437

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 70: 437-449, January-December 2022 (Published Jun. 20, 2022)

Impact of plot size on tropical forest

structure and diversity estimation

Alan Bernardes da Silveira¹; https://orcid.org/0000-0002-4884-1875

Samuel de Padua Chaves e Carvalho²*; https://orcid.org/0000-0002-5590-9049

Marcos Felipe Nicoletti³; https://orcid.org/0000-0003-4732-0119

Carlos Alberto Silva4; https://orcid.org/0000-0002-7844-3560

Ronaldo Drescher1; https://orcid.org/0000-0001-9549-6501

Mariana Peres de Lima Chaves e Carvalho1; https://orcid.org/0000-0002-9641-4579

Joao Paulo Sardo Madi5; https://orcid.org/0000-0002-9817-2657

Larissa Regina Topanotti6; https://orcid.org/0000-0001-5066-4196

Walmes Marques Zeviani7; https://orcid.org/0000-0001-5503-8565

Valdir Carlos Lima de Andrade8; https://orcid.org/0000-0002-5559-9124

1. Postgraduate program of Environmental and Forest Science, Federal University of Mato Grosso, Cuiabá, Brazil;

alan@onfbrasil.com.br, ronaldodrescher@gmail.com, marianaperes212@gmail.com

2. Postgraduate program of Forest Science, University of Brasilia, Brasilia, Brazil; sam.padua@gmail.com

(*Correspondence)

3. College of Agriculture and Veterinary, University of Santa Catarina, Lages, Brazil; marcos.nicoletti@udesc.br

4. School of Forest, Fisheries, and Geomatics Sciences, University of Florida, Gainesville, United States of America;

c.silva@ufl.edu

5. Postgraduate program of Forest Engineering, Federal University of Parana, Curitiba, Brazil;

joaosardomadi@gmail.com

6. Department of Forest Economics and Sustainable Land-use Planning, Faculty of Forest Sciences and Forest Ecology,

University of Göttingen, Göttingen, Germany; larissa.topanotti@uni-goettingen.de

7. Departament of Statistics, Federal University of Parana, Curitiba, Brazil; walmeszeviani@gmail.com

8. Departament of Forest Engineering, Federal Univerisity of Tocantins, Gurupi, Brazil; vclandradeuft@gmail.com

Received 12-X-2021. Corrected 31-III-2022. Accepted 15-VI-2022.

ABSTRACT

Introduction: Inventories are essential for forest management, but, in the Amazon region, the absence of

standardization produces information loss, low accuracy, and inconsistent measurements. This prevents valid

comparisons and compromises the use of information in networks and software. Sampling unit size is of key

importance in the inventory of native forests, particularly regarding accuracy and costs.

Objective: To identify a plot size that provides adequate precision for dendrometric parameters in the Amazon.

Methods: In Cotriguaçu, Mato Grosso, Brazil, we tested four plot sizes with six repetitions each: 2 500, 5 000,

7 500, and 10 000 m². We measured diameter at breast height, population density, basal area, and biomass. We

applied Shannon and Jaccard indexes; Weibull 2P and Gamma functions to fit the diametric distribution; and the

Akaike Information Criterion for the best model.

Results: There was a directly proportional relationship between plot area and population similarity, but diversity

did not indicate significant alterations. Plot size did not affect dendrometric attributes and diametric distribution.

Larger plot areas led to lower coefficients of variation and smaller confidence intervals. The Gamma function

was the best model to represent the distributions of different plot sizes.

https://doi.org/10.15517/rev.biol.trop.2022.48640

TERRESTRIAL ECOLOGY

438 Revista de Biología Tropical, ISSN: 2215-2075 Vol. 70: 437-439, January-December 2022 (Published Jun. 20, 2022)

In the management of native forests, forest

inventory is an essential procedure for forest

planning. According to de Araujo (2006), in

areas where forest management is required,

forest inventory is a key to understanding

the composition and structure of the forest,

through the gathering of information from

natural regeneration to the adult stage; it allows

to determine the potential and aptitude for the

management of the area.

Forest inventory is responsible for quanti-

fying available resources, assessing their qual-

ity, and providing information that supports

the decision-making process to implement a

management system (Farias, 2012). According

to the same author, this activity also aims to

monitor managed species to obtain information

about possible impacts on the regeneration of

such species and the remaining fauna and flora.

Forest inventories play several roles in the

management of native forests. They provide

information about the stock of non-timber

species (Farias, 2012), forest carbon (Higuchi

et al., 2004; Vianna et al., 2010), and forest

structure (Cavalcanti et al., 2009; Cavalcanti

et al., 2011; Ubialli et al., 2009), among other

purposes. In the Amazon forest, according to

de Oliveira et al. (2014), inventories are mainly

used for volume estimation, as a tool to evalu-

ate the technical and economic feasibility of

forest management plans in private properties.

Inventories performed in the Amazon

rainforest adopt several methodologies (Silva,

1980; Silva & Lopes, 1984; Silva et al., 2005),

besides the network of permanent plots as

Rainfor - Amazon Forest Inventory Network

and TmFO - Tropical Managed Forests Obser-

vatory, especially with the emphasis on the size

and shape of the plot and also in the definition

of the minimum inclusion diameter, which ends

up being a problem. This variation in the size

and shape of the plots as well as the sampling

intensity can lead to uncertainties in forest

inventory (de Oliveira et al., 2014).

In this sense, absence of standardization

in forest inventory can generate certain con-

sequences. They refer to information loss,

low accuracy, and inconsistent measurement

standards which compromise the use of the

collected information in national or interna-

tional networks and processing software. Thus,

standardization of the methods would allow the

researcher to benefit from common expertise,

enhance research activities, favor the valoriza-

tion and dissemination of results, and optimize

the available resources.

Thus, when considering the hypothesis

that plot with 10 000 m² of área are the bet-

ter to realize the forest inventory diagnostics

in amazonia Pantropical, we suggested other

three variations in this size (2 500, 5 000 and

7 500 m²) compared with the 10 000 m² size,

to provide adequate precision for the estimation

of dendrometric parameters and diversity in an

area of Mixed Submontane Forest, in the legal

Brazilian Amazon region, Mato Grosso State.

MATERIALS AND METHODS

Study area: This study was conducted at

the São Nicolau farm, owned by the company

ONF Brazil. The farm is located in the legal

Amazon region (ME, 2019), Cotriguaçu city,

northwest of the Mato Grosso State, around

the coordinates 58.24 S and 9.84 W (Fig.

1). The farm has a total area of 10 287 ha,

which 7 500 ha is of native forest. Of these

7 500 ha of native forest, 1 800 ha have been

used as a Private Reserve of Natural Heritage

(RPPN Peugeot ONF), while the remnant of

the native forest area was used for sustainable

timber production.

Conclusions: For similar forests, we recommend the 2 500 m² plot to evaluate diameter at breast height, popula-

tion density, basal area, and biomass.

Key words: sustainable forest management; sampling; floristic analysis; structural analysis; tropical ecology.

439

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 70: 437-449, January-December 2022 (Published Jun. 20, 2022)

The property is located in a region with

an “A” type climate, according to the Köppen

method (Alvares et al., 2013). The region has

average annual temperatures between 25 °C

and 27 °C, an average annual rainfall of 2 500

mm, and relative humidity of 85 %.

The original forest type of the area under

study, according to MME (1982), compris-

es the Tropical Open Rainforest typology,

sub-montane formation with Palm trees. Its

understory is quite dense, mainly due to the

regeneration of palm trees. The region of the

Open Ombrophyla Rainforest is called the

“transition zone” between the Amazon and the

rest of the country.

Experimental design: The study was con-

ducted in the Annual Production Unit 1 (UPA

1) of the Forest Management Plan of Fazenda

São Nicolau, with 227 ha of area. Using QGIS

(QGIS Development Team, 2018), a grid of

size 100 m x 100 m was allocated throughout

the study area. After that, the first plot was

randomly chosen and the subsequent five were

systematically defined, with a six sample plot

allocated in this study area (Fig. 2).

After the selection and installation of plots,

they were divided into four subplots of 2 500

m² each, which allowed the definition of a spa-

tial gradient of 2 500 m², 5 000 m², 7 500 m²,

and 10 000 m² across the dimensions of 25 m x

100 m, 50 m x 100 m, 75 m x 100 m and 100 m

x 100 m, respectively (Fig. 3). The design

resulting from the divisions of the allocated

plots had four plot sizes, each with six repeti-

tions, in which the variables of interest were

analyzed. The number of samples in the study

follow the requirements established by Decree

2152/14 of the Secretary of State for the Envi-

ronment for the State of Mato Grosso, Brazil.

Data collection: Trees with diameters at

breast height (DBH) ≥ 20 cm (Silva & Lopes,

1984) were measured and identified within

the plots. From the trees not identified in loco,

vegetative material was collected, pressed, and

Fig. 1. Location of the study area (orange polygon) on the Mato Grosso State, Brazil.

440 Revista de Biología Tropical, ISSN: 2215-2075 Vol. 70: 437-439, January-December 2022 (Published Jun. 20, 2022)

Fig. 2. Study area in details with the location of the six randomized plots (in brown).

Fig. 3. A didactic scheme exemplifying how the plots were allocated in the field.

441

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 70: 437-449, January-December 2022 (Published Jun. 20, 2022)

dried in an oven with forced ventilation at 60

°C for 48 hours, according to the herboriza-

tion methodology indicated by the Technical

Manual of Brazilian Vegetation (ME, 2012).

The material for the identification of these

species was then sent to the Herbarium IAN -

EMBRAPA, in Belém, Para State, Brazil.

Metrics calculated: Initially, the floristic

similarity between the different plot sizes was

calculated, using the Jaccard Index (Equation

1), according to Magurran (1988).

(1)

where: SJ = Jaccard similarity index; a = total

number of species present in sample “a”; b =

total number of species present in sample “b”;

c = total number of species common to samples

“a” and “b”.

The diversity was obtained as proposed by

Magurran (1988), by Shannon-Weaver Index

(Equation 2).

(2)

in which: H’ = Shannon-Weaver diversity

index; ni = number of individuals of the i-th

species in the sample; N = total number of indi-

viduals in the sample; ln = Napierian logarithm

(base e).

The Shannon’s indexes were compared

by Hutcheson’s Test with 5 % of signifi-

cance (Hugo & David, 2021; Hutcheson, 1970;

Magurran, 1988).

The horizontal structure of the vegetation

was evaluated through tree density (N.ha-1)

and basal area (m².ha-1) calculated at plot level

(Equation 3).

(3)

where: AB = Basal Area (m²/ha); DBH =

Diameter at Breast Height (cm); EF = Expan-

sion Factor (EF = 4 for 2 500 m² plot size; 2

for 5 000 m² plot size; and 1.33 for 7 500 m²

plot size).

Above Ground Biomass (AGB) was also

estimated at the individual tree-level using the

equation proposed by Chave et al. (2005) for

humid forests (Equation 4). After obtaining

this variable at the tree-level, the individual

biomass of the trees per plot was summed and

expanded to 1 hectare.

where: AGB = Above Ground Biomass (kg/

tree); DBH = Diameter at Breast Height (cm);

ln = Napierian logarithm (base e); r = wood

basic density.

The basic wood density values for each

species to apply the Chave et al. (2005) equa-

tion were obtained from the Global Wood

Density Database (Dryed) (Zanne, 2009). The

confidence intervals were obtained for biomass

(Mg.ha-1) and stocking density (n.ha-1).

The coefficient of variation was calculated

to verify the dispersion of the data (Equation 5).

(5)

where: CV = coefficient of variation (%); sd =

standard deviation, = mean.

(4)

Subsequently, in all plot size conditions,

the Weibull probability density function of two

parameters and Gamma (Equation 6, Equation

7) were fitted. The AIC – Akaike Information

Criteria was used to evaluate the best model

to represent the data. (Equation 8). These

functions were chosen because of the perfor-

mance to describe curves with different shapes

and flexibility.

(6)

where: f(x) = density function of variable x; x

= class center diameter; b = scale parameter; c

= shape parameter.

442 Revista de Biología Tropical, ISSN: 2215-2075 Vol. 70: 437-439, January-December 2022 (Published Jun. 20, 2022)

(7)

where: f(x) = density function of variable x; x

= class center diameter; b = scale parameter; a

= shape parameter; .

AIC = -2ln(mv)+2p (8)

where: ln = logaritimic of the value; mv =

maximum likehood value; p = number of the

parameters of the model evaluated.

Finally, dendrometric parameters were

compared by the Dunnett’s Test assumpting the

10 000 m² plot size with a control treatment.

Further details can be obtained in Dunnett

(1955) and Signorell et al. (2021).

All analyses were performed in R Soft-

ware (R Core Team, 2019).

RESULTS

Considering all the plots, a total of 877

individuals with DBH ≥ 20 cm were recorded

in the forest inventory, with an average of 146

individuals per hectare. These individuals were

distributed in 36 families and 119 species.

The richest species-diverse families were

the following: Fabaceae with 34 species, Mora-

ceae with 10, Malvaceae with 8, followed

by Lecythidaceae, Meliaceae, and Sapotaceae,

with 5 species each. Regarding the number of

individuals per family, the most representative

were as follows: Burseraceae (254), Moraceae

(137), Fabaceae (108), and Sapotaceae (85),

which represented 66.5 % of the total sampled

individuals. More floristic informations can be

obtained in Silveira (2019).

The different plot sizes generated values of

Shannon-Weaver index (H’) (Table 1), and they

presented certain proximity among them. For

the similarity, the Jaccard Index (Sj) (Table 1),

it was observed that the largest plots presented

higher index values than the smaller ones,

when compared to the 10 000 m² plot (refer-

ence control plot size in this study).

Table 2 shows the results obtained con-

cerning mean diameter, mean population

density (number of trees per hectare), basal

area, and biomass, for the different plot sizes,

extrapolated to one hectare. The main struc-

tural parameter that differed in the plots was

biomass, where the smallest plot size showed

a higher value compared with the other plot

sizes, however without statistical difference as

showed in Table 2.

The statistics obtained for the density and

biomass variables for the different plot sizes are

shown in Table 3. The plot sizes that showed

the lowest standard deviation and coefficient

of variation were 10 000 m² and 7 500 m²,

respectively, which implies that these two sizes

would result in a better statistical performance.

TABLE 1

Shannon’s s (H’) and Jaccard’ Index for each plot size

Plot Size (m2) H’ SJ*

2 500 3.24 0.568

5 000 3.36 0.814

7 500 3.40 0.941

10 000 3.42

Note: The values of the Jaccard’s Index were obtained from

a comparison with the control plot - 10 000 m².

TABLE 2

Mean parameters of dendrometric atributes for each plot size investigated in this study

Variables Plot Size

2 500 m² 5 000 m² 7 500 m² 10 000 m²

DBH (cm)ns 36.91 36.64 36.34 36.24

Density (N ha-1)ns 160.00 146.66 145.83 146.16

Biomass (Mg ha-1)ns 344.67 288.61 279.54 278.32

Basal Area (m² ha-1)ns 22.38 19.83 19.34 19.09

*ns = no signficance for the Dunnett test (a = 5 %). The control was the 10 000 m² plot size. DBH = diameter at breast

height.

443

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 70: 437-449, January-December 2022 (Published Jun. 20, 2022)

Although there were no significant differ-

ences in these results (Table 3). Fig. 4 shows

that there is a tendency to reduce the coefficient

of variation according to the increase in the

plot area.

Through the AIC value, the best model

was Gamma (Table 4), admitting that the 2

units it’s sufficient to select one than other

model (Burnham & Anderson, 2002). So, the

Gamma model was used to represent the dia-

metric distribution for all plot size.

In the analysis of the diametric structure

obtained in the different plot sizes, the fitting

of the Gamma function is show in histograms

for each of the plot sizes, 2 500, 5 000 or 7 500

m², in comparison with the reference size,

10 000 m² (Fig. 4). The “inverted J” behavior,

characteristic of native forests, can be observed

in all of them.

It was also observed that the values of

the parameters obtained, in the fitting of the

diametric model (Table 5) for the different

TABLE 3

Statistics for density and biomass variables for each plot size

Variables Estimators Plot Size

2 500 m² 5 000 m² 7 500 m² 10 000 m²

Density (N.ha-1) Mean 146.19 146.41 146.47 146.11

Standard Deviation 18.55 13.61 10.65 9.09

CV (%) 12.56 9.22 7.22 6.18

Lower C.I. 109.08 119.19 125.16 127.92

Upper C.I. 183.29 173.63 167.78 164.29

Biomass (Mg.ha-1) Mean 284.85 282.40 285.76 284.00

Standard Deviation 19.66 14.28 10.72 9.47

CV (%) 6.46 4.76 3.53 3.18

Lower C.I. 245.54 253.83 264.32 265.06

Upper C.I. 324.17 310.96 307.20 302.94

Note: CV - coefficient of variation; C.I. - confidence interval (95 %).

Fig. 4. Coefficient of variation (CV) of populational density and biomass for each plot size.

444 Revista de Biología Tropical, ISSN: 2215-2075 Vol. 70: 437-439, January-December 2022 (Published Jun. 20, 2022)

plot sizes, were very close to each other. These

results contribute to the similarity across the

curves observed in the Fig. 5.

DISCUSSION

The results obtained regarding the compo-

sition of families corroborate with other stud-

ies carried out in the Amazon region, such as

those by Carim et al. (2013), Condé and Tonini

(2013), and Pereira et al. (2011). In these stud-

ies, the number of individuals and the species

richness contribute to the sovereignty of the

inventoried families, which gives the physiog-

nomic characteristics of the forest.

In the study conducted by Condé and

Tonini (2013), the families with the highest

number of species also represented a greater

richness of individuals. Their result differs

from the results obtained here since the Burs-

eraceae family corresponds for 29 % of the

total sampled individuals yet with the concen-

tration of a single species, Protium autissimum.

According to Amaral et al. (2000), abundance

patterns are quite variable for species and

families in general.

rectangular plots of 400 m² and 2 500 m².

Ubialli et al. (2009) also commented that a

plot size of 2 500 m² installed in a rectangular

shape produces accurate estimates for phytoso-

ciological studies when considering the most

important species groups from the economic

and phytosociological point of view, regardless

of the sampling process, but with a sampling

intensity of 10 %.

The choice of the size of the sampling

units in those cases in which the prediction

of the parameters is not compromised by the

reduction of the plot size can be made due to

the operationality of the location and implanta-

tion. In this case, smaller units are advisable

when time is a constraint, as suggested by Pél-

lico Netto and Brena (1997).

The Shannon diversity index (H’) of the

different plot sizes was higher than 3.11, which

according to Saporreti Jr. et al. (2003), indi-

cates well-preserved area. Based on this crite-

rion, the sampled forest can be characterized as

a well-preserved area.

In all the different plot sizes, a similar

assessment of the forest community diver-

sity was obtained. This same index was higher

when compared to savanna area in Brazil

TABLE 4

AIC value for the models fitted

Function Plot size (m²) AIC

Gamma 2 500 1 992.832

Weibull - 2p 2 056.68

Gamma 5 000 3 623.105

Weibull - 2p 3 750.546

Gamma 7 500 5 386.31

Weibull - 2p 5 581.771

Gamma 10 000 7 157.747

Weibull - 2p 7 419.148

TABLE 5

Gamma parameters for the four fitted models for each plot size

Plot Size (m²) Shape Standard Error Scale Standard Error p-value

2 500 5.1849 0.4588 0.1396 0.0130 < 0.0001

5 000 5.3388 0.3492 0.1464 0.0100 < 0.0001

7 500 5.4341 0.2910 0.1504 0.0084 < 0.0001

10 000 5.5578 0.2579 0.1542 0.0075 < 0.0001

A study carried out in the Ecotonal Forest

in the northern region of Mato Grosso, Ubialli

et al. (2009) concluded that, according to the

plot size, the distribution pattern of the species

can change, and plots with larger dimensions

up to 1 ha are indicated to obtain the real pat-

tern of the spatial distribution of species in

the area. To carry out phytosociological stud-

ies using a sampling intensity of 10 % and a

random process, the estimates were accurate

for the most economical and phytosociologi-

cal important species, especially when using

445

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 70: 437-449, January-December 2022 (Published Jun. 20, 2022)

(Finger & Finger, 2015; Guilherme et al., 2004;

Rodrigues-Sousa et al., 2015), with values

4.03, 3.86 and 3.85, respectively. The differ-

ence could be explained by the minimum diam-

eter measured, where in these three studies was

10 cm. However, same values for the Shannon

index was obtained by Apgaua et al. (2014).

The forest researched by the authors was

Tropical Forest with a dry season. The value

obtained was 3.6. According to Rodrigues-

Sousa et al. (2015), the Shannon index, when

evaluated for each indivudal, was 3.55 in

Riparian Forest.

In the evaluation of similarity, by Jac-

card index, the plots with the highest similarity

were the ones with 5 000 and 7 500 m², and

the one of 2 500 m² presented the lowest value

compared with the standard plot size of 10 000

m². However, for all plot sizes, the Jaccard

index was higher than 0.5, which, according

to Kent and Coker (1992), indicates high

similarity.

The mean values of the basal area for the

category DBH ≥ 20 cm were close to those

presented by Higuchi et al. (2004), in other

locations of the Brazilian Amazon. The values

presented by these authors ranged from 17.72

to 23.08 m².ha. The predicted biomass values

are close to those found in other studies as well.

Alves et al. (2003), found estimates of roughly

290 Mg.ha-1 for primary forests in Rondônia.

For the same type of forest, Brown et al. (1995)

obtained values of 285 Mg.ha-1. In discrepancy

with such results, Nascimento and Laurance

Fig. 5. Comparative analyses about distributions fitting by Gamma to four plot sizes. A. 10 000 m² x 2 500 m²; B. 10 000

m² x 5 000 m²; C. 10 000 m² x 7 500 m²; D. All plot sizes.

446 Revista de Biología Tropical, ISSN: 2215-2075 Vol. 70: 437-439, January-December 2022 (Published Jun. 20, 2022)

(2002) obtained an average biomass above

ground much higher, 397.7 Mg.ha-1, in intact

forests in the central Amazon.

It was observed that the effect of reducing

the size of plots on the prediction of biomass

and population density was not statistically

significant according to the Dunnett Test (Table

2). This behavior can be evidenced by observ-

ing the confidence intervals (Table 3) for

the different sizes of the sampling units for

the aforementioned dendrometric attributes,

in which the mean values are always contem-

plated. Still, it is worth mentioning that the

largest plots presented a smaller range ampli-

tude than the other plot sizes. Pinto et al. (2021)

evaluated the influence of the plot size in many

forests types and concluded that the bests sizes

were between 600 and 1 400 m² for the bio-

diversity parameters.

It was possible to verify that the estimation

of biomass and population density, considering

the dimensions of the plots, reached plausible

results, with coefficient of variation varying

from 6.18 to 12.56 % for density and 3.18 to

6.46 % for biomass (Fig. 5). This result corrob-

orates the findings of de Oliveira et al. (2014),

found that, for sampling of individuals with

DAP ≥ 20 cm in the Amazon forest, an uncer-

tainty level below 10 % was obtained from

sampling units with dimensions of 1 200 m².

Small plot sizes show a tendency to present

higher values of coefficient of variation. In this

study, the highest values of coefficient of varia-

tion for both biomass and population density

estimation was found in the 2 500 m² plot size.

The study by Cavalcanti et al. (2009), which

conducted in a forest characterized as Open

Forest and considered individuals with DAP ≥

40 cm, obtained a high coefficient of variation

for sampling units with 2 500 m². The authors

observed that the coefficient tended to decrease

as the area of the sampling unit increased up to

2 ha. This behavior was expected since larger

plots tend to better capture forest variability

(Wagner et al., 2010).

The diametric distribution in all sizes of

the sampling units followed the pattern of an

inverted J (Fig. 4). According to Campbell

et al. (1986), the inverted J is a characteristic

behavior for dryland and flooded Amazon

forests According to Silva et al. (2008), this

behavior represents a decrease in individuals

as the diametric classes increase. As stated by

Phillips et al. (1994), when a diametric struc-

ture presents in its first-class more than 65 % of

the sampled individuals, it exhibits the natural

dynamic of mortality and recruitment of new

individuals due to the death and tree fall.

When analyzed the parameters of the

Gamma function the results showed values

close to each other. This shows that the size of

the sampling units did not significantly reflect

the diametric structure of the forest. That is, all

the variability in DBH was captured for all plot

size, regardless of size.

After verifying the similarities between the

curves, we did an exercise just to exemplify

the use of these fitted curves. Trees larger than

or equal to 50 cm of DBH were predicted, as

this value is the minimum harvesting diam-

eter established by the environmental agency

(SEMA-MT) through decree 2152/2014. The

results obtained from the proportion of trees

above the established diameter were 22, 23, 23,

and 25 % for plots with 10 000, 7 500, 5 000,

and 2 500 m² respectively, values are also very

close to each other when comparing the small-

est plots with the 1 ha plot.

The diversity analysis of populations using

the Shannon index showed that the change in

the size of the plots did not show a significant

difference, implying that the use of smaller

plots captured the floristic diversity of the

forest. In the forest structure analysis, the

plot sizes did not influence the prediction of

biometric attributes of population density and

biomass, as well as the diametric structure.

Considering the level of inclusion DBH ≥ 20

cm, all dimensions of the analyzed plots in this

study could be used in forest inventories and,

despite the values in coefficient of variation

and confidence interval, plots with areas of

2 500 m² is recommended.

Ethical statement: the authors declare

that they all agree with this publication and

447

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 70: 437-449, January-December 2022 (Published Jun. 20, 2022)

made significant contributions; that there is no

conflict of interest of any kind; and that we fol-

lowed all pertinent ethical and legal procedures

and requirements. All financial sources are

fully and clearly stated in the acknowledge-

ments section. A signed document has been

filed in the journal archives.

ACKNOWLEGMENTS

To ONF Brazil for all supports to develop

this study.

RESUMEN

Impacto del tamaño de la parcela en la estructura y

estimación de la diversidad de los bosques tropicales

Introducción: Los inventarios son fundamentales para la

gestión forestal, pero en la Amazonía la ausencia de estan-

darización produce pérdida de información, baja precisión

y mediciones inconsistentes. Esto impide comparaciones

válidas y compromete el uso de información en redes y

programas. El tamaño de la unidad de muestreo es de

importancia clave en el inventario de bosques nativos, par-

ticularmente en lo que respecta a la precisión y los costos.

Objetivo: Identificar un tamaño de parcela que proporcio-

ne una precisión adecuada para los parámetros dendromé-

tricos en la Amazonía.

Métodos: En Cotriguaçu, Mato Grosso, Brasil, probamos

cuatro tamaños de parcela con seis repeticiones cada una:

2 500, 5 000, 7 500 y 10 000 m². Medimos diámetro a

la altura del pecho, densidad de población, área basal y

biomasa. Se aplicaron los índices de Shannon y Jaccard;

Funciones Weibull 2P y Gamma para adaptarse a la distri-

bución diametral; y el Criterio de Información de Akaike

para el mejor modelo.

Resultados: Hubo una relación directamente proporcional

entre el área de parcela y la similitud poblacional, pero la

diversidad no indicó alteraciones significativas. El tamaño

de la parcela no afectó los atributos dendrológicos y la

distribución diametral. Las áreas de parcela más grandes

dieron lugar a coeficientes de variación más bajos e inter-

valos de confianza más pequeños. La función Gamma fue

el mejor modelo para representar las distribuciones de

diferentes tamaños de parcela.

Conclusiones: Para bosques similares, recomendamos la

parcela de 2 500 m² para evaluar diámetro a la altura del

pecho, densidad de población, área basal y biomasa.

Palabras clave: gestión forestal sostenible; muestreo;

análisis florístico; análisis estructural; ecología tropical.

REFERENCES

Alvares, C. A., Stape, J. L., Sentelhas, P. C., de Moraes

Gonçalves, J. L., & Sparovek, G. (2013). Köppen’s

climate classification map for Brazil. Meteorolo-

gische Zeitschrift, 22, 711-728. https://doi.org/

10.1127/0941-2948/2013/0507

Alves, D., Soares, J. V., Amaral, S., Mello, E., Almei-

da, S., da Silva, O. F., & Silveira, A. (2003).

Biomass of primary and secondary vegetation

in Rondônia, Western Brazilian Amazon. Glo-

bal Change Biology, 9(3), 451-461. https://doi.

org/10.1046/j.1365-2486.1997.00081.x

Amaral, I. L., Matos, F. D. A., & Lima, J. (2000). Compo-

sição florística e estrutural de um hectare de floresta

densa de terra firme no Rio Uatumã, Amazônia, Bra-

sil. Acta Amazonica, 30, 377-392.

Apgaua, D. M. G., Coelho, P. A., Santos, R. M., Santos, P.

F., & Oliveira-Filho, A. T. (2014). Tree community

structure in a seasonally dry tropical forest remnant,

Brazil. Cerne, 20(2), 173-182.

de Araujo, H. J. B. (2006). Inventário florestal a 100% em

pequenas áreas sob manejo florestal madeireiro. Acta

Amazônica, 36(4), 447-464.

Brown, I. F., Martinelli, L. A., Thomas, W. W., Moreira,

M. Z., Ferreira, C. A. C., & Victoria, R. A. (1995).

Uncertainty in the biomass of Amazonian forests:

An example from Rondônia, Brazil. Forest Eco-

logy and Management, 75, 175-189. https://doi.

org/10.1016/0378-1127(94)03512-U

Burnham, K. P., & Anderson, D. R. (2002). Model Selec-

tion and Multimodel Inference: A Pratical Informa-

tion-Theoretic Approach. Springer.

Campbell, D. C., Daly, D. C., Prance, G. T., & Maciel, U.

N. (1986). Quantitative ecological inventory of terra

firme and várzea tropical forest on the rio Xingu,

Brazilian Amazon. Brittonia, 38(4), 369-393. https://

doi.org/10.2307/2807085

Carim, M. J. V., Guillaumet, J-L. B., Guimarães, J. R. S.,

& Tostes, L. C. L. (2013). Composição e Estrutura

de Floresta Ombrófila Densa do extremo Norte do

Estado do Amapá, Brasil. Biota Amazônica, 3(2),

1-10. http://dx.doi.org/10.18561/2179-5746/biotaa-

mazonia.v3n2p1-10

Cavalcanti, F. J. B., Machado, S. D. A., & Hosokawa, R.

T. (2009). Tamanho de unidade de amostra e intensi-

dade amostral para espécies comerciais da Amazônia.

Floresta, 39(1), 207-214. http://dx.doi.org/10.5380/

rf.v39i1.13740

Cavalcanti, F. J. B., Machado, S. D. A., Osokawa, R. T.,

& da Cunha, U. S. (2011). Comparação dos valores

estimados por amostragem na caracterização da

estrutura de uma área de floresta na Amazônia com

as informações registradas no censo florestal. Revista

448 Revista de Biología Tropical, ISSN: 2215-2075 Vol. 70: 437-439, January-December 2022 (Published Jun. 20, 2022)

Arvore, 35(5), 1061-1068. https://doi.org/10.1590/

S0100-67622011000600012

Chave, J., Andalo, C., Brown, S., Cairns, M. A., Chambers,

J. Q., Eamus, D., Folster, H., Higuchi, N., Kira, T.,

Lescure, J.-P., Nelson, B. W., Ogawa, H., Puig, H.,

Riera, B., & Yamakura, T. (2005). Tree allometry and

improved estimation of carbon stocks and balance in

tropical forests. Oecologia, 145, 87-99. https://doi.

org/10.1007/s00442-005-0100-x

Condé, T. M., & Tonini, H. (2013). Fitossociologia de uma

Floresta Ombrófila Densa na Amazônia Setentrional,

Roraima, Brasil. Acta Amazônica, 43(3), 247-260.

https://doi.org/10.1590/S0044-59672013000300002

Dunnett, C. W. (1955). A multiple comparison procedure

for comparing several treatments with a control.

Journal of the American Statistical Association, 50,

1096-1121.

Farias, L. L. (2012). Tamanho e forma de parcela amostral

para inventários de espécies não madeireiras da

Amazônia central (Master’s thesis). National Institute

of Amazon Research, Brazil. https://bdtd.inpa.gov.br/

handle/tede/1125

Finger, Z., & Finger, F. A. (2015). Fitossociologia em

comunidades arbóreas remanescentes de cerrado

sensu stricto no Brasil central. Floresta, 45(4), 769-

780. http://dx.doi.org/10.5380/rf.v45i4.30860

Guilherme, F. A., Morellato, L. P., & Assis, M. A. (2004).

Horizontal and vertical tree community structure in

a lowland Atlantic Rain Forest, Southeastern Brazil.

Brazilian Journal of Botany, 27(4), 725-737. https://

doi.org/10.1590/S0100-84042004000400012

Higuchi, N., Chambers, J., Dos Santos, J., Ribeiro, R. J.,

Pinto, A. C. M., da Silva, R. P., Rocha, R. M., &

Tribuzy, E. S. (2004). Dinâmica e balanço do carbono

da vegetação primária da Amazônia central. Floresta,

34, 295-304.

Hugo, S., & David, R. D. (2021). ecolTest: Community

Ecology Tests (Version 0.0.1, R Package). https://

github.com/hugosal/ecolTest

Hutcheson, K. (1970). A Test for Comparing Diversities

Based on the Shannon Formula. Journal of Theoreti-

cal Biology, 29, 151-54.

Kent, M., & Coker, P. (1992). Vegetation Description

Analyses. Behaven Press.

Magurran, A. E. (1988). Ecological Diversity and Its Mea-

surement. Princeton University Press.

ME (Ministry of Economic). (2012). Manual Técnico da

Vegetação Brasileira. Ministry of Economic, Brazil.

ME (Ministry of Economic). (2019). Biomas e Sistema

Costeiro-Marinho do Brasil: Compatível com a Esca-

la 1:250 000. Ministry of Economic, Brazil.

MME (Ministry of Minas and Energy). (1982). Project

RADAM Brazil: Survey of Natural Resources. Minis-

try of Minas and Energy, Brazil.

Nascimento, H. E. M., & Laurance, W. F. (2002). Total

aboveground biomass in central Amazonian rainfo-

rests: a landscape-scale study. Forest Ecology and

Management, 168, 311-321. https://doi.org/10.1016/

S0378-1127(01)00749-6

de Oliveira, M. M., Higuchi, N., Celes, C. H., & Higuchi, F.

G. (2014). Tamanho e formas de parcelas para inven-

tários florestais de espécies arbóreas na Amazônia

central. Ciência Florestal, 24(3), 645-653. https://doi.

org/10.5902/1980509815744

Péllico Netto, S., & Brena, D. A. (1997). Inventário Flo-

restal. Curitiba Federal University of Paraná Press.

Pereira, L. A., Sobrinho, F. A. P., & Costa Neto, S. V.

(2011). Florística e Estrutura de Uma Mata de Terra

Firme na Reserva de Desenvolvimento Sustentável

Rio Iratapuru, Amapá, Amazônia Oriental, Brasil.

Floresta, 4(1), 113-122. http://dx.doi.org/10.5380/

rf.v41i1.21191

Phillips, O. L., Hall, P., Gentry, A. H., Sawyer, S. A., &

Vásquez, R. (1994). Dynamics and species richness

of tropical rain forests. Proceedings of the National

Academy of Sciences of the USA, 91(7), 2805-2809.

Pinto, L. O. R., Souza, C. R., Terra, M. C. N. S.,

Mello, J. M., Calegario, N., & Arcebi Junior, F.

W. (2021). Optimal plot size for carbon-diversity

sampling in tropical vegetation. Forest Ecology and

Management, 482, 118778. https://doi.org/10.1016/j.

foreco.2020.118778

QGIS Development Team. (2018). QGIS Geographic

Information System (version 3.0, software). Open

Source Geospatial Foundation Project. http://qgis.

osgeo.org

R Core Team. (2019). R: A language and environment

for statistical computing (version 3.6.0, Computer

Software). R Foundation for Statistical Computing.

https://www.R-project.org

Rodrigues-Souza, J., Prado Júnior, J. A., Vale, V. S.,

Schiavin, I., de Oliveira, A. P., & Arantes, C. S.

(2015). Secondary forest expansion over a savanna

domain at an ecological reserve in the Southeastern

Brazil after 15 years of monitoring. Brazilian Journal

of Botany, 38(2), 311-322. https://doi.org/10.1007/

s40415-015-0146-x

Saporetti Jr, A. W., Meira Neto, J. A. A., & Alma-

do, R. P. (2003). Fitossociologia de cerrado sensu

stricto no município de Abaeté-MG. Revista

Árvore, 27(3), 413-419. https://doi.org/10.1590/

S0100-67622003000300020

Signorell, A., Aho, K., Alfons, A., Anderegg, N., Ara-

gon, T., Arachchige, C., Arppe, A., Baddeley, A.,

449

Revista de Biología Tropical, ISSN: 2215-2075, Vol. 70: 437-449, January-December 2022 (Published Jun. 20, 2022)

Barton, K., Bolker, B., Borchers, H. W., Caeiro, F.,

Champely, S., Chessel, D., Chhay, L., Cooper, N.,

Cummins, C., Dewey, M., Doran, H. C., ... Zeileis,

A. (2021). DescTools: Tools for descriptive statistics

(Version 0.99.44, Computer Software). https://andri-

signorell.github.io/DescTools

Silva, J. N. M. (1980). Eficiência de diversos tamanhos

e formas de unidades de amostras aplicadas em

inventário florestal na região do baixo Tapajós

(Unpublished master’s thesis). Federal University of

Parana, Brazil.

Silva, J. N. M., & Lopes, J. C. A. (1984). Inventário Flo-

restal Contínuo em Florestas Tropicais: A Metodolo-

gia utilizada pela EMBRAPA-CPATU na Amazônia

Brasileira. Embrapa-CPATU. https://www.embra-

pa.br/busca-de-publicacoes/-/publicacao/381691/

inventario-florestal-continuo-em-florestas-tropicais-

a-metodologia-utilizada-pela-embrapa-cpatu-na-

amazonia-brasileira

Silva, J. N. M., Lopes, J. C. A., de Oliveira, L. C., da Silva,

S. M. A., de Carvalho, J. O. P., Costa, D. H. M.,

Melo, M. S., & Tavares, M. J. M. (2005). Diretrizes

para Instalação e Medição de Parcelas Permanen-

tes em Florestas Naturais da Amazônia Brasileira.

Embrapa Amazonia Oriental. https://www.embrapa.

br/busca-de-publicacoes/-/publicacao/954902/dire-

trizes-para-instalacao-e-medicao-de-parcelas-perma-

nentes-em-florestas-naturais-da-amazonia-brasileira

Silva, K. E., Matos, F. D. A., & Ferreira, M. M. (2008).

Composição florística e fitossociologia de espécies

arbóreas do Parque Fenológico da Embrapa Amazônia

Ocidental. Acta Amazonica, 38(2), 213-222. https://

doi.org/10.1590/S0044-59672008000200004

Silveira, A. B. (2019). Influência do tamanho de parcelas

na estimativa de parâmetros biométricos em uma

área de manejo florestal sustentável no Mato Grosso

(Master’s thesis). Federal University of Mato Grosso,

Brazil. www.ufmt.br/curso/ppgcfa.

Ubialli, J. A., Filho, A. F., do Machado, S. A., & Arce, J.

E. (2009). Comparação de métodos e processos de

amostragem para estudos fitossociológicos em uma

floresta ecotonal na região norte matogrossense.

Floresta, 39(3), 511-523. http://dx.doi.org/10.5380/

rf.v39i3.15351

Vianna, A. L. M., Vidal, E. J., Gorenstein, M. R., & Batista,

J. L. F. (2010). Tamanho ótimo de parcela para esti-

mativa de estoque de carbono da parte aérea lenhosa

de fitofisionomias florestais na Amazônia oriental.

Floresta, 40(2), 447-456. http://dx.doi.org/10.5380/

rf.v40i2.17839

Wagner, F., Rutishauser, E., Blanc, L., & Herault, B.

(2010). Effects of Plot Size and Census Inter-

val on Descriptors of Forest Structure and

Dynamics. Biotropica 42(6), 1-8. https://doi.

org/10.1111/j.1744-7429.2010.00644.x

Zanne, A. E., Lopez-Gonzalez, G., Coomes, D. A., Ilic,

J., Jansen, S., Lewis, S. L., Miller, R. B., Swenson,

N., Wiemann, M. C., & Chave, J. (2009). Towards

a worldwide wood economics spectrum [Data set].

https://doi.org/10.5061/dryad.234.