Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de...

15
Biomass estimations in forests of different disturbance history in the Atlantic Forest of Rio de Janeiro, Brazil Andre ´ Lindner Dietmar Sattler Received: 21 October 2010 / Accepted: 9 August 2011 / Published online: 18 August 2011 Ó Springer Science+Business Media B.V. 2011 Abstract Tropical forests are large reservoirs of biomass and there is a need for infor- mation on existing carbon stocks in these ecosystems and especially the effects of logging on these stocks. Reliable estimates of aboveground biomass stocks within the Atlantic Forest are rarely available. Past human disturbance is an important factor affecting forest structure variation and biomass accumulation among tropical forest ecosystems. To sup- port the efforts of improving the quality of estimations of the current and future biomass carbon storage capacity of this disturbed forest region we tested a non-experimental small scale approach to compare the aboveground tree biomass (AGB) of forest sites. Three sites with known disturbance histories have been investigated: complete cut down, selective logging and conservation since 70 years. The woody plant community (dbh C 10 cm) was censused and canopy openness in conjunction with leaf area index has been obtained by hemispherical photographs at each site. Estimates of aboveground tree biomass have been carried out using an allometric equation for moist tropical forests already applied for the study area. Additionally, a FAO standard equation has been employed for crosschecking our results. We identified significant differences in recent AGB of the three compared forest sites. With 313 (±48 Mg ha -1 ) the highest AGB-values have been found in the preserved forest area within a National Park, followed by 297 (±83) Mg ha -1 at the former clear cut site. Lowest AGB has been calculated for the area with past selective logging: 204 (±38) Mg ha -1 . Values calculated with the FAO standard equation showed the same trend but at a lower AGB level. Our results based an a small scale approach suggest that biomass productivity can recover in a forest which was completely cleared 60 years ago to reach AGB values up to a level that almost represents the situation in a preserved forest. A. Lindner (&) Department of Systematic Botany and Functional Biodiversity, Institute for Biology I, University of Leipzig, Johannisallee 21-23, 04103 Leipzig, Germany e-mail: [email protected] D. Sattler Department of Physical Geography and Geo-Ecology, Institute for Geography, University of Leipzig, Johannisallee 19, 04103 Leipzig, Germany e-mail: [email protected] 123 New Forests (2012) 43:287–301 DOI 10.1007/s11056-011-9281-9

Transcript of Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de...

Page 1: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

Biomass estimations in forests of different disturbancehistory in the Atlantic Forest of Rio de Janeiro, Brazil

Andre Lindner • Dietmar Sattler

Received: 21 October 2010 / Accepted: 9 August 2011 / Published online: 18 August 2011� Springer Science+Business Media B.V. 2011

Abstract Tropical forests are large reservoirs of biomass and there is a need for infor-

mation on existing carbon stocks in these ecosystems and especially the effects of logging

on these stocks. Reliable estimates of aboveground biomass stocks within the Atlantic

Forest are rarely available. Past human disturbance is an important factor affecting forest

structure variation and biomass accumulation among tropical forest ecosystems. To sup-

port the efforts of improving the quality of estimations of the current and future biomass

carbon storage capacity of this disturbed forest region we tested a non-experimental small

scale approach to compare the aboveground tree biomass (AGB) of forest sites. Three sites

with known disturbance histories have been investigated: complete cut down, selective

logging and conservation since 70 years. The woody plant community (dbh C 10 cm) was

censused and canopy openness in conjunction with leaf area index has been obtained by

hemispherical photographs at each site. Estimates of aboveground tree biomass have been

carried out using an allometric equation for moist tropical forests already applied for the

study area. Additionally, a FAO standard equation has been employed for crosschecking

our results. We identified significant differences in recent AGB of the three compared

forest sites. With 313 (±48 Mg ha-1) the highest AGB-values have been found in the

preserved forest area within a National Park, followed by 297 (±83) Mg ha-1 at the former

clear cut site. Lowest AGB has been calculated for the area with past selective logging: 204

(±38) Mg ha-1. Values calculated with the FAO standard equation showed the same trend

but at a lower AGB level. Our results based an a small scale approach suggest that biomass

productivity can recover in a forest which was completely cleared 60 years ago to reach

AGB values up to a level that almost represents the situation in a preserved forest.

A. Lindner (&)Department of Systematic Botany and Functional Biodiversity, Institute for Biology I,University of Leipzig, Johannisallee 21-23, 04103 Leipzig, Germanye-mail: [email protected]

D. SattlerDepartment of Physical Geography and Geo-Ecology, Institute for Geography, University of Leipzig,Johannisallee 19, 04103 Leipzig, Germanye-mail: [email protected]

123

New Forests (2012) 43:287–301DOI 10.1007/s11056-011-9281-9

Page 2: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

Selective logging may slow down AGB accumulation and the effect is measurable after

several decades.

Keywords Aboveground live biomass � Tropical forests � Atlantic Forest � Land use �Forest succession

Introduction

Despite the fact that the Atlantic Forest is probably the region with the highest species

diversity and degree of endemism throughout South America (Tabarelli et al. 2005; Silva

and Casteleti 2003), this forest was one of the largest of the subcontinent and covered

about 1.2 million km2. Its remaining remnants cover nowadays, depending on the defini-

tion of ‘‘Atlantic Forest’’ and the spatial methods employed, about 7–16% of its original

extent (Galindo-Leal and de Gusmao Camara 2003; Tabarelli et al. 2005; SOS Mata

Atlantica 2009; Ribeiro et al. 2009). However, the spatial distribution of these remnants

gives a more meaningful representation of the situation than the total amount of remaining

forest. According to Ribeiro et al. (2009), the remnants of the so called ‘‘Mata Atlantica’’

are currently distributed in more than 245,000 forest fragments of which 83.4% are smaller

than 50 ha and only 0.03% are larger than 10,000 ha. In addition to their degree of

fragmentation, these remaining forests are frequently to be found in varying stages of

degeneration caused by anthropogenic disturbance. Forest types with different disturbance

history are very heterogeneous habitats, continuously changing in structure and varying in

species composition (e.g., Williams-Linera et al. 1998; Laurance et al. 2002; Galindo-Leal

and de Gusmao Camara 2003). If these forest remnants cover 7–16% of the original extent

of the Mata Atlantica biome, do they represent 7–16% of the original forests biomass?

Probably not, because there are indications that the biomass of a large number of small

forest fragments left over after habitat fragmentation is much less then in a continuous

forest of the same overall size (Groeneveld et al. 2009).

Compared to other neotropical forests (e.g., Amzonian Rain Forests), reliable estimates

of aboveground biomass stocks within the Atlantic Forest are rarely available. The few

existing studies estimating aboveground biomass in this region have been carried out

mainly in relatively undisturbed and protected Atlantic Forest of different elevations (e.g.,

Rolim et al. 2005; Alves et al. 2010). Two studies known to the authors have been carried

out in secondary forests of the Mata Atlantica region using direct measurements by har-

vesting and weighting dry biomass (Burger and Delitti 2008; Tiepolo et al. 2002). Nev-

ertheless, particularly the effects of different disturbance histories on forest structure and

biomass stocks seem to be essential in the Amazon (Nogueira et al. 2008; Kellner and

Asner 2009), but are poorly understood in fragments of the Atlantic Forest (Kindel et al.

1999).

There is an alarming loss of biomass due to logging activities (Houghton 2005) and

nowadays about 60% of all tropical forests are secondary or degraded (ITTO 2002). Hence,

it becomes increasingly important to evaluate biomass stocks of recovering disturbed forest

(Frolking et al. 2009; Saldarriaga and Uhl 1991; Uhl and Buschbacher 1985). In our study

region, many of the forest remnants are in fact secondary re-growth of formerly cut down

or strongly disturbed Atlantic Forest (Pedreira et al. 2009).

With this study we address the hypothesis that even small scale, non-experimental

approaches using existing information on past disturbance can be used to reveal differences

in aboveground tree biomass (AGB) accumulation. Using the existing heterogeneous

288 New Forests (2012) 43:287–301

123

Page 3: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

situation we compared the AGB of forest sites with three distinct disturbance histories:

complete cut down, selective logging and conservation since 70 years. In addition to

historic information we used structural parameters such as stem density per hectare, tree

height classes, leaf area index (LAI) and canopy openness as possible proxies for the

disturbance history. Because of the fact that large tree individuals are known to provide a

disproportionally high contribution regarding aboveground biomass in tropical forests

(Vieira et al. 2004; Midgley and Niklas 2004; Chambers et al. 2001), we analyzed forest

composition with a special regard on these structural elements. As indirect AGB estima-

tions are carried out using allometric models, the results depend essentially from the

calculation method applied. To verify the AGB values calculated for the three different

forest sites with an allometric model based on dbh only, we performed an additional

standard FAO calculation including average wood density.

Study area

The study was conducted in two areas within the Atlantic Forest (Mata Atlantica) of the

state of Rio de Janeiro, Brazil (Fig. 1). The first research area is the 5,500 ha NGO reserve

‘‘Reserva Ecologica de Guapiacu—REGUA’’ (22� 25053 S, 42� 45020 W), located in the

municipality of Cachoeiras de Macacu about 100 km northeast of Rio de Janeiro. Mean

annual temperature is about 23�C with a mean annual rainfall of about 2,560 mm (Kurtz

and de Araujo 2000). The recently enlarged national park ‘‘Parque Nacional da Serra dos

Orgaos’’ (PARNASO), founded in 1939 and now covering an area of 20,024 ha within the

municipalities of Petropolis, Guapimirim, Mage and Teresopolis, was our second research

area (22�2702400 S, 42�5904800 W). Annual rainfall is 2,821 mm in combination with high

relative humidity and a mean annual temperature of 17.8�C (Rizzini 1954; Guimaraes and

Arle 1984). The regional climate is characterized by a hot and rainy season from October

to March and a cooler and drier season from April to September (Kurtz and de Araujo

Fig. 1 Study area in south-eastern Brazil in the state of Rio de Janeiro. REGUA—NGO reserve ‘‘ReservaEcologica de Guapiacu’’, located in the municipality of Cachoeiras de Macacu. PARNASO—national park‘‘Parque Nacional da Serra dos Orgaos’’, research site located in the municipality of Teresopolis

New Forests (2012) 43:287–301 289

123

Page 4: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

2000). The natural vegetation can be classified as dense ombrophilous forest (Veloso et al.

1991) and is typical for the lower and medium elevations 50–1,500 m asl. of the coastal

mountain range of the Rio de Janeiro state (Morellato and Haddad 2000; Oliveira-Filho

and Fontes 2000). Within the Mata Atlantica Biome, the dense ombrophilous forest hosts

more than half (60%) of the plant species richness and the vast majority (80%) of endemic

plants (Stehmann et al. 2009). Characteristic canopy tree species of the lower elevations of

this forest (up to 500 m asl.) are Vochysia tucanorum (Vochysiaceae), Miconia thaezans,Tibouchina granulosa (Melastomataceae), Xylopia brasiliensis (Annonaceae) and Pla-thymenia foliosa (Fabaceae), while montane elevations (up to 1,500 m asl.) are charac-

terized by Vochysia laurifolia (Vochysiaceae), Talauma ovata (Magnoliaceae), Carinianaestrellensis (Lecythidaceae), Cedrela angustifolia (Meliaceae) and Clethra brasiliensis(Clethraceae). Myrtaceae, Rubiaceae and Melastomataceae are the most abundant and

typical families both in the sub-canopy and the understory layer (Rambaldi et al. 2003). In

the study area there are several stages of succession of forest and non-forest vegetation

representing degradation after disturbance. The very early stages of forest succession

following disturbance can be dominated by only a few woody species such as Miconiaalbicans (Melastomataceae), Attalea humilis (Arecaceae) and Gochnatia polymorpha(Asteraceae) as reported by Lima et al. (2006). The comparison of sites within these two

areas may be disadvantageous because they are representing an elevation difference of

around 900 m. We had to accept this because of the fact that the occurrence of mature

forest with low or no anthropogenic influence is mainly restricted to conservation units in

higher elevations of the Atlantic Forest of this region.

Methods

Plot design, census and forest structure

At the NGO reserve area our research plots were established in two different forest sites: in

continuous forest where selective logging took place until the late 1970s at 500 m asl. and

in forests re-growing after clear-cut about 60 years ago at 235 m asl. (Locke, N. pers.

comm.). A 100 9 100 m (1 ha) research plot with a 10 9 10 m grid of marked points was

set up at each study site. Five 20 9 20 m subplots were randomly chosen within those

plots with the constraint that subplots do not overlap. Data from the preserved forest areas

in the national park was acquired in five 20 9 20 m plots distributed at elevations between

1,165 and 1,300 m asl. In each 20 9 20 m plot at every investigation site the woody plant

community was censused according to Condit (1998). All woody plants with a diameter at

breast height (dbh) C 10 cm were tagged (including multiple stemmed individuals

exceeding 10 cm in sum). Diameter in breast height was measured by using a calliper or a

measuring tape when the diameter exceeded 50 cm. The stem-height of each individual

was estimated using four height classes:

1. \5 m

2. 5–10 m

3. 10–20 m

4. [20 m.

Canopy openness and LAI were measured with hemispherical photographs, taken in

series at the end of the dry and wet season respectively (2004/05 PARNASO, 2007/08

REGUA) to capture seasonal effects and inter-annual variability. This variability is a

290 New Forests (2012) 43:287–301

123

Page 5: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

typical pattern for forests of the region and needs to be taken into account as it develops

when forests are disturbed and/or close to edges (Sattler et al. 2007). However, the vari-

ability does not seriously affect biomass calculations due to the comparably low contri-

bution (\2%) of seasonally varying leaf biomass to total AGB (Chave et al. 2008).

For hemispherical photography we used a NikonTM Coolpix 4500 digital camera with a

NikonTM FC-E8 fisheye converter mounted on a leveled tripod at 1.30 m above ground.

Photographs were only taken under overcast conditions to avoid overexposure which could

subsequently lead to inaccurate LAI and canopy openness calculations. In the NGO reserve

area photos were taken at every grid-point within the 1 ha plots (100 photos per plot per

series), whereas at the preserved forest sites 16 photos were taken per series in each of the

20 9 20 m plots.

Data analysis

We used an allometric equation (1) given in Chave et al. (2001) for moist tropical forests

and already applied for the Atlantic forest (Rolim et al. 2005), where AGB for each tree

was calculated as: (1) ln(AGB) = -2.19 ? 2.54 (ln dbh), valid for a dbh range from 10 to

150 cm. As palm trees are especially abundant in the National Park plots, AGB of palms

was calculated separately using an allometric model suggested by Hughes et al. (1999) and

modified by Vieira et al. (2008): (2) exp((5.7236 ? 0.9285 (ln dbh2))*1.05001)/103.

Following Alamgir and Al-Amin (2008) models using basal area (dbh respectively) alone

were found to be the best predictor of biomass stock in trees because of high coefficient of

determination. Aboveground tree biomass is measured in kg and dbh in cm. As the

equation by Chave et al. (2001) is based on only one independent variable (dbh) we also

applied the standard equation by the Food and Agriculture Organisation of the United

Nations (3) as described in Garzuglia and Saket (2003), which additionally includes wood

density and a biomass expansion factor. This methodological comparison was employed to

reconfirm our AGB calculation with a different and commonly used calculation approach.

Aboveground biomass for each tree was then calculated as:

AGB ¼ BV � BEF; with BV ¼ VOB10 �WD and BEF ¼ Expf3:213� 0:506 � lnðBVÞgð1Þ

where BV = dry woody biomass of inventoried volume, VOB10 = stem volume of living

trees dbh C 10 cm, WD = wood density and BEF = biomass expansion factor.

The total basal area (calculation based on individual dbh) per 20 9 20 m plot was

scaled up to get the basal area per hectare per forest site. The stem volume (VOB10) was

calculated using five-level stem height classes. Since the specific wood density data of

many of the censused tree species could not be acquired or showed only small variation

(Chave et al. 2009), we used a constant WD = 0.603 g cm3 as suggested by Vieira et al.

(2008) for the Atlantic forest of South East Brazil. A One-Way-ANOVA with Student–

Newman–Keuls multiple comparison procedure was used to analyze differences in AGB

between the differently disturbed sites on the one hand, and differences between the results

of both AGB estimation methods for each site on the other.

The WinScanopy 2005ab software (Regent Instruments Inc. 2005) was used to analyze

the hemispherical photographs and to derive data about canopy openness (CO) and leaf

area index (LAI). Canopy openness is defined as the proportion of open sky area in a 180�hemisphere monitored from a centre point. Pixels of digital images are to be classified as

either ‘‘canopy’’ or ‘‘sky’’ based on the grayscale threshold value. Leaf are index is the

New Forests (2012) 43:287–301 291

123

Page 6: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

total one-sided area of leaf tissue per ground unit. The LI-COR LAI2000 modified algo-

rithm (assuming random leaf distribution, transmittance is equivalent to gap fraction,

modelled by the Poisson model) was used to calculate LAI values (Welles and Norman

1991).

A One-Way ANOVA on ranks was used to check for significant differences in CO and

LAI between the different forest types. Due to the different sample size of hemispherical

photographs a Dunn’s all multiple comparison procedure was performed to differentiate

between the groups. Additionally, we compared the seasonal difference at each site in

those canopy structure parameters with a repeated-measure ANOVA on ranks. Statistical

analyses were performed using the SigmaStat software package (version 3.0.1, SPSS Inc.,

2003).

Remarks on methodology

As there are many possible environmental an biological factors influencing the develop-

ment of a forests AGB, the realistic prediction of the AGB development of different forest

sites over time is highly uncertain (Asner et al. 2009). A way out of this unsatisfactory

situation would be an experimental plot design which is very complicated due to the

required time scale and the inevitable destruction of forest. Nonetheless, our small scale

non-experimental approach provides valuable results on a field of research in which data is

still scarce. Among others (McWilliam et al. 1993; DeWalt and Chave 2004; Vilela et al.

2000) our approach confirms the feasibility of small scale studies to assess questions of

forest structure and biomass in tropical forests.

Using the standard equations by the FAO (Garzuglia and Saket, 2003) for reconfirming

our AGB calculation with a different approach we applied a constant wood density of

0.603 g cm3. Needless to say that wood gravity differs between species and therefore

between sites with a different species composition, but our approach has been validated by

Wiemann and Williamson (2002), Chave et al. (2006) and Vieira et al. (2008).

Results

Biomass and stand structure

Stem density and tree height was significantly higher in the preserved forest plots, whereas

both parameters did not differ between the sites with different disturbance history in the

NGO reserve area. We noted a non-significant trend of higher basal area of the forest stand

in the preserved forest plots.

The composition of the forest regarding dhb-size classes of trees revealed a different

pattern (Fig. 2), especially in the fraction up to a dbh of 30 cm. There were significantly

less small trees (dbh 10–20 cm) in the area which was completely cleared 60 years ago,

whereas trees with a dbh of 20–30 cm were significantly less represented in the preserved

forest plots. We found no differences in the number of trees per plot with a dbh C 30 cm

between investigation areas. When looking at the distribution of the total basal area within

different dbh-classes (Fig. 3) the same pattern was identified. In general big trees

(dbh C 30 cm) do contribute more to total basal area in the preserved forest plots, even

though no significances were found due to a high variability (Lindner 2010). Nevertheless

we identified a distinct coherence between the basal area of the largest tree and the total

basal area per plot (Fig. 4).

292 New Forests (2012) 43:287–301

123

Page 7: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

Biomass calculations using two different approaches resulted in a comparable trend

pattern. We identified the highest AGB-values in the preserved forest area within the

National Park (313 ± 48 Mg ha-1). Intermediate results for AGB were found at the for-

mer complete cut down site (297 ± 83 Mg ha-1) and lowest AGB was calculated for the

area with past selective logging (204 ± 38 Mg ha-1). The results indicate that AGB

values for the complete cut down site in the NGO reserve area and those in the preserved

forest area are not significantly different (q = 0.72, P = 0.62). In contrast, the low AGB

Fig. 2 Diameter distribution of all individuals dbh C 10 cm sampled in five plots per forest type (completecut down: n = 109, selective logging: n = 133, conservation area: n = 207)

Fig. 3 Diameter distribution of total basal area (ba) of all individuals C 10 cm dbh sampled in five plots of20 9 20 m per forest type (complete cut down: ba = 6.10 m2, selective logging: ba = 4.72 m2,conservation area: ba = 7.60 m2)

New Forests (2012) 43:287–301 293

123

Page 8: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

value for the forest which was selectively logged is significantly different to the other sites

(clear cut q = 3.42, P = 0.03; preserved forest q = 4.14, P = 0.03). The crosscheck

performed by applying the FAO calculation method revealed lower AGB for both the

selectively logged and the complete cut down site but higher AGB in the preserved forest

(Table 1). As the FAO calculation does not treat the palm trees separately it leads to an

overestimation of AGB in the preserved forest due to high palm abundance at these sites.

Canopy structure and dynamics

We detected several differences in canopy structure parameters depending on location and

climate seasonality (Table 1). The lowest LAI of all sites was measured in the former clear

cut plots, whereas a higher LAI was recorded in plots in the selectively logged area and the

preserved forest plots of the national park. This interval becomes even more obvious and

significant during the dry season. Seasonal changes in LAI were significant only for plots

in the preserved forest and the selectively logged area. Despite higher LAI measures we

found higher canopy openness in the preserved forest plots too. Canopy openness was

subject to strong seasonal fluctuations at both sites in the NGO reserve area REGUA,

which is in contrast to the preserved forest plots where no significant seasonal changes in

this parameter were found.

Vegetation composition at REGUA NGO reserve area

Although no standard floristic survey was conducted, we recognized different character-

istics in terms of tree species composition in the investigated sites. Especially two species

showed particularly noticeable patterns of dominant occurrence. The Jussara-palm tree

Euterpe edulis (Arecaceae) represented 38% of all individuals and 10.5% of total basal

area in the forest plots of the national park and was the dominating species with a

dbh B 20 cm (mean dbh 11.6 ± 1.1 cm). This palm tree is endemic to the Atlantic Forest,

where it provides an economically important resource for palm heart harvesting (Galetti

and Fernandez 1998). Harvesting the apical meristemic cone (palm heart) means killing the

Fig. 4 The relationship between basal area of the largest tree and total basal area in the complete data set

294 New Forests (2012) 43:287–301

123

Page 9: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

individual. Thus, it is no surprise that E. edulis was completely absent in the disturbed

NGO reserve plots. The other conspicuous tree species was Piptadenia gonoacantha(Mimosaceae), dominating the tree community with a dbh C 20 cm in the plots in NGO

reserve area which were completely cleared 60 years ago (mean dbh 43.4 ± 11.6 cm).

This pioneer tree species is typical for early and mid successional stages of the Atlantic

Forest (Borem and Oliveira-Filho 2002; Thier and Wesenberg 2009) and represented

18.5% of all individuals and 44% of total basal area in the NGO plot with past clear-cut

and could only be recorded here. Surprisingly, no individuals of P. gonoacantha with a

dbh B 20 cm or, in other words, no actual regeneration of this species were found.

Discussion

Aboveground tree biomass, forest structure and disturbance history

Our results in biomass calculations based on a small scale approach are well in range of

published data for other moist neotropical forests (Chave et al. 2001; DeWalt and Chave

2004; Urquiza-Haas et al. 2007; Groeneveld et al. 2009). Compared to AGB estimates for

Table 1 Stand structure, aboveground tree biomass (AGB) and canopy structure parameters of all studysites

REGUA PARNASO P

Complete cutdown Selective logging Preserved forest

Stand structurea

AGB (Mg ha-1)b 297 (±83) a 204 (±38) b 313 (±48) a 0.031

AGB (Mg ha-1)c 207 (±40) a 155 (±18) b 336 (±27) c \0.001

P 0.06 0.031 0.42

Stem density per ha 545 (±33) a 665 (±128) a 1,010 (±219) b \0.001

Basal area (m2 ha-1) 30.5 (±6.9) 23.6 (±2.5) 38.0 (±13.1) 0.07

Tree hight (classes) 2.89 (±0.15) a 2.81 (±0.14) a 3.00 (±0.12) b \0.001

Canopy structure

LAI

Wet season 3.49 (±0.34) 3.66 (±0.58) 3.64 (±0.47) 0.078

Dry season 3.55 (±0.29) a 3.93 (±0.54) b 4.08 (±0.70) b \0.001

P 0.11 \0.001 \0.001

CO (%)

Wet season 6.45 (±0.58) a 6.36 (±0.64) a 6.91 (±1.06) b 0.012

Dry season 5.82 (±0.39) a 5.24 (±0.53) b 6.57 (±1.00) c \0.001

P \0.001 \0.001 0.12

Letters (a, b, c) indicate the horizontal grouping within the table resulting of a One-Way ANOVA withStudent–Newman–Keuls multiple comparison procedure for the AGB estimates and a One-Way ANOVA onranks with Dunn’s all multiple comparison procedure for the canopy structure parameters

Tree height classes: (1) \5 m; (2) 5–10 m; (3) 10–20 m; (4) [20 ma Interpolated from five 400 m2 plots per forest typeb Calculated using allometric equation (Chave et al. 2001)c Calculated using FAO-standards (Garzuglia and Saket 2003)

New Forests (2012) 43:287–301 295

123

Page 10: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

other disturbed and undisturbed forest sites in Brazils Atlantic Forest, our results coincide

as well with data for the respective forest condition, even though data published for mature

and/or undisturbed forests are from lower elevations (Table 2). The FAO equation

crosscheck of our AGB calculation revealed once more the influence of the methodological

approach in estimating AGB of a given forest site (e.g., Brown et al. 1989 and Chave et al.

2008) as the values obtained with this calculation (Table 1) are up to 30% lower than the

ones calculated with the allometric equation of Chave et al. (2001). Nevertheless, the

relative trend of AGB within the differently disturbed forest sites remains the same. The

selectively logged forest contains the lowest AGB within the investigated areas in our

study and this result is independent from the applied biomass calculation method. Here the

selective removal of trees with a diameter appropriate for timber production indicates an

underrepresentation of trees with dbh C 50 cm (Fig. 3), although this demographic dif-

ference was not significant it is still reflected in the outcome of the calculated AGB.

Significantly more AGB was estimated for the forest regrowth after complete clearing

60 years ago, which is mainly due to the contribution of a single mid-successional species:

P. gonoacantha. Within 60 years after complete clear cut the dominance of this species

within a re-growing forest led to a higher AGB than in a selectively logged forest. Overall

tree diversity is of course much lower at the clear-cut site. Furthermore, the fate of this

forest and further succession is rather unpredictable, especially after the disappearance of

P. gonoacantha which is likely to happen because neither seedlings nor young individuals

of this species were found within the plots.

Big trees (dbh C 30 cm) contribute more into total basal area and hence in calculated

AGB, especially for the protected forest area. For the National Park ‘‘Parque Nacional da

Serra dos Orgaos’’ a contribution of 78% to total AGB is reported for such big trees

(Lindner 2010), which confirms their importance in terms of carbon storage capacity.

We also recognized a ‘‘gap’’ in dbh size class distribution between the different areas:

there were obviously less individuals with a dbh 20–30 cm in the national park area. For

Table 2 Atlantic forest AGB estimates from literature in comparison to data of this study

Reference Forest type Forestcondition

Elevationa.s.l. (m)

DBHrange(cm)

AGBestimation

AGB(Mg ha-1)

Alveset al.(2010)

Lowland to lowermontane Atlanticrain forest

Latesecondary/mature

100–1,000 5–156 Allometricmodel

243.7(±33.2)

Cunhaet al.(2009)

Atlantic moist forest Secondary(40 years)

600–900 10–70 Allometricmodel

166.8

Rolimet al.(2005)

Lowland, semi-deciduous Atlanticmoist forest

Relativelyundisturbed

28–65 C 10 Allometricmodel

334.5(±11.3)

Burger(2005)

Atlantic moist forest Secondary(±30 years)

570 1.6–47.8 Destructivesampling

245

Thisstudy

Montane, denseombrophilous forest

Latesecondary/mature

1,165–1,300 10–70 Allometricmodel

313.5(±48.2)

Thisstudy

Submontane, denseombrophilous forest

Secondary 235–500 10–70 Allometricmodel

250.4(±77.9)*

* Mean value and STD from merged data of both selective logging and complete cutdown sites

296 New Forests (2012) 43:287–301

123

Page 11: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

our study, this difference might be due to an earlier successional stage indicating that this

size class provides a major contribution to total biomass within areas exposed to distur-

bance. The higher stem density in the protected forest plot may be due to the higher

elevation (Moser et al. 2007). No significant difference in stem density was found in the

disturbed NGO reserve plots, independent from disturbance history. This confirms the

results by Vilela et al. (2000) which investigated the effect of selective logging in a

seasonally dry Brazilian Atlantic Forest. The most trees with a dbh B 20 cm have been

recorded within the plots of the protected forest area and were mainly represented by E.edulis. The frequent occurrence of the Jussara-palm tree made the most noticeable dif-

ference when compared the NGO plots, where it was completely missing. In general high

density stands seem to be important in terms of carbon sequestration (Redondo-Brenes

2007).

A factor indirectly influencing forest biomass calculations is the abundance of lianas.

They play a key role for forest dynamics and processes. Lianas can reduce the host tree

growth rate and suppress gap phase regeneration (Schnitzer and Bongers 2002). Moreover it

is known that lianas increase the mortality of host trees (Ingwell et al. 2010). Although

lianas represent a small fraction of AGB that has not been taken into account in our study,

their negative effects on tree growth can contribute to the low stand biomass in the selective

logging plot where lianas tend to be more abundant than in the other plots (Faske 2009).

General implications

As incipiently mentioned there is a general uncertainty about biomass calculation and

therefore for carbon stock estimations in tropical forests (Houghton et al. 2001; Fearnside

et al. 2009). Nevertheless, information and data availability differs regionally with some

regions (e.g., Amazonian Forests) being investigated more intensely. Especially in the

Atlantic Forest studies are scarce. Our study provides data that help to fill the data-gap for

the Atlantic Forest and contributes knowledge for a largely under-sampled area, especially

with regard to forests representing different disturbance histories.

While the estimation of biomass directly from remote sensing data still has been proved

illusive as a tool for large-scale estimates (Fearnside et al. 2009), reference points on the

ground are essential to improve the results of meta-studies (Saatchi et al. 2007) and to

supply data for ecological models like FORMIND (Kohler and Huth 1998; Groeneveld

et al. 2009). Most of the remaining area of the Atlantic Forest is highly fragmented

(Ribeiro et al. 2009) and the influence of fragmentation and its effect on biomass condi-

tions is an important aspect to study (Saunders et al. 1991; Groeneveld et al. 2009).

However, disturbance history of forest patches and continuous forest is often not known

and generalization of ecosystem responses may lead to indistinct outcomes for upcoming

investigations. Midgley and Niklas (2004) drew attention on the fact that different dis-

turbance regimes determine the max size that trees can achieve and therefore strongly

influence the total AGB of a forest. Therefore, small scale structural approaches as tested in

our study are able to form an initiating framework of more detailed results and help to

improve estimates on biomass and carbon storage.

Conclusions

Our small scale, non-experimental approach using existing information on past disturbance

and actual forest stand characteristics was applicable to reveal significant differences in

New Forests (2012) 43:287–301 297

123

Page 12: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

AGB formation within the studied forest sites. Furthermore, the results of this study lead to

the conclusion that biomass productivity can recover in a forest which was completely

cleared 60 years ago, even if the details of past and further successional development

remain rather unexplained. On the other hand the outcome of our study shows that

‘‘gentle’’ influences like selective logging can have, even if stopped several decades ago. In

other studies selective logging is described to have little effect on forest structure and

dynamics (Deccker and de Graaf 2003), but our results suggest otherwise. These findings

can contribute to the discussion on the value of different forest management strategies

within the context of the reductions in emissions from deforestation in developing coun-

tries (REDD) debate (Gibbs et al. 2007; Shevliakova et al. 2009; Reyer et al. 2009).

Acknowledgments We are grateful to Jens Wesenberg for kindly providing the floristic and foreststructure data for the national park area. We are thankful to Christian Wirth and anonymous reviewers forvaluable comments on earlier versions of the manuscript. Furthermore we would like to thank the nationalpark ‘‘Serra dos Orgaos’’, the Brazilian Institute of Environment and Renewable Natural Resources(IBAMA) and the National Counsel of Technological and Scientific Development (CNPq) for issuingresearch permission. We are grateful to the ‘‘Reserva Ecologica de Guapiacu’’ (REGUA) and the wholestaff, especially to Mr. Nicholas Locke for logistic support and the permission to work on the property. Forfunding we are thankful to the German Federal Ministry of Education and Research (BMBF). This study wasconducted within the framework of the Brazilian-German joint research project ‘‘Climate change, landscapedynamics, land use and natural resources in the Atlantic Forest of Rio de Janeiro’’ (FKZ 01LB0801B).

References

Alamgir M, Al-Amin M (2008) Allometric models to estimate biomass organic carbon stock in forestvegetation. J For Res 19:101–106

Alves LF, Vieira SA, Scaranello MA, Camargo PB, Santos FAM, Joly CA, Martinelli LA (2010) Foreststructure and live aboveground biomass variation along an elevational gradient of tropical Atlanticmoist forest (Brazil). For Ecol Manag 260:679–691

Asner GP, Hughes RF, Varga TA, Knapp DE, Kennedy-Bowdoin T (2009) Environmental and bioticcontrols over aboveground biomass throughout a tropical rain forest. Ecosystems 12:261–278

Borem RAT, Oliveira-Filho AT (2002) Influencia do solo e topografia sobre as variacoes da composicaoflorıstica e estrutura da comunidade arboreo-arbustiva de uma floresta estacional semidecidual emIngaı, MG. Revista Brasileira de Botanica 25:195–213

Brown S, Gillespie AJR, Lugo AE (1989) Biomass estimation methods for tropical forests with applicationsto forest inventory data. For Sci 35:881–902

Burger D (2005) Modelos alometricos para a estimativa da fitomassa de Mata Atlantica na Serra do Mar, SP.PhD thesis, Universidade de Sao Paulo, Sao Paulo

Burger DM, Delitti WBC (2008) Allometric models for estimating the phytomass of a secondary AtlanticForest area of southeastern Brazil. Biota Neotropica 8:131–136

Chambers JQ, dos Santos J, Ribeiro RJ, Higuchi N (2001) Tree damage, allometric relationships, and above-ground net primary production in central Amazon forest. For Ecol Manag 152:73–84

Chave J, Riera B, Dubois MA (2001) Estimation of biomass in a neotropical forest of French Guiana: spatialand temporal variability. J Trop Ecol 17:79–96

Chave J, Muller-Landau HC, Baker TR, Easdale TA, ter Steege H, Webb CO (2006) Regional and phy-logenetic variation of wood density across 2456 Neotropical species. Ecol Appl 16:2356–2367

Chave J, Olivier J, Bongers F, Chatelet P, Forget PM, van der Meer P, Norden N, Riera B, Charles-Dominique P (2008) Above-ground biomass and productivity in a rain forest of eastern South America.J Trop Ecol 24:355–366

Chave J, Coomes DA, Jansen S, Lewis SL, Swenson NG, Zanne AE (2009) Data from: towards a worldwidewood economics spectrum. Dryad Digit Repos. doi:10.5061/dryad.234

Condit R (1998) Tropical forest census plots. Springer, BerlinCunha GD, Gama-Rodrigues AC, Gama-Rodrigues EF, Velloso ACX (2009) Biomassa e estoque de

carbono e nutrientes em florestas montanas da mata atlantica na regiao norte do estado do Rio deJaneiro. Revista Brasileira De Ciencia Do Solo 33(5):1175–1185

298 New Forests (2012) 43:287–301

123

Page 13: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

Deccker M, de Graaf NR (2003) Pioneer and climax tree regeneration following selective logging withsilviculture in Suriname. For Ecol Manag 172:183–190

DeWalt SJ, Chave J (2004) Structure and biomass of four lowland neotropical forests. Biotropica 36:7–19Faske M (2009) Untersuchungen zur Vegetationsstruktur der Lianen in unterschiedlichen Sukzessionssta-

dien der fragmentierten Mata Atlantica. Dipl. Thesis, University of LeipzigFearnside PM, Righi CA, Graca PMLA, Keizer EWH, Cerri CC, Nogueira EM, Barbosa RI (2009) Biomass

and greenhouse-gas emissions from land-use change in Brazil’s Amazonian ‘‘arc of deforestation’’: thestates of Mato Grosso and Rondonia. For Ecol Manag 258:1968–1978

Frolking S, Palace MW, Clark DB, Chambers JQ, Shugart HH, Hurtt GC (2009) Forest disturbance andrecovery: a general review in the context of spaceborne remote sensing of impacts on abovegroundbiomass and canopy structure. J Geophys Res Biogeosci 114:G00E02. doi:10.1029/2008JG000911

Galetti M, Fernandez JC (1998) Palm heart harvesting in the Brazilian Atlantic forest: changes in industrystructure and illegal trade. J Appl Ecol 35:294–301

Galindo-Leal C, de Gusmao Camara I (2003) Atlantic Forest hotspot status: an overview. In: Galindo-LealC, de Gusmao Camara (eds) The Atlantic forest of South America. Island Press, Washington, DC,pp 3–11

Garzuglia M, Saket M (2003) Wood volume and woody biomass: review of FRA 2000 estimates. ForestryResources Assessment Programme: Working Paper 68

Gibbs HK, Brown S, Niles JO, Foley JA (2007) Monitoring and estimating previous tropical forest carbonstocks: making REDD a reality. Environ Res Lett 2:1–13

Groeneveld J, Alves LF, Bernacci LC, Catharino ELM, Knogge C, Metzger JP, Putz S, Huth A (2009) Theimpact of fragmentation and density regulation on forest succession in the Atlantic rain forest. EcolModel 220:2450–2459

Guimaraes AE, Arle M (1984) Mosquitos no Parque Nacional da Serra dos Orgaos, Estado do Rio doJaneiro, Brasil. I—Distribuicao Estacional. Memorias do Instituto Oswaldo Cruz 79:309–323

Houghton RA (2005) Aboveground forest biomass and the global carbon balance. Glob Change Biol11:945–958

Houghton RA, Lawrence KT, Hackler JL, Brown S (2001) The spatial distribution of forest biomass in theBrazilian Amazon: a comparison of estimates. Glob Change Biol 7:731–746

Hughes RF, Kauffman JB, Jaramillo VJ (1999) Biomass, carbon, and nutrient dynamics of secondary forestsin a humid tropical region of Mexico. Ecology 8(6):1892–1907

Ingwell LL, Wright SJ, Becklund KK, Hubbell SP, Schnitzer SA (2010) The impact of lianas on 10 years oftree growth and mortality on Barro Colorado Island, Panama. J Ecol 98:879–887

ITTO (2002) Guidelines for the restoration management and rehabilitation of degraded and secondarytropical forests. International Tropical Timber Organization, Yokohama

Kellner JR, Asner GP (2009) Convergent structural responses of tropical forests to diverse disturbanceregimes. Ecol Lett 12:887–897

Kindel A, Barbosa PMS, Perez DV, Garay I (1999) Efeito do extrativismo seletivo de especies arboreas dafloresta Atlantica de Tabuleiros na materia organica e outros atributos do solo. Revista Brasileira daCiencia do Solo 25:551–563

Kohler P, Huth A (1998) The effects of tree species grouping in tropical rainforest modelling: simulationswith the individual-based model FORMIND. Ecol Model 109:301–321

Kurtz BC, de Araujo DSD (2000) Composicao florıstica e estrutura do componente arboreo de um trecho deMata Atlantica na Estaocao Ecologica Estadual do Paraıso, Cachoeiras de Macacu, Rio de Janeiro,Brasil. Rodriguesia 51:69–112

Laurance WF, Lovejoy TE, Vasconcelos HL, Bruna EM, Didham RK, Stouffer PC, Gascon C, BierregaardRO, Laurance SG, Sampaio E (2002) Ecosystem decay of Amazonian forest fragments: a 22-yearinvestigation. Conserv Biol 16:605–618

Lima HC, SdVA Pessoa, Guedes-Bruni RR, Moraes LFD, Granzotto SV, Iwamoto S, Di Ciero J (2006)Caracterizacao fisionomico-florıstica e mapeamento da vegetacao da Reserva Biologica de Poco dasAntas, Silva Jardim, Rio de Janeiro, Brasil. Rodriguesia 57:369–389

Lindner A (2010) Biomass storage and stand structure in a conservation unit in the Atlantic Rainforest—therole of big trees. Ecol Eng 36:1769–1773

McWilliam ALC, Roberts JM, Cabral OMR, Leitao MVBR, Decosta ACL, Maitelli GT, Zamparoni CAGP(1993) Leaf-area index and aboveground biomass of terra-firme rain-forest and adjacent clearings inAmazonia. Funct Ecol 7:310–317

Midgley JJ, Niklas KJ (2004) Does disturbance prevent total basal area and biomass in indigenous forestsfrom being at equilibrium with the local environment? J Trop Ecol 20:595–597

Morellato LPC, Haddad CFB (2000) Introduction: the Atlantic Forest. Biotropica 32:786–792

New Forests (2012) 43:287–301 299

123

Page 14: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

Moser G, Hertel D, Leuschner C (2007) Altitudinal change in LAI and stand leaf biomass in tropicalmontane forests: a transect study in Ecuador and a pan-tropical meta-analysis. Ecosystems 10:924–935

Nogueira EM, Fearnside PM, Nelson BW, Barbosa RI, Keizer EWH (2008) Estimates of forest biomass inthe Brazilian Amazon: new allometric equations and adjustments to biomass from wood-volumeinventories. For Ecol Manag 256:1853–1867

Oliveira-Filho AT, Fontes MAL (2000) Patterns of floristic differentiation among Atlantic Forests inSoutheastern Brazil and the influence of climate. Biotropica 32:793–810

Pedreira BCG, Fidalgo ECC, Abreu MB (2009) Mapeamento do uso e cobertura da terra da bacia hid-rografica do rio Guapi-Macacu, RJ. Anais XIV Simposio Brasileiro de Sensoriamento Remoto, Natal,INPE, pp 2111–2118

Rambaldi DM, Magnanini A, Ihla A, Lardosa E, Figueiredo P, Oliveira RF (2003) A Reserva da Biosfera daMata Atlantica no Estado do Rio de Janeiro. Serie Estado do Rio de Janeiro. Caderno no 22. ConselhoNacional da Reserva da Biosfera da Mata Atlantica, Sao Paulo

Redondo-Brenes A (2007) Growth, carbon sequestration, and management of native tree plantations inhumid regions of Costa Rica. New For 34:253–268

Regent Instruments Inc (2005) WinScanopy 2005ab for hemispherical image analyses. Available from:http//www.regentinstruments.com

Reyer C, Guericke M, Ibisch PL (2009) Climate change mitigation via afforestation, reforestation anddeforestation avoidance: and what about adaptation to environmental change? New For 38:15–34

Ribeiro MC, Metzger JP, Martensen AC, Ponzoni FJ, Hirota MM (2009) The Brazilian Atlantic Forest: howmuch is left, and how is the remaining forest distributed? Implications for conservation. Biol Conserv142:1141–1153

Rizzini CT (1954) Flora organensis—Lista preliminar dos cormophyta da Serra dos Orgaos. Arquivos doJardim Botanico do Rio de Janeiro 13:118–246

Rolim SG, Jesus RM, Nascimento HEM, Couto HTZ, Chambers JQ (2005) Biomass change in an Atlantictropical moist forest: the ENSO effect in permanent sample plots over a 22-year period. Oecologia142:238–246

Saatchi SS, Houghton RA, Dos Santos Alvala RC, Soares JV, Yu Y (2007) Distribution of aboveground livebiomass in the Amazon Basin. Glob Change Biol 13:816–837

Saldarriaga JG, Uhl C (1991) Recovery of forest vegetation following slash-and-burn agriculture in theupper Rio Negro of Colombia and Venezuela. In: Gomez-Pompa A, Whitmore TC, Hadley M (eds)Tropical rain forest: regeneration and management. Blackwell, New York, pp 303–312

Sattler D, Lindner A, Morawetz W (2007) A funcao da sazonalidade no levantamento estrutural da florestamontana tropical. In: Cronemberger C, Viveiros de Castro E (eds) Ciencia e Conservacao na Serra dosOrgaos. MMA-IBAMA, Brasilia, pp 105–116

Saunders DA, Hobbs RJ, Margules CR (1991) Biological consequences of ecosystem fragmentation: areview. Conserv Biol 5:18–32

Schnitzer SA, Bongers FA (2002) The ecology of lianas and their role in forests. Trends Ecol Evol17:223–230

Shevliakova E, Pacala SW, Malyshev S, Hurtt GC, Milly PCD, Caspersen JP, Sentman LT, Fisk JP, Wirth,C, Crevoisier C (2009) Carbon cycling under 300 years of land use change: Importance of the sec-ondary vegetation sink. Glob Biogeochem Cycles 23:GB2022. doi:10.1029/2007GB003176

Silva JMC, Casteleti CHM (2003) Status of the biodiversity of the Atlantic Forest of Brazil. In: Galindo-Leal C, Camara I (eds) The Atlantic Forest of South America: biodiversity status, trends, and outlook.Center for Applied Biodiversity Science and Island Press, Washington, DC, pp 43–59

SOS Mata Atlantica/INPE (2009) Atlas dos remanescentes florestais da Mata Atlantica perıodo2005–2008—Relatorio parcial. Sao Paulo

Stehmann JR, Forzza RC, Salino A, Sobral M, Costa DP, Kamio LHY (2009) Diversidade Taxonomica naFloresta Atlantica. In: Stehmann JR, Forzza RC, Salino A (eds) Plantas da Floresta Atlantica. Rio deJaneiro. Jardim Botanico de Rio de Janeiro, pp 3–37

Tabarelli M, Pinto LP, Silva MC, Hirota M, Bede L (2005) Challenges and opportunities for biodiversityconservation in the Brazilian Atlantic Forest. Conserv Biol 19:695–700

Thier O, Wesenberg J (2009) Diversity, floristic composition and similarity of three forest fragments in theMata Atlantica of Rio de Janeiro. In: Gaese H, Torrico-Albino JC, Wesenberg J, Schlueter S (eds)Biodiversity and land use systems in the fragmented Mata Atlantica of Rio de Janeiro. Cuvillier,Gottingen, pp 245–258

Tiepolo G, Calmon M, Feretti AR (2002) Measuring and monitoring carbon stocks at the Guaraquecabaclimate action project, Parana, Brazil. Ext Serie Taiwan For Res Inst 153:98–115

Uhl C, Buschbacher R (1985) A disturbing synergism between cattle ranch practices and selective treeharvesting in the eastern Amazon. Biotropica 17:265–268

300 New Forests (2012) 43:287–301

123

Page 15: Biomass Estimations in Forests of Different Disturbance History in the Atlantic Forest of Rio de Janeiro, Brazil

Urquiza-Haas T, Dolma PM, Peres CA (2007) Regional scale variation in forest structure and biomass in theYucatan Peninsula, Mexico: effects of forest disturbance. For Ecol Manag 247:80–90

Veloso HP, Rangel-Filho AL, Lima JCA (1991) Classificacao da vegetacao brasileira, adaptada a umsistema universal. IBGE/CDDI, Departamento de Documentacao e Biblioteca, Rio de Janeiro

Vieira S, de Camargo PB, Selhorst D, da Silva R, Hutyra L, Chambers JQ, Brown IF, Higuchi N, dos SantosJ, Wofsy SC, Trumbore SE, Martinelli LA (2004) Forest structure and carbon dynamics in Amazoniantropical rain forests. Oecologia 140:468–479

Vieira SA, Alves LF, Aidar M, Araujo LS, Baker T, Batista JLF, Campos MC, Camargo PB, Chave J, DelittiWBC, Higuchi N, Honorio E, Joly CA, Keller M, Martinelli LA, Mattos EAD, Metzker T, Phillips O,Santos FAMD, Shimabukuro MT, Silveira M, Trumbore SE (2008) Estimation of biomass and carbonstocks: the case of the Atlantic Forest. Biota Neotropica 8:21–29

Vilela EA, Oilveira-Filho AT, Carvalho DA, Guilhermes FAG, Appolinario V (2000) Caracterizacaoestrutural de floresta riparia do Alto Rio Grande, em Madre de Deus de Minas, MG. Cerne 6:41–54

Welles JM, Norman JM (1991) Instrument for indirect measurement of canopy architecture. Agron J83:818–825

Wiemann MC, Williamson GB (2002) Geographic variation in wood specific gravity: effects of latitude,temperature and precipitation. Wood Fiber Sci 34:96–107

Williams-Linera G, Domiguez-Gastelu V, Garcia-Zurita ME (1998) Microenvironment and floristics ofdifferent edges in a fragmented tropical rainforest. Conserv Biol 12:1091–1102

New Forests (2012) 43:287–301 301

123