Estimation of Diameters by Stump Measurements for
Natural and Artificial Small-Leaved Linden (Tilia cordata)
Aydar Gabdelkhakov*, Liubov Blonskaya,
Ildar Sabirzyanov, Regina Baiturina and Ilnur Mullagaleev
Department of Forestry and
Landscape Design, Federal State Budgetary Educational Establishment of Higher
Education «Bashkir State Agrarian University», Ufa, Russian Federation
*For correspondence: a_gabdelkhakov@rambler.ru
Received 28 April
2021; Accepted 17 June 2021; Published 28 September 2021
Abstract
This study aimed at to assess the relationship between
stump diameter (DST) at the height of 0.1 m above ground level and diameter at
1.3 m (DBH – diameter at breast height) for eight different
age classes (IV–XI); of plantings of small-leaved linden (Tilia cordata Mill.). From 4,523 pairs of DBH and DST
measurements, several simple linear models representing the DBH – DST
relationship have been developed and evaluated. The field data processing was
carried out using the methods generally accepted in forest inventory and
variation statistics. For the dependence DBH = b×DST
for each trial plot and pooled samples, the values of the coefficients b,
significance, errors (standard, relative, systematic and random) were
established. Age classes were compared according to the F-criterion, it was
concluded that they differ significantly from each other. Verification of the
data obtained with the standards developed in the 1980s showed that their
accuracy is acceptable for forest managers when assessing tree volume removed
in local conditions. However, for an accurate assessment, for example,
scientific research, tracking forest management history, etc., a differentiated
assessment is necessary, considering the origin, age and conditions of the
habitats. The research results can help model and plan management in stands of
the same age with trees removed for various reasons. ©
2021 Friends Science Publishers
Keywords: Forecast; Tree diameter;
Small-leaved linden; Diameter at 1.3 m; Diameter at stump height; Linear model;
F-criterion
Introduction
In forestry, volumetric tables
are used to determine the stock of forest stands. Depending on the number of
parameters, they are divided into three types: one, two and several indicators.
In tables with one input, it is enough to know the diameter at the height of 1.3 m (DBH), with two inputs – DBH and the tree's height (or the
category of heights) with an average shape of the trunk. With several inputs
and the previous indicators, the form factor is considered, and in exceptional
cases and other forest biological characteristics
of the tree (Verkhunov and Chernykh
2007; Sağlam et
al. 2016).
The volumetric tables with one
input are used more often because DBH is easy to measure. The variability of
the volumes of the trunks within the same diameter when using local assortment
tables is taken into account due to the group of diameters represented by 4 cm
steps of thickness, the group of heights in grades of heights with an average
shape factor.
Various reasons may require the
restoration of diameters and heights of already felled trees: a description of
the structure (Costa et al. 2019) and
restoration of the taxation characteristics of the stand before felling (Kulla et al.
2017). This may be an assessment of the damage caused by illegal felling (Abdullaeva and Khairov 2019) or
the results of catastrophic events (Di Cosmo and Gasparini 2020). This may
require tracking the history of management activities (García-Cuevas et al. 2017; Paramonov
et al. 2020), and a review of the intensity
of fell, the acquisition of skills in eye taxation at the preparatory stage of
forest inventory etc. (Milios et al. 2016; Asrat et al. 2020; Mugasha 2021). In such cases, only the stump parameters have
to be used. A high correlation dependence allows models for finding the DBH of
felled trees by stump
diameter (DST) at the height of 0.1 m as an independent variable (Ercanli et al.
2015).
The number of studies in many world countries on DBH/DST ratios of
various woody species is not decreasing. Research in this direction is
being conducted over many parts of the earth including North (Westfall 2010;
Pond and Froese 2014; García-Cuevas et
al. 2017) and South (Costa et al.
2019) America; Western (Diéguez-Aranda et al. 2003), Southern (Milios et al. 2016; Di Cosmo and Gasparini
2020), Central (Bruchwald 2001; Kulla
et al. 2017) and Eastern (Abdullaeva and Khairov 2019; Paramonov et al.
2020) Europe, Western Asia (Sasanifar et al. 2015; Şahin
et al. 2019; Şenyurt
et al. 2020) and Africa (Asrat et al. 2020; Chukwu
et al. 2020; Mugasha
2021). In Turkey alone, about two dozen studies have
been conducted on the relationships between DST and DBH (Şahin
et al. 2019). These works considered
one (Weiss 2013; Chukwu et al. 2020) to 30 common tree species (Asrat et al. 2020; Di Cosmo and Gasparini 2020) in the studied regions. Both coniferous (Şenyurt et al. 2020) and deciduous (Sasanifar et al.
2015) trees, which have primary and secondary commercial value were studied
(Costa et al. 2019; Paramonov et al.
2020). Özdemir et
al. (2020) determine the DST/DBH ratio for pure stands of rock oak (Quercus petraea (Matt.) Liebl), and Sakici and Özdemir (2017) for
mixed stands of oriental beech (Fagus orientalis) and Kazdag fir (Abies nordmanniana
subsp. equi-trojani) plant. To solve the problem
of finding DBH depending on DST, many authors considered the possibility of
using both simple linear and nonlinear equatins (Milios et al. 2016; Sağlam
et al. 2016; Özdemir
et al. 2020). They also used the
stump height's effect on the accuracy of DBH determination (Diéguez-Aranda
et al. 2003; Pond and Froese 2014; Sakici and Özdemir 2017) and
artificial neural networks to model the relationship and differentiate DBH from
DST (Sakici and Ozdemir
2018; Şenyurt et al. 2020).
Lack of empirical information on the size of felled trees can impede the
conviction of illegal loggers. In the legal proceedings in the Russian
Federation, the assessment of damage by volume of wood is carried out at 4 cm
steps of thickness and the first category of heights in the bark (Abdullaeva and Khairov 2019). To describe the structure of stands and restore their inventory
characteristics before logging, more precise values of the diameters and
heights of the removed trees are needed (Bruchwald
2001; Kulla et
al. 2017). It is noted that biases in the definition of DBH lead to
distortions in the assessment of growing stock (Pond and Froese 2014).
Based on regression models, diagrams or tables (VNIITslesresurs
1991; Corral-Rivas et al. 2007; Milios et al. 2016), works on DBH estimation by DST are applied both everywhere and
locally. Simultaneously, the lack of accuracy is noted since the conditions of
growing stands, the category of heights, the entirety of stands etc., are not
considered. Consequently, the most suitable option for different geographic
regions is developing local models and standards for assessing the felled stock
that meet the existing requirements and conditions for tree growth (Westfall, 2010; Sağlam et al., 2016; Kulla
et al. 2017).
Small-leaved linden (Tilia cordata Mill.) is a widespread species; its
stands occupy 22% of the forested lands of the territory of the Republic of
Bashkortostan (1148.4 thousand hectares) with a total reserve of 209.3 thousand
m3 (Sultanova et al. 2020). Linden is widespread not only in Russia but also in the
temperate zone of Europe. However, no previous scientific studies have been
traced in the literature that would show a DST – DBH relationship, which
creates uncertainty in obtaining biometric parameters of trees removed from the
forest and urban environment. For this reason, our study aimed at to determine
the relationship between DST and DBH for eight different age classes of natural
and artificial stands of linden and to develop a
predictive model for DBH based on DST measurements of distant trees. The tasks
undertaken to achieve this goal were to a) develop a simple linear model with a single
coefficient for the transition from DST to DBH; b) evaluate the significance
and errors of regression models; c) tabulate the generalized model for the
definition of DBH from DST and d) verify the existing regulations.
Materials
and Methods
The studies were carried out in the territory of Ufa
and the Ufa municipal district. The area is characterized on an average by
coordinates 54°70′ N 55°90′ E etc.; at an altitude of 150 m above
sea level (Fig. 1), the climate is quite humid-continental. Average annual air
temperature is 3.0°C, January is characterized by an average temperature of
-14.5°C, July 19.5°C with an absolute maximum of 40°C and an absolute minimum
of -50°C. The average annual precipitation is in the range of 500–600 mm,
during the growing season about 350 mm. Under these conditions of growth, the
small-leaved linden grows according to quality classes I to III. In this work,
natural and artificial stands of small-leaved linden ten temporary test plots
(TPs) of various age classes (III-XI) were studied. Also, trees of VII class of
age of free growth, planted on the streets of Ufa, were studied (Martynova et al.
2020). The duration of age classes for small-leaved linden was ten years.
The TPs were planted with a size of 0.1 ha or
more, depending on the stand's specific entirety, so that each of them was a
homogeneous plantation. At each site, DBH and DST of all growing trees were
>3.9 cm with bark in two mutually perpendicular directions, which were
measured with an accuracy of 1 cm. The stump height was considered to be no
more than 10 cm when cutting trees with DBH thinner than 30 cm (for thicker
ones - no more than one-third of DST) from the soil surface and when the roots
are exposed - from the root collar. Heights were measured according to the data
of taxation descriptions and lists of forest cultures. The age was determined
by counting the annual rings of the model trees. The rest of the forest stands
were calculated using the counted trees (Verkhunov
and Chernykh 2007).
The main densitometric characteristics of the
investigated plantations are shown in Table 1. A total
of 4994 trees were measured. Statistical processing of the obtained research
results was carried out using computer programs Microsoft Excel and Statistica. A simple linear function without a free
coefficient was tested by the least-squares method for each TP, for each age
class as a whole (III–XI) and pooled samples to estimate DBH by DST.
The applicability of the
obtained equations was assessed by the coefficient of determination (R2),
standard (Se), relative (Sо), systematic (Qp) and random (Qs)
errors (formulas 1–5):
(1)
(2)
(3)
(4)
(5)
where yi - are actual DBH;
ŷi – are calculated DBH values
calculated by substituting DST values into the regression equations;
ȳ - is the arithmetic mean of yi;
n - is the sample size;
p - is the number of equation
parameters (in our case, p = 1).
A "nonlinear complementary
sum of squares" method was used to determine the likelihood of differences
in the equation DBH = f (DST) between age classes. This method is based on
creating complete and reduced models, which is used to detect differences
between tree species and geographic regions (Corral-Rivas et al. 2007; Özçelík et al. 2010), age classes (Özdemir et al.
2020). Each age class is defined using a different set of parameters in the
full model, while in the shorthand model, all age classes are defined with the
same parameters. The equality of the complete and reduced models is checked
using the F-test:
(6)
where SSER and dfR are the errors sums and the freedom degree
of the reduced model, and SSEF and dfF are the sums of the squares of the errors and the degree of freedom of
the entire model.
The null hypothesis of the
model`s equality is rejected if the F, calculated by formula 6, takes a higher
value than the tabular Fst with a
probability of 95% and the corresponding number of degrees of freedom.
Consequently, there is a statistical difference between the age classes for the
models. Conversely, suppose the null hypothesis is adopted. In that case, it is
concluded that there is no significant difference between the age classes for
the DST to DBH models, and a single equation can describe this relationship.
Additionally, the comparison of
the series was carried out by calculating the standard deviation (σ, %):
(7)
where ai and bi are pairwise
compared DBH data;
n - is the number of compared pairs, pcs.
Results
Statistical processing of the starting material is
presented in Table 2. The study range is for DST from 4 to 66 cm and from 3 to
55 cm - for DBH. The distribution series of DBH stands for SP to corresponded
to the normal distribution: the coefficients of asymmetry and kurtosis were
within their main twofold errors (except for the stands for SP 2, 14). A normal
distribution characterizes DST for 13 stands, for seven - by nominal values of
extension (kurtosis - negative) and rows' asymmetry. The volume of material was
sufficient for a reliable characterization of the average values since the
experiment's accuracy did not exceed 3%. The coefficient of variation for TPs
varied from 20 to 44% for DBH and from 20 to 45% for DST.
The obtained statistical
indicators testified the reliability of the empirical material and gave the
right, on their basis, to reveal the dependencies and patterns of changes in
DBH on DST. Based on empirical data, simple linear functions of the transition
from DST to DBH were calculated for all test plots, age groups and altitude
categories (Table 3). The gradation of the heights categories was adopted
following the current assortment tables for forest stands of the Cis-Urals.
All equations were characterized
by relatively high coefficients of determination (R2> 0.7),
except for the model for trees of open growth of the urban environment
(0.45). Coefficient b was in the range of 0.8167–0.8727 for natural stands,
0.7787–0.8561 for artificial stands, and 0.7917 for free growth trees. It was
significant for all models (p <0.01). The Fisher criterion's calculated values significantly exceeded the critical ones (F> Fst), indicating the models' reliability. The Se
value was in the range of 0.9–2.9 cm, Sо - 0.07–0.15, Qp varies
from –5.6 to 3.4%, and Qe did not exceed
15%. It also testified the adequacy of the obtained equations.
Graphical analysis and similar
values of the b coefficients showed uniformity in plots for trial plots - they
merge into one line, despite the difference in the heights. A dense correlation
field and a visible form of connection gave the basis to combine data by age
classes, two categories of heights, and a single sample. Simple linear
equations were also compiled, and their statistical indicators were found
(Table 3).
The results of comparing the
equations by age class using the F-test are shown in Table 4. Comparison of the
equations of each pair of TPs by age classes and in general showed that there
was a statistical difference between them at the significance level α = 0.05 (F > Fst = 3.9), except for natural stands VIII (TP3 and TP4) and X (TP6 and
TP8; TP6 and TP9), as well as forest cultures of VI (TP15 and TP16) age
classes.
The DBH/DST ratio depends
significantly on the shape of the trunk in the butt part, not on the entire
shape of the trunk and the tree's age. Despite the differences in the model's
age classes, we used the generalized equation.
Table 1: Main dendrometric
characteristics of TPs
TP number |
Age (years) |
Average height (m) |
Average diameter (cm) |
Total basal area (m2 ha-1) |
Height class |
Natural forest stand |
|||||
1 |
28 |
14.0 |
12.4 |
25.56 |
2 |
2 |
38 |
15.0 |
10.4 |
24.12 |
1 |
3 |
70 |
19.8 |
24.8 |
33.09 |
2 |
4 |
75 |
21.5 |
25.2 |
38.73 |
2 |
5 |
85 |
19.8 |
26.0 |
40.91 |
2 |
6 |
97 |
22.1 |
32.5 |
33.79 |
2 |
7 |
100 |
23.4 |
30.6 |
37.43 |
1 |
8 |
100 |
22.6 |
29.0 |
42.90 |
1 |
9 |
100 |
24.4 |
32.0 |
36.25 |
1 |
10 |
110 |
21.0 |
26.0 |
32.52 |
2 |
Artificial forest stand |
|||||
11 |
39 |
16.0 |
14.2 |
42.07 |
2 |
12 |
48 |
15.0 |
12.6 |
27.58 |
1 |
13 |
50 |
18.7 |
16.1 |
31.81 |
1 |
14 |
52 |
17.0 |
14.1 |
33.92 |
1 |
15 |
55 |
13.9 |
13.3 |
17.49 |
2 |
16 |
55 |
18.7 |
14.6 |
25.24 |
1 |
17 |
63 |
20.0 |
19.3 |
32.89 |
1 |
18 |
63 |
20.0 |
19.6 |
36.71 |
1 |
19 |
71 |
22.0 |
22.2 |
38.62 |
1 |
20 |
79 |
26.0 |
25.2 |
34.11 |
1 |
Open urban trees |
|||||
21 |
65 |
13.6 |
36.4 |
- |
<4 |
Table 2:
Summarized descriptive statistics of tree diameters on TPs
Parameters |
Distribution
series statistics* |
||||||
Х, cm |
S |
Хmin, cm |
Хmax, cm |
As |
Ex |
||
Natural
forest stand |
|||||||
DST |
min |
11,7 |
4,42 |
4 |
22 |
-0,78 |
-0,82 |
max |
37,2 |
11,21 |
16 |
66 |
0,62 |
1,04 |
|
DBH |
min |
9,6 |
3,79 |
3 |
19 |
-0,33 |
-0,73 |
max |
31,6 |
10,34 |
15 |
55 |
0,59 |
0,19 |
|
Artificial
forest stand |
|||||||
DST |
min |
12,8 |
4,02 |
4 |
25 |
0,06 |
-0,30 |
max |
30,4 |
7,63 |
12 |
51 |
1,01 |
6,11 |
|
DBH |
min |
10,7 |
3,53 |
4 |
21 |
0,03 |
-0,42 |
max |
24,7 |
6,35 |
12 |
42 |
1,93 |
5,61 |
|
Open
urban trees |
|||||||
DST |
44,5 |
9,25 |
10 |
70 |
-0,22 |
0,92 |
|
DBH |
35,6 |
7,43 |
9 |
65 |
-0,10 |
1,55 |
Notes. X is the arithmetic mean, cm; S - standard deviation, cm; Xmin - minimum value, cm; Xmax
- maximum value, cm; As is the coefficient of asymmetry; Ex - kurtosis
coefficient
Fig. 1: The
spatial location of the data collection area
Table 3: Statistical indicators of the relationship model DBH
= b×DST*
TP number / age class |
n, pcs |
b |
R2 |
Errors of equations for TPs, sampling by the height category of and
combined sampling |
σ |
||||||||||||
Se |
Sо |
Qp |
Qs |
||||||||||||||
TP |
h.c. |
total |
TP |
h.c. |
total |
TP |
h.c. |
total |
TP |
h.c. |
total |
h.c. |
total |
||||
Natural forest stand |
|||||||||||||||||
1/III |
176 |
0.8223 |
0.932 |
1.1 |
1.1 |
1.1 |
0.08 |
0.08 |
0.08 |
-0.8 |
1.5 |
1.1 |
8.4 |
8.2 |
8.3 |
5.0 |
4.9 |
2/IV |
228 |
0.8254 |
0.949 |
0.9 |
0.9 |
0.9 |
0.10 |
0.10 |
0.10 |
0.6 |
2.3 |
2.4 |
9.6 |
9.4 |
9.4 |
3.8 |
4.0 |
3/VIII |
169 |
0.8290 |
0.902 |
1.9 |
1.9 |
1.9 |
0.07 |
0.07 |
0.07 |
-0.8 |
0.6 |
0.6 |
7.5 |
7.4 |
7.4 |
3.2 |
3.1 |
4/VIII |
243 |
0.8167 |
0.789 |
2.2 |
2.3 |
2.3 |
0.09 |
0.09 |
0.09 |
-0.9 |
2.0 |
2.0 |
9.0 |
8.8 |
8.8 |
6.6 |
6.4 |
3+4/VIII |
412 |
0.8216 |
0.847 |
2.1 |
2.7 |
2.2 |
0.09 |
0.08 |
0.08 |
-0.9 |
1.5 |
1.4 |
8.5 |
8.3 |
8.3 |
5.2 |
5.1 |
5/IX |
235 |
0.8501 |
0.871 |
1.9 |
1.9 |
1.9 |
0.07 |
0.08 |
0.07 |
-0.6 |
-1.6 |
-1.7 |
7.3 |
7.3 |
7.3 |
2.1 |
2.3 |
6/X |
108 |
0.8579 |
0.851 |
2.6 |
2.7 |
2.7 |
0.08 |
0.08 |
0.09 |
-0.4 |
-2.4 |
-2.4 |
8.2 |
8.4 |
8.4 |
4.0 |
4.1 |
7/X |
177 |
0.8277 |
0.864 |
2.8 |
2.9 |
2.9 |
0.09 |
0.08 |
0.08 |
-1.5 |
-0.1 |
0.1 |
8.5 |
8.4 |
8.4 |
3.2 |
3.4 |
8/X |
141 |
0.8606 |
0.881 |
2.1 |
2.2 |
2.2 |
0.07 |
0.08 |
0.08 |
0.3 |
-2.2 |
-2.1 |
7.3 |
7.5 |
7.4 |
5.0 |
4.7 |
9/X |
149 |
0.8405 |
0.825 |
2.9 |
2.9 |
2.9 |
0.09 |
0.09 |
0.09 |
-0.1 |
-0.1 |
0.1 |
9.2 |
9.2 |
9.2 |
0.2 |
0.1 |
6+7+8+9/X |
575 |
0.8442 |
0.854 |
2.7 |
- |
2.7 |
0.08 |
- |
0.08 |
-0.6 |
- |
-0.9 |
8.4 |
- |
8.5 |
- |
0.8 |
10/XI |
146 |
0.8727 |
0.974 |
1.7 |
1.9 |
1.9 |
0.08 |
0.09 |
0.09 |
1.7 |
-1.9 |
-2.0 |
8.3 |
8.6 |
8.6 |
7.3 |
7.4 |
1 height
category |
695 |
0.8399 |
0.960 |
- |
2.3 |
2.3 |
- |
0.09 |
0.09 |
- |
0.3 |
0.4 |
- |
8.9 |
8.9 |
- |
- |
2 height
category |
1077 |
0.8415 |
0.942 |
- |
1.9 |
1.9 |
- |
0.07 |
0.07 |
- |
-0.3 |
-0.3 |
- |
7.6 |
7.6 |
- |
- |
Total (1-10) |
1772 |
0.8409 |
0.952 |
- |
- |
1.5 |
- |
- |
0.06 |
- |
- |
-0.2 |
- |
- |
5.9 |
- |
- |
Artificial forest stand |
|||||||||||||||||
11/IV |
145 |
0.7787 |
0.695 |
2.1 |
21 |
2.3 |
0.15 |
0.15 |
0.14 |
-2.2 |
-0.9 |
3.4 |
14.7 |
14.6 |
13.9 |
19.5 |
12.3 |
12/V |
173 |
0.8322 |
0.886 |
1.2 |
1.2 |
1.2 |
0.10 |
0.10 |
0.10 |
-0.1 |
-0.5 |
-1.2 |
9.7 |
9.7 |
9.8 |
0.9 |
2.2 |
13/V |
236 |
0.8561 |
0.968 |
1.1 |
1.2 |
1.2 |
0.07 |
0.08 |
0.08 |
0.1 |
-3.2 |
-3.9 |
7.3 |
7.6 |
7.6 |
6.5 |
7.8 |
12+13/V |
409 |
0.8484 |
0.953 |
1.1 |
1.2 |
1.2 |
0.08 |
0.09 |
0.09 |
0.3 |
-2.1 |
-2.7 |
8.4 |
8.7 |
8.7 |
4.8 |
6.1 |
14/VI |
263 |
0.8250 |
0.935 |
1.2 |
1.2 |
1.2 |
0.10 |
0.10 |
0.10 |
-1.9 |
-1.5 |
-2.1 |
10.1 |
10.0 |
10.1 |
0.8 |
0.6 |
15/VI |
341 |
0.7933 |
0.923 |
1.5 |
1.4 |
1.5 |
0.12 |
0.12 |
0.12 |
-1.5 |
-2.2 |
2.2 |
12.1 |
12.1 |
11.6 |
1.3 |
8.3 |
16/VI |
263 |
0.7859 |
0.923 |
1.5 |
1.7 |
1.7 |
0.11 |
0.10 |
0.10 |
-3.2 |
2.1 |
1.5 |
10.7 |
10.1 |
10.2 |
12.1 |
10.5 |
14+15+16/VI |
867 |
0.7975 |
0.927 |
1.4 |
- |
1.5 |
0.12 |
- |
0.11 |
-2.6 |
- |
0.6 |
11.2 |
- |
10.9 |
- |
7.0 |
17/VII |
267 |
0.8165 |
0.853 |
2.4 |
2.4 |
2.4 |
0.12 |
0.12 |
0.12 |
-1.3 |
0.2 |
-0.4 |
11.7 |
11.6 |
11.6 |
3.2 |
1.8 |
18/VII |
204 |
0.8528 |
0.860 |
2.2 |
2.2 |
2.3 |
0.10 |
0.11 |
0.11 |
-1.9 |
-4.9 |
-5.6 |
9.4 |
9.7 |
9.7 |
5.8 |
7.0 |
17+18/VII |
471 |
0.8318 |
0.851 |
2.4 |
2.3 |
2.4 |
0.11 |
0.11 |
0.11 |
-1.6 |
-2.0 |
-2.7 |
11.0 |
11.1 |
11.2 |
0.8 |
2.1 |
19/VIII |
333 |
0.8469 |
0.861 |
2.1 |
2.1 |
2.2 |
0.09 |
0.10 |
0.10 |
-1.3 |
-3.5 |
-4.2 |
9.2 |
9.4 |
9.4 |
4.4 |
5.7 |
20/VIII |
152 |
0.8067 |
0.756 |
2.6 |
2.7 |
2.6 |
0.11 |
0.11 |
0.11 |
-1.5 |
1.2 |
0.6 |
10.9 |
10.6 |
10.7 |
5.9 |
4.4 |
19+20/VIII |
485 |
0.8311 |
0.834 |
2.3 |
2.3 |
2.3 |
0.10 |
0.10 |
0.10 |
-1.7 |
-2.0 |
-2.7 |
10.0 |
10.0 |
10.1 |
0.6 |
2.0 |
1 height
category |
1891 |
0.8287 |
0.924 |
- |
1.9 |
1.9 |
- |
0.10 |
0.10 |
- |
-1.4 |
-2.0 |
- |
10.1 |
10.2 |
- |
- |
2 height
category |
486 |
0.7884 |
0.882 |
- |
1.7 |
1.8 |
- |
0.13 |
0.13 |
- |
-1.8 |
2.5 |
- |
12.9 |
12.4 |
- |
- |
Total (11-20) |
2377 |
0.8233 |
0.922 |
- |
- |
1.9 |
- |
- |
0.11 |
- |
|
-1.1 |
- |
- |
11.7 |
- |
- |
Total (1-20) |
4149 |
0.8339 |
0.949 |
- |
- |
1.9 |
- |
- |
0.10 |
- |
- |
-1.0 |
- |
- |
9.8 |
- |
- |
Open urban trees |
|||||||||||||||||
21/VII |
374 |
0.7917 |
0.451 |
5.5 |
- |
- |
0.15 |
- |
- |
-2.3 |
- |
- |
14.6 |
- |
- |
- |
- |
17+18+21/VII |
845 |
0.8016 |
0.857 |
4.1 |
- |
- |
0.13 |
- |
- |
-3.5 |
- |
- |
13.0 |
- |
- |
- |
- |
Total (11-21) |
2751 |
0.8097 |
0.921 |
- |
- |
2.7 |
- |
- |
0.12 |
- |
|
-2.4 |
- |
- |
11.5 |
- |
- |
Total (1-21) |
4523 |
0.8242 |
0.935 |
- |
- |
2.5 |
- |
- |
0.10 |
- |
|
-1.1 |
- |
- |
10.4 |
- |
- |
Notes. b - is the value of the
coefficient of the linear equation; h.c. – height category
The validity of combining
empirical material was confirmed by minor errors of the compared equations and
analytically - by calculating the standard deviation of the series: the degree
of difference between series was non-significant. It did not exceed 4% on
average, except for TP11 and TP16 (Table 3). However, comparing free growth
trees in urban conditions with natural and artificial stands revealed a significant
difference (14 and 105%, respectively).
Discussion
This study was carried out using mathematical-statistical analysis and descriptive interpretation of the
DBH and DST indices in tree bark of linden small-leaved. On an average, for TP,
the DST variation coefficients were 26.9 and 32.3%, and for DBH they were 27.6
and 31.8% for natural and artificial stands, respectively. The greater
variation in trunk diameters in forest cultures is explained by their rare
density and, accordingly, less pronounced differentiation processes.
The results of the regression
analysis showed a linear relationship between these metrics, which can be
applied to assess small trees. Some authors reported that simple linear models,
taking into account the fit statistics, are most suitable for modeling the DBH
– DST relationship (Özçelík et al. 2010; García-Cuevas et al. 2017). It
was found that the coefficients of all models in age classes and generalized
samples were quite closer to each other and explained the change in DBH by
70–97% for natural and artificial stands. Kulla et al. (2017) in their species-specific
models showed an overall DBH variance of 95% for European beech, 96% for Norway
spruce and 97% for Silver fir and Scots pine. Ercanli et al.
(2015) used mixed-effects models to predict DBH according to DST Fagus Orientalis
Lipsky with a coefficient of determination of 0.99. Diéguez-Aranda
et al. (2003) developed a linear
model for the Eucalyptus globulus Labill and Betula
alba L., explaining 92 and 81%, respectively, of the total variance of DBH.
Extrapolation of large stump diameters of linden trees in free growth
conditions should be approached cautiously, as their forecasts were more
volatile (R2=45%). The linear models created in our study allow the DBH to be estimated
with standard error values ranging from 0.9 to 2.9 cm for trees growing in the
forest and 5.5 cm for free growing trees. Errors Qp
do not exceed 6%, and Qs in most cases is below 10%. This indicates
that the models with the R2 coefficient of more than 0.7 give quite
satisfactory estimates for the corresponding age group and the origin of the
stands.
Table 4: The results of the F test, which
determine the differences in age classes, describing the relationship DBH = b×DST
TP numbers |
Age classes |
n |
SSER |
SSEF |
F |
1+2 |
III+IV |
304 |
377 |
377 |
0.1 |
3+4 |
VIII |
412 |
1782 |
1768 |
3.2 |
1+3+4 |
III+VIII |
588 |
1988 |
1974 |
4.1 |
1+5 |
III+IX |
411 |
1046 |
1020 |
10.3 |
6+7 |
X |
285 |
2234 |
2149 |
11.2 |
6+8 |
X |
249 |
1368 |
1368 |
0.1 |
6+9 |
X |
257 |
2032 |
2004 |
3.4 |
7+8 |
X |
318 |
2169 |
2066 |
15.9 |
7+9 |
X |
326 |
2721 |
2004 |
115.8 |
6+7+8+9 |
X |
575 |
4210 |
4070 |
19.7 |
1+6+7+8+9 |
III+X |
751 |
4433 |
4275 |
27.6 |
1+10 |
III+XI |
322 |
688 |
611 |
40.2 |
1+3+4+5+6+10 |
III+VIII+IX+Х +XI |
1077 |
4061 |
3715 |
100.2 |
2+3+4 |
IV+VIII |
640 |
1954 |
1939 |
4.8 |
2+5 |
IV+IX |
463 |
1004 |
986 |
8.5 |
2+6+7+8+9 |
IV+Х |
803 |
4392 |
4241 |
28.6 |
2+7+8+9 |
IV+Х |
695 |
3627 |
3515 |
21.9 |
2+10 |
IV+XI |
374 |
638 |
577 |
39.1 |
3+4+5 |
VIII+IХ |
647 |
2709 |
2583 |
31.4 |
3+4+6+7+8+9 |
VIII+Х |
987 |
6121 |
5838 |
47.8 |
3+4+10 |
VIII+ХI |
558 |
2438 |
2174 |
67.5 |
5+6+7+8+9 |
IX+Х |
810 |
5030 |
4885 |
24.1 |
5+10 |
IX+ХI |
381 |
1262 |
1221 |
12.8 |
6+7+8+9+10 |
Х+ХI |
721 |
4705 |
4643 |
9.6 |
Total (1-10) |
III+IV+VIII+IX+Х+XI |
1772 |
4062 |
3715 |
165.6 |
12+13 |
V |
409 |
526 |
511 |
12.1 |
11+12+13 |
IV+V |
554 |
1353 |
1172 |
85.3 |
14+15 |
VI |
604 |
1110 |
1077 |
18.5 |
14+16 |
VI |
526 |
1034 |
984 |
26.7 |
15+16 |
VI |
604 |
1345 |
1342 |
1.1 |
14+15+16 |
VI |
867 |
1754 |
1702 |
26.8 |
11+14+15+16 |
IV+VI |
1012 |
2430 |
2363 |
28.5 |
11+15 |
IV+VI |
486 |
1385 |
1379 |
2.3 |
12+13+14+15+16 |
V+VI |
1276 |
2488 |
2212 |
158.6 |
12+13+14+16+17+18+19+20 |
V+VI+VII +VIII |
1891 |
6951 |
6493 |
133.1 |
17+18 |
VII |
471 |
2607 |
2526 |
15.1 |
11+17+18 |
IV+VII |
616 |
3381 |
3187 |
37.3 |
12+13+17+18 |
V+VII |
880 |
3156 |
3036 |
34.4 |
14+15+16+17+18 |
VI+VII |
1338 |
4504 |
4232 |
85.9 |
19+20 |
VIII |
485 |
2617 |
2473 |
28.1 |
11+19+20 |
IV+VIII |
630 |
3394 |
3134 |
52.0 |
12+13+19+20 |
V+VIII |
894 |
3171 |
2984 |
55.9 |
14+15+16+19+20 |
VI+VII |
1352 |
4533 |
4179 |
114.3 |
17+18+19+20 |
VII+VIII |
956 |
5224 |
4999 |
43.0 |
Total (11-20) |
IV–VIII |
2377 |
8533 |
7877 |
197.7 |
12+13 |
V |
409 |
526 |
511 |
12.1 |
11+12+13 |
IV+V |
554 |
1353 |
1172 |
85.3 |
2+11 |
IV+IV |
373 |
876 |
832 |
19.4 |
3+4+19+20 |
VIII+VIII |
897 |
4416 |
4241 |
36.8 |
Total (1-20) |
III–XI |
4149 |
16634 |
15308 |
359.3 |
17+18+21 |
VII |
845 |
14235 |
13847 |
23.6 |
Total (11-21) |
IV–VIII |
2751 |
20294 |
19193 |
157.6 |
Total (1-21) |
III–XI |
4523 |
29014 |
26629 |
404.9 |
The R2 variability is
inversely correlated with the sum of the cross-sectional areas of stands and is
–0.63 (p = 0.053) for natural lime forests and –0.60 (p = 0.068) for artificial
ones. It is due to the trees' conicity, depending on the density of the stands
and the trees cenotic position. The DBH/DST ratio
varies from one stand to another because the conditions and structures of the
stand are site-specific, i.e., in terms of stand density, site and soil
properties, with significant variability (Ercanli et al. 2015). It is also consistent with
Milios et al.
(2016), who indicated that higher tree conicity results from low forest
density. Free growing trees or large dominant trees have more wood increment at
the base, while oppressed trees or trees growing in stands with high
completeness, without being dominant, have a smaller trunk thickness.
A significant correlation was
noted between the age and the coefficient b
of the linear equation for natural lime forests (r = 0.685; p = 0.029). This
also indicates an increase in the taper of tree trunks with the age of the
stands. However, for artificial stands such a relationship is not found (r =
0.076; p = 0.835). Apparently in forest crops with a regular planting step and
row spacing the taper is not pronounced.
In our study, the stump height
was not considered a predictive variable; it was assumed that all trees at the
cut height have the correct trunk; that is, they are not strongly deformed due
to root butt swell. There are conflicting reports in the literature on this
subject. Pond et al. argue that in
cases where the stump height is not included in the model, and there is high
variability in the stump height, the model's predictive power is low (Pond and
Froese 2014). Research by Diéguez-Aranda et al. (2003) showed that the stump
height did not significantly improve the forecasts for the five studied
species. Only in the case of features at the base of the birch trunk was it
advisable to consider this variable.
Comparison of DBH/DST ratios between TP, age classes, altitude
categories based on the whole, and reduced models found that, with a few
exceptions, there was a significant difference between them. Therefore, each
age class should be represented by separate regression equations, and caution must
be exercised when applying the obtained regression equations to estimate
DBH/DST ratios in natural and artificial stands, especially free-growing trees.
The results may not be applicable due to differences in origin, density and
habitat.
Even though the results of the F tests revealed significant differences
between the equations for assessing DBH by age classes and the origin of
stands, the resulting generalized model DBH = 0.8339 DST (R2 = 0.95,
which is relatively higher) was used for comparison with the data of the
forests of Urals (VNIITslesresurs 1991). It is the
only regulatory document in Russia on the translation of DBH from DST for
small-leaved linden. Verification of theoretical data of generalized model with
the reference book data showed an overestimation of the results of entire range
of diameters for the latter. At the same time, the maximum differences were
observed in the group of small-sized trunks, not exceeding 5%, and for others,
<2.5%. In absolute terms, for the thickest trunks (56 cm), the difference
did not exceed 1 cm, while the predicted DBHs remained identical in thickness
steps for the entire standard range.
Conclusion
Analysis of simple linear equations without a free parameter indicated
their high adequacy, indicating no need to use complex models. When evaluating
models between age classes and natural and artificial stands, trees of free
growth, the presence of a statistically significant difference according to the
F-criterion was revealed. Despite this, the generalized equation DBH =
0.8339•DST explained 95% variability of DBH versus DST. The values of errors
for the combined material varied within acceptable limits and indicated a
similar pattern of DBH changes for I and II categories of heights. Verification
of the data obtained with the standards developed in the 1980s showed that
their accuracy was acceptable for forest managers when assessing the volume of
illegally removed trees in local conditions. However, for an accurate
assessment, a differentiated assessment is necessary, considering the origin,
age and conditions of habitats. Lack of previous work on DBH and DST allometry
on this species indicates that this is actually the first study of its kind.
Acknowledgements
This
research did not receive any specific grant from funding agencies in the
public, commercial, or not-for-profit sectors.
Author Contributions
AG devised and supervised the project, formulated the
main conceptual ideas and proof outline, and performed the computations. AG, IM
and IS established the test plots and planted the trees. LB and RB wrote the
manuscript and revised it after peer review. All authors discussed the results
and contributed to the final manuscript
Conflict of Interest
The
authors declare that they have no conflicts of interest.
Data Availability
Data will be available on a reasonable request.
Ethics Approval
The authors declare that the work is written with due consideration of
ethical standards. The study was conducted in accordance with the ethical
principles approved by the Ethics Committee of Federal State Budgetary
Educational Establishment of Higher Education “Bashkir State Agrarian
University” (Protocol № 6 of 13.06.2020).
References
Abdullaeva NA,
IKH Khairov (2019). To the question of studying the
patterns of transition from the diameter of the tree stump to the diameter at
the chest's height. In: National Conference on the Results of Scientific and
Industrial Work of Teachers and Students in the Field of Landscape Architecture
and Forestry, pp:3–7. Amirit, Saratov, Russian
Federation
Asrat Z, T
Eid, T Gobakken, M Negash
(2020). Modelling and quantifying tree biometric properties of dry Afromontane
forests of south-central Ethiopia. Trees 34:1411–1426
Bruchwald A
(2001). Tables for Determining the Breast Height and Thickness of the Tree and
the Base of the Trunk Diameter, Advertising and Publishing Agency A. Grzegorczyk, Warsaw, Poland
Chukwu
O, FN Ogana, JU Nwatu
(2020). Comparison of volumes estimated from breast height and stump diameters:
Application to Tectona Grandis data. For Sci 66:551–555
Corral-Rivas
JJ, M Barrio-Anta, OA Aguirre-Calderón, U Dieguez-Aranda
(2007). Use of stump diameter to estimate diameter at breast height and tree
volume for major pine species in El Salto, Durango (Mexico). Forestry 80:29–40
Costa E,
M Siqueira, G Marangon, M
Teixeira Martins, C Schons (2019). Equações para variáveis dendrométricas em Árvores singulares de Pinus taeda L.
Agric Acad 6:155–161
Di
Cosmo L, P Gasparini (2020). Predicting diameter at breast height from stump
measurements of removed trees to estimate cuttings, illegal loggings and
natural disturbances. SEEFOR 11:41–49
Diéguez-Aranda
U, FC Dorado, MB Anta, MB Murias (2003). Estimación del diámetro normal y del volumen del tronco a partir de las dimensiones
del tocón para seis especies forestales
comerciales de Galicia. For Syst 12:131–139
Ercanli I,
A Gunlu, EZ Başkent
(2015). Mixed effect models for predicting breast height diameter from stump
diameter of Oriental beech in Göldağ. Sci Agric 72:245–251
García-Cuevas
X, J Hernández-Ramos, JJ García-Magaña, A
Hernández-Ramos, V Herrera-Ávila, A González-Peralta, EJ Garfias-Mota
(2017). Predicción de diámetro normal, altura y volumen de Abies religiosa
a partir del diámetro del tocón.
Madera Bosques 23:61–70
Kulla L, M Bošeľa, V Murgaš, J Roessiger, V Šebeň (2017).
Reconstructing past forest status using inventory and tree ring data to suTPort uneven-aged forest management. Cent Eur For J 63:203–211
Martynova M, R Sultanova, G Odintsov, R Sazgutdinova, E Khanova (2020). Growth of Tilia cordata Mill. in urban forests. SEEFOR 11:51–59
Milios E, KG
Kitikidou, V Dalakouras, E Pipinis (2016). Diameter at breast height estimated from
stumps in Quercus frainetto
in the region of Evros in Northeastern Greece. Cerne 22:337–344
Mugasha W
(2021). Stump diameter: A potential predictor of removed biomass through tree
cutting. Tanz J For Nat Conserv
90:96–109
Özçelík R, JR
Brooks, MJ Diamantopoulou, HV Wiant
Jr (2010). Estimating breast height diameter and volume from stump diameter for
three economically important species in Turkey. Scand
J For Res 25:32–45
Özdemir E, A Şahin, GA Özdemir (2020).
Determination of relationship stump diameter – diameter at breast height for
sessile oak (Quercus petraea (Matt.) Liebl) stands in Northern Marmara Region. Turk J For
21:25–34
Paramonov AA,
SV Tretyakov, SV Koptev, SV
Yaroslavtsev, IV Tsvetkov
(2020). Research of the trunk lower part form of the tree-like willow in the North-Taiga
area and development of transition tables from the stump diameter to the
diameter at breast height. In: IOP
Conference Series: Earth and Environmental Science, Vol. 574, No. 1, p:012060. IOP Publishing, Bristol, England
Pond NC, RE Froese (2014).
Evaluating published approaches for modelling diameter at breast height from
stump dimensions. Intl J For Res 87:683–696
Sağlam F, OE
Sakıcı, M Seki (2016) Stump diameter and
diameter at breast height relationship for Chestnut (Csatanea sativa Mill.) stands in Kastamonu
coastal region. In: International Forestry Symposium (IFS 2016), pp:7–10
Şahin A, A Kahriman, A Göktürk (2019).
Estimating diameter at breast height (DBH) from diameter at stump height (DST)
in triple mixed stands in the region of Artvin in
Turkey. Forestist 69:61–67
Sakici OE, G
Özdemir (2017). Stump diameter and diameter at breast
height relationships for oriental beech and Kazdağı
fir at mixed stands in Karabük. Kastamonu
univ. Orman Fak Derg 17:298–306
Sakici OE, G
Ozdemir (2018). Stem taper estimations with
artificial neural networks for mixed Oriental beech and Kazdaği
fir stands in Karabük region, Turkey. Cerne 24:439–451
Sasanifar S, M Namiranian, N Zargham (2015).
Determining mathematical relations among important characteristics of Fraxinus
excelsior (Case study: Research-Training Forest of Kheyroud).
ERR 3:47–57
Şenyurt M,
İ Ercanlı, A Günlü,
F Bolat, S Bulut (2020).
Artificial neural network models for predicting relationships between diameter
at breast height and stump diameter: Crimean pine stands at ÇAKÜ Forest. BOSQUE 41:25–34
Sultanova R, M Martynova, S Konashova, E Khanova, V Yanbaeva (2020).
Cutting practices in mature stands of Tilia cordata
Mill. Cent Eur For J 66:151–158
Verkhunov PM,
VL Chernykh (2007). Forest Inventory: Textbook. PSTU,
Volga State Technological University, Yoshkar-Ola, Russia
VNIITslesresurs
(1991). Forest Inventory Reference Book for the Urals' Forests (Normative
Materials for the Perm, Chelyabinsk, Sverdlovsk, Kurgan Regions and the Bashkir
Autonomous Soviet Socialist Republic). VNIITslesresurs,
Moscow, Russia
Weiss
AA (2013). Standards for the restoration of aspen plantations (Populus tremula)
in the conditions of Central Siberia. Bull BSAU 4:109–111
Westfall JA (2010). New models for predicting diameter at breast height
from stump dimensions. North J Appl For 27:21–27