross-sectional area (CSA) growth of Eucalyptus grandis, sequential water balance and climate (top to bottom) at Itatinga. Tree size class (DBH): Suppressed = 10 cm; Average = 17.5 cm; Dominant = 22.5 cm.

The comparison between calculated and observed cumulated CSA showed optimum performance (Pi > 0.92) for all growth regressions (Fig. 7). Calculated cumulated CSA was slightly underestimated for P. caribaea var. hondurensis.

4. Discussion In tropical regions, the seasonality of tree bole radial growth has been directly correlated with the hydrological regime. However, in subtropical climates, as rainfall is more evenly distributed across

the year, the wide range in air temperature over the year plays an important role influencing the cambial activity of trees (Worbes, 1995, 1999; Botosso et al., 2000; Brienen and Zuidema, 2005; Lisi et al., 2008; Callado et al., 2013; Shimamoto et al., 2016). Our experimental sites are distributed across a transition region of different climatic types, with a wide variation of air temperature, night-time temperature, relative humidity, and rainfall (Alvares et al., 2013a,b, 2015, 2016), resulting in tree growth responding to different meteorological variables. However, considering the behavior of the curves (Figs. 2–5), all species at all sites were similarly responsive to meteorological seasonality.

156

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

Fig. 4. Cumulated and current bole cross-sectional area (CSA) growth of Pinus caribaea var. hondurensis, sequential water balance and climate (top to bottom) at Nova Ponte. Tree size class (DBH): Suppressed = 12 cm; Average = 15.8 cm; Dominant = 18 cm.

Vapor pressure deficit (VPD) was the only meteorological variable negatively affecting growth at all sites and species (Table 3). The pattern of negative impact of VPD on stomatal conductance, decreasing CO2 uptake, and consequently reducing carbohydrates production and bole growth is observed across tree species and meteorological conditions (Lin et al., 2015). Over the study period at all sites, nearly zero growth was directly related to high values of VPD and soil water deficit. Maximum and minimum air temperatures were also related to bole growth of all species at all sites (Table 3). Temperature has a

great impact on several physiological processes, notably on leaf heating processes, increasing transpiration, and on the enzymatic dynamics of the photosynthesis and respiration (Lin et al., 2015; Atkin et al., 2015). At Anhembi, pioneers were more responsive than non-pioneers to meteorological variables, showing higher absolute growth rates over the studied period (Fig. 2). Studies developed at the same experimental site showed that growth differences results from the successional status of the species, due to higher photosynthetic rates, total chlorophyll, and nitrogen content in pioneers,

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

157

Fig. 5. Cumulated and current bole cross-sectional area (CSA) growth of Pinus taeda, sequential water balance and climate (top to bottom) at Telêmaco Borba. Tree size class (DBH): Suppressed = 12.6 cm, Average = 14.3 cm; Dominant = 17 cm.

compared to non-pioneers (Campoe et al., 2014). The variability in growth, as a function of within year meteorological seasonality, due to variation in leaf level photosynthetic processes was observed in eight tropical tree species from different successional status planted in Panamá (Craven et al., 2011). In general, under tropical meteorological conditions, water availability due to rainfall is the most important variable directly affecting carbon uptake processes, resulting in variation of growth rates (Wagner et al., 2014; Venegas-González et al., 2016).

Higher responsiveness to resources availability and productivity of dominant trees, compared to suppressed trees, is a pattern observed by several tree species under different silvicultural managements or field conditions. Larger trees are more productive due to their ability to absorb higher amounts of the available resource and additionally higher capacity to convert the absorbed resource into biomass (Binkley et al., 2013). Using MAESTRA model (Medlyn, 2004) and an approach focusing on single trees within forest stands, several studies confirmed this theory for Eucalyptus

158

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

Fig. 6. Relationship between bole cross sectional area (CSA) change and average minimum air temperature for Pinus taeda at Telêmaco Borba experimental site. Bars represent standard deviation of daily minimum air temperature for each measurement period.

Fig. 7. Relationship between observed and estimated cumulated bole cross sectional area (CSA) for all species. Estimated cumulated bole CSA are results of the growth regressions at each site. The comparison was tested by Performance index (Pi).

(Binkley et al., 2010; Campoe et al., 2013a; Forrester et al., 2013), Pinus (Campoe et al., 2013b), and Acacia (le Maire et al., 2013). Dominant Eucalyptus trees at Itatinga were highly responsive to within and among year meteorological seasonality, in accordance to resources availability, compared to the suppressed trees, which showed an almost flat growth line over the period (Fig. 3). Dominant Eucalyptus trees showed higher responsiveness to rainfall and VPD seasonality and productivity as a result of increased ability for light absorption (higher leaf area) and also higher capacity to convert absorbed light into biomass production (higher light use efficiency), compared to suppressed trees (Campoe et al., 2013a). At both experimental sites with pine plantations (Nova Ponte and Telêmaco Borba), dominant trees were also more productive and responsive to meteorological variations than sup-

pressed trees. Studies show that the higher growth and responsiveness of the dominant pine trees is a result of increased light absorption and light use efficiency (Campoe et al., 2013b). At Nova Ponte, the experimental site with higher soil water deficit occurring in two consecutive years and lasting for months, VPD was negatively related to bole growth of all pine size classes (Table 3). Over both periods, growth stopped as a result of water limitation (Fig. 4). Similarly to our study, Ward et al. (2015) found that water stress simulated by throughfall exclusion reduced wood productivity of pine plantations in southeastern USA. Climate at the other pine site (Telêmaco Borba) provided soil water availability almost across the whole study period (Fig. 5), resulting in significant relation between bole growth and rainfall, and additionally to minimum and maximum air temperature (Table 3). No relation with VPD was observed. In a similar Köppen’s climate type, Sixel et al. (2015) studied the sustainability of wood productivity of P. taeda and found that the high forest productivity (38 m3 ha
1 year
1) indicates that southern Brazilian conditions were suitable for tree growth. No drought was observed in the region, and the seasonal air temperature was considered to be the main limiting factor. Different tree species show specific temperature ranges (optimum temperature) to express maximum potential growth. Air temperature beyond the specific range have a negative impact on growth (Yin et al., 1995). There are few references on basal temperatures for tree species (Rawal et al., 2014; Watt et al., 2014), however, in Brazil, these ranges of basal minimum temperature are still unknown. Under field conditions in tropical regions, the basal minimum temperatures can be determined, especially in areas of high altitude and high latitude during specific seasons, i.e., between late fall and mid-winter. Basal minimum temperature is an important parameter in process-based models such as 3-PG (Bryars et al., 2013), and our results enable improved parameterization for P. taeda in southern Brazil. Understand the impacts of within year meteorological seasonality on growth of different tree species in different regions is a key step to understand limits on wood production and restoration. Simulations suggest future climatic changes, with significant variations in quantity and distribution of rainfall, and also increase in air temperature in South America (Chou et al., 2014). If the simulations are confirmed, the productivity of the studied forest plantations has high probability to be negatively affected. These predicted changes will reduce soil water availability and increase VPD, reducing growth of all studied species. Additionally to simulated future climatic changes, most of the new forest plantations in Brazil are moving towards north and northeast regions of the country (Gonçalves et al., 2013). These regions currently show difficulties related to water availability, and intense and prolonged droughts (Alkama and Cescatti, 2016). Improving our understanding on how different tree species respond to meteorological variation is essential to ensure long-term viability of forest plantations without compromising productivity. Long-term studies assessing the responses of tree species to meteorological variability (more than one year) are highly recommended and very important to increase the probability to capture larger ranges of meteorological conditions. These datasets are valuable for calibration and validation of ecophysiological models, improving the ability to model and predict the impact of climate on tree growth.

5. Conclusions High frequency of DBH measurements can provide accurate growth data to understand the impact of meteorological variability on forest production within and among years, regardless climatic

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

region, forest type (eucalypts, pine, and Brazilian native species) and tree groups (successional status and tree size). The studied trees, from different sizes and successional status, showed different levels of responses to meteorological variables; therefore, is very important to select individuals representing all the different classes of trees. All species responded to multiple meteorological variables related to air temperature and water availability, and dominant trees were more responsive, compared to suppressed trees. Acknowledgments To the Forestry Science and Research Institute (IPEF) which coordinates the Forest Restoration project, the Eucflux project and the PPPIB project. We thank to the Brazilian National Institute of Meteorology (INMET) for the meteorological data provided. We also thank to Duratex, Floragro, and Klabin staff for their technical support, and to Kevin Hall for the English review. References Alkama, R., Cescatti, A., 2016. Biophysical climate impacts of recent changes in global forest cover. Science 351, 600–604. Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop evapotranspiration: guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 56. FAO, Rome. Alvares, C.A., Mattos, E.M., Sentelhas, P.C., Miranda, A.C., Stape, J.L., 2015. Modeling temporal and spatial variability of leaf wetness duration in Brazil. Theor. Appl. Climatol. 120, 455–467. Alvares, C.A., Sentelhas, P.C., Mattos, E.M., Miranda, A.C., Moraes, W.B., Silva, P.H.M., Furtado, E.L., Stape, J.L., 2016. Climatic favourability zones for Eucalyptus rust in Brazil. Forest Pathol. http://dx.doi.org/10.1111/efp.12301. Alvares, C.A., Stape, J.L., Sentelhas, P.C., Gonçalves, J.L.M., Sparovek, G., 2013a. Köppen climate classification map for Brazil. Meteorol. Z. 22, 711–728. Alvares, C.A., Stape, J.L., Sentelhas, P.C., Gonçalves, J.L.M., 2013b. Modeling monthly mean air temperature for Brazil. Theor. Appl. Climatol. 113, 407–427. Alvarez, J., Allen, H.L., Albaugh, T.J., Stape, J.L., Bullock, B.P., Song, C., 2012. Factors influencing the growth of radiate pine plantations in Chile. Forestry 86, 13–26. Atkin, O.K. et al., 2015. Global variability in leaf respiration in relation to climate, plant functional types and leaf traits. New Phytol. 206, 614–636. Battie-Laclau, P., Laclau, J.P., Domec, J.C., Christina, M., Bouillet, J.P., Piccolo, M.C., Gonçalves, J.L.M., Moreira, R.M., Krusche, A.V., Bouvet, J.M., Nouvellon, Y., 2014. Effects of potassium and sodium supply on drought-adaptive mechanisms in Eucalyptus grandis plantations. New Phytol. 203, 401–413. Binkley, D., Stape, J.L., Ryan, M.G., 2004. Thinking about efficiency of resource use in forests. Forest Ecol. Manage. 193 (1–2), 5–14. Binkley, D., Stape, J.L., Bauerle, W.L., Ryan, M.G., 2010. Explaining growth of individual trees: light interception and efficiency of light use by Eucalyptus at four sites in Brazil. Forest Ecol. Manage. 259, 1704–1713. Binkley, D., Campoe, O.C., Gspaltl, M., Forrester, D.I., 2013. Light absorption and use efficiency in forests: why patterns differ for tree and stands. Forest Ecol. Manage. 288, 5–13. Boisvenue, C., Running, S.W., 2006. Impacts of climate change on natural forest productivity: evidence since the middle of the 20th century. Glob. Change Biol. 12, 862–882. Booth, T.H., 2013. Eucalypt plantations and climate change. Forest Ecol. Manage. 301, 28–34. Botosso, P.C., Vetter, R.E., Tomazello Filho, M., 2000. Periodicidade e taxa de crescimento de árvores de cedro (Cedrela odorataL., Meliaceae), jacareuba (Calophyllum angulare A.C. Smith, Clusiaceae) e muirapiranga (Eperua bijiga Mart. ExBenth, Leg. Caesalpinioideae) de floresta de terra firme, em ManausAM. In: Roig, F.A. (Ed.), Dendrocronología Em América Latina. EDIUNC, Mendoza, pp. 357–380. Boyer, W.D., 1970. Shoot growth patterns of young Loblolly pine. Forest Sci. 16, 471–482. Brienen, R.J.W., Zuidema, P.A., 2005. Relating tree growth to rainfall in Bolivian rain forests: a text for six species using tree ring analysis. Oecologia 146, 1–12. Bryars, C., Maier, C., Zhao, D., Kane, M., Borders, B., Will, R., Teskey, R., 2013. Fixed physiological parameters in the 3-PG model produced accurate estimates of loblolly pine growth on sites in different geographic regions. Forest Ecol. Manage. 289, 501–514. Callado, C.H., Roig, F.A., Tomazello-Filho, M., Barros, C.F., 2013. Cambial growth periodicity studies of South American woody species – a review. IAWA J. 34, 213–230. Callado, C.H., Silva-Neto, S.J., Scarano, F.R., Costa, C.G., 2004. Radial growth dynamics of Tabebuia umbellate (Sond.) Sandwith (Bignoniaceae), a floodtolerant tree from the Atlantic forest swamps in Brazil. IAWA J. 25, 175–183. Campoe, O.C., Iannelli, C., Stape, J.L., Cook, R.L., Mendes, J.C.T., Vivian, R., 2014. Atlantic forest tree species responses to silvicultural practices in a degraded

159

pasture restoration plantation: from leaf physiology to survival and initial growth. Forest Ecol. Manage. 313, 233–242. Campoe, O.C., Stape, J.L., Nouvellon, Y., Laclau, J.-P., Bauerle, W.L., Binkley, D., Le Maire, G., 2013a. Stem production, light absorption and light use efficiency between dominant and non-dominant trees of Eucalyptus grandis across a productivity gradient in Brazil. Forest Ecol. Manage. 288, 14–20. Campoe, O.C., Stape, J.L., Albaugh, T.J., Allen, H.L., Fox, T.R., Rubilar, R., Binkley, D., 2013b. Fertilization and irrigation effects on tree level aboveground net primary production, light interception and light use efficiency in a loblolly pine plantation. Forest Ecol. Manage. 288, 43–48. Campoe, O.C., Stape, J.L., Laclau, J.-P., Marsden, C., Nouvellon, Y., 2012. Stand-level patterns of carbon fluxes and partitioning in a Eucalyptus grandis plantation across a gradient of productivity, in São Paulo State, Brazil. Tree Physiol. 32, 696–706. Campoe, O.C., Stape, J.L., Mendes, J.C.T., 2010. Can intensive management accelerate the restoration of Brazil’s Atlantic forests? Forest Ecol. Manage. 259, 1808– 1814. Chou, S.C., Lyra, A., Mourão, C., Dereczynski, C., Pilotto, I., Gomes, J., Bustamante, J., Tavares, P., Silva, A., Rodrigues, D., Campos, D., Chagas, D., Sueiro, G., Siqueira, G., Marengo, J., 2014. Assessment of climate change over South America under RCP 4.5 and 8.5 downscaling scenarios. Am. J. Clim. Change 3, 512–525. Cardoso, F.C.G., Marques, R., Botosso, P.C., Marques, M.C.M., 2012. Stem growth and phenology of two tropical trees in contrasting soil conditions. Plant Soil. 354, 269–281. Craven, D., Dent, D., Braden, D., Ashton, M.S., Berlyn, G.P., Hall, J.S., 2011. Seasonal variability of photosynthetic characteristics influences growth of eight tropical tree species at two sites with contrasting precipitation in Panama. Forest Ecol. Manage. 261, 1643–1653. Cristiano, P.M., Madanes, N., Campanello, P.I., Francescantonio, D., Rodriguez, S.A., Zhang, Y.J., Carrasco, L.O., Goldstein, G., 2014. High NDVI and potential canopy photosynthesis of South American subtropical forests despite seasonal changes in leaf area index and air temperature. Forests 5, 287–308. Donagemma, G.K., Campos, D.V.B., Calderano, S.B., Teixeira, W.G., Viana, J.H.M., 1997. Manual De Métodos De Análise De Solo, second ed. Centro Nacional de Pesquisa de Solos, Rio de Janeiro. Drew, D.M., O’Grady, A.P., Downes, G.M., Read, J., Worledge, D., 2008. Daily patterns of stem size variation in irrigated and unirrigated Eucalyptus globulus. Tree Physiol. 28, 1573–1581. Ferez, A.P., Campoe, O.C., Mendes, J.C.T., Stape, J.L., 2015. Silvicultural opportunities for increasing carbon stock in restoration of Atlantic forests in Brazil. Forest Ecol. Manage. 350, 40–45. Forrester, D.I., Collopy, J.J., Beadle, C.L., Baker, T.G., 2013. Effect of thinning, pruning and nitrogen fertiliser application on light interception and light-use efficiency in a young Eucalyptus nitens plantation. Forest Ecol. Manage. 288, 21–30. Gonçalves, J.L.M., Alvares, C.A., Higa, A.R., Silva, L.D., Alfenas, A.C., Stahl, J., Ferraz, S. F.B., Lima, W.P., Brancalion, P.H.S., Hubner, A., Bouillet, J.-P., Laclau, J.-P., Nouvellon, Y., Epron, D., 2013. Integrating genetic and silvicultural strategies to minimize abiotic and biotic constraints in Brazilian eucalypt plantations. Forest Ecol. Manage. 301, 6–27. Gonçalves, J.L.M., Silva, L.D., Behling, M., Alvares, C.A., 2014. Management of industrial plantations. In: Borges, J.G., Diaz-Balteiro, L., McDill, M.E., Rodriguez, L.C.E. (Eds.), The Management of Industrial Forest Plantations – Theoretical Foundations and Applications. Springer, Dordrecht, pp. 91–119. Grogan, J., Schulze, M., 2012. The impact of annual and seasonal rainfall patterns on growth and phenology of emergent tree species in Southeastern Amazonia, Brazil. Biotropica 44, 331–340. Janowiak, M.K., Swanston, C.W., Nagel, L.M., Brandt, L.A., Butler, P.R., Handler, S.D., Shannon, P.D., Iverson, L.R., Mattews, S.N., Prasad, A., Peters, M.P., 2014. A practical approach for translating climate change adaptation principles into forest management actions. J. Forest. 112, 424–433. King, G., Fonti, P., Nievergelt, D., Buntgen, U., Frank, D., 2013. Climatic drivers of hourly to yearly radius variations along a 6 °C natural gradient. Agric. Forest Meteorol. 168, 36–46. Le Maire, G., Nouvellon, Y., Christina, M., Ponzoni, F.J., Gonçalves, J.L.M., Bouillet, J.P., Laclau, J.-P., 2013. Tree and stand light use efficiencies over a full rotation of single- and mixed-species Eucalyptus grandis and Acacia mangium plantations. Forest Ecol. Manage. 288, 31–42. Lisi, C.S., Tomazello Filho, M., Botosso, P.C., Roig, F.A., Maria, V.R.B., Ferreira-Fedele, L., Voigt, A.R.A., 2008. Tree-ring formation, radial increment periodicity, and phenology of tree species from a seasonal semi-deciduous forest in southeast Brazil. IAWA J. 29, 189–207. Lin, Y. et al., 2015. Optimal stomatal behaviour around the world. Nature Clim. Change 5, 459–464. Medlyn, B.E., 2004. Forests at the land–atmosphere interface. In: Mencuccini, M., Grace, J., Moncrieff, J., McNaughton, K.G. (Eds.), A MAESTRO Retrospective. CABI, Edinburgh, pp. 105–122. Morellato, L.P.C., Haddad, C.F.B., 2000. Introduction: the Brazilian Atlantic Forest. Biotropica 32, 786–792. Pereira, A.R., 2005. Simplificando o balanço hídrico de Thornthwaite-Mather. Bragantia 64, 311–313. Rawal, D.S., Kasel, S., Keatley, M.R., Aponte, C., Nitschke, C.R., 2014. Environmental effects on growth phenology of co-occurring Eucalyptus species. Int. J. Biometeorol. 58, 427–442. Rodrigues, R.R., Lima, R.A.F., Gandolfi, S., Nave, A.G., 2009. On the restoration of high diversity forests: 30 years of experience in the Brazilian Atlantic forest. Forest Ecol. Manage. 142, 1242–1251.

160

O.C. Campoe et al. / Forest Ecology and Management 380 (2016) 149–160

Rodríguez-García, A., Martín, J.A., López, R., Mutke, S., Pinillos, F., Gil, L., 2015. Influence of climate variables on resin yield and secretory structures in tapped Pinus pinaster Ait. in central Spain. Agric. Forest Meteorol. 202, 83–93. Rolim, G.D.S., Sentelhas, P.C., Barbieri, V., 1998. Planilhas no ambiente EXCEL TM para oscálculos de balanços hídricos: normal, seqüencial, de cultura e de produtividade real e potencial. Revista Brasileira de Agrometeorologia 6, 133– 137. Rowland, L., Malhi, Y., Silva-Espejo, J.E., Farfan-Amezquita, F., Halladay, K., Doughty, C.E., Meir, P., Phillips, O.L., 2014. The sensitivity of wood production to seasonal and interannual variations in climate in a lowland Amazonian rainforest. Oecologia 174, 295–306. Sette Jr., C., Tomazello Filho, M., Lousada, J.L., Laclau, J.-P., 2012. Sazonalidade do incremento em diâmetro do tronco de árvores de Eucalyptus grandis pelo uso de dendrômetros. Ciência Florestal. 22, 763–777. Shimamoto, C.Y., Botosso, P.C., Amano, E., Marques, M.C., 2016. Stem growth rhythms in trees of a tropical rainforest in Southern Brazil. Trees 30, 99–111. Sixel, R.M.M., Arthur Junior, J.C., Gonçalves, J.L.M., Alvares, C.A., Andrade, G.R.P., Azevedo, A.C., Stahl, J., Moreira, A.M., 2015. Sustainability of wood productivity of Pinus taeda based on nutrient export and stocks in the biomass and in the soil. Rev. Bras. Cienc. Solo 39, 1416–1427. Spanner, M.A., Pierce, L.L., Running, S.W., Peterson, D.L., 1990. The seasonality of AVHRR data of temperate coniferous forests: relationship with leaf area index. Remote Sens. Environ. 33, 97–112. Stape, J.L., Binkley, D., Ryan, M.G., Fonseca, S., Loos, R.A., Takahashi, E.N., Silva, C.R., Silva, S.R., Hakamada, R.E., Ferreira, J.M.D., Lima, A.M.N., Gava, J.L., Leite, F.P., Andrade, H.B., Alves, J.M., Silva, G.G.C., Azevedo, M.R., 2010. The Brazil Eucalyptus potential productivity project: influence of water, nutrients and stand uniformity on wood production. Forest Ecol. Manage. 259, 1684– 1694. Tetens, O., 1930. Uber einige meteorologische Begriffe. Z. Geophys. 6, 297–309. Thornthwaite, C.W., 1948. An approach toward a rational classification of climate. Geogr. Rev. 38, 55–94. Thornthwaite, C.W., Mather, J.R., 1955. The Water Balance. Drexel Institute of Technology – Laboratory of Climatology, Centerton.

Urrutia-Jalabert, R., Rossi, S., Deslauriers, A., Malho, Y., Lara, A., 2015. Environmental correlates of stem radius change in the endangered Fitzroya cupressoides forests of southern Chile. Agric. Forest Meteorol. 200, 209–221. van Raij, B., Andrade, J.C., Cantarella, H., Quaggio, J.A., 2001. Análise Química Para Avaliação Da Fertilidade De Solos Tropicais. Instituto Agronômico de Campinas, Campinas. Venegas-González, A., Chagas, M.P., Júnior, C.R.A., Alvares, C.A., Roig, F.A., Tomazello Filho, M., 2016. Sensitivity of tree ring growth to local and large-scale climate variability in a region of Southeastern Brazil. Theor. Appl. Climatol. 123, 233– 245. Wagner, F., Rossi, V., Aubry-Kientz, M., Bonal, D., Dalitz, H., Gliniars, R., Stahl, C., Trabucco, A., Hérault, B., 2014. Pan-tropical analysis of climatic effects on seasonal tree growth. PLoS one 9, 1–15. Ward, E.C., Domec, J.C., Laviner, M.A., Fox, T.R., Sun, G., McNulty, S., King, J., Noormets, A., 2015. Fertilization intensifies drought stress: water use and stomatal conductance of Pinus taeda in a midrotation fertilization and throughfall reduction experiment. Forest Ecol. Manage. 355, 72–82. Watt, M.S. et al., 2014. Using seasonal measurements to inform ecophysiology: extracting cardinal growth temperatures for process-based growth models of five Eucalyptus species/crosses from simple field trials. New Zeal. J. For. Sci. 44, 1–11. Willmott, C.J., Robeson, S.M., Matsuura, K., 2012. A refined index of model performance. Int. J. Climatol. 32, 2088–2094. Worbes, M., 1995. How to measure growth dynamics in tropical trees – a review. IAWA J. 16, 337–351. Worbes, M., 1999. Annual growth rings, rainfall-dependent growth and long-term growth patterns of tropical trees from the Caparo Forest Reserve in Venezuela. J. Ecol. 87, 391–403. Wunder, J., Fowler, A.M., Cook, E.R., Pirie, M., McCloskley, S.P.J., 2013. On the influence of tree size on the climate–growth relationship of New Zealand kauri (Agathis australis): insights from annual, monthly and daily growth patterns. Trees 27, 937–948. Yin, X., Kropff, M.J., McLaren, G., Visperas, R.M., 1995. A nonlinear model for crop development as a function of temperature. Agric. Forest Meteorol. 77, 1–16.