Keywords

6.1 Introduction

NFI estimators of totals and densities (values per surface unit) described in Chap. 5 require that the value of the quantitative variables measured in the sample plots are expressed with reference to the square plane surface of 1 km2. The total value of a variable in the sample plot (e.g., the number or volume of the trees) must be divided by the surface of the sample plot and multiplied by the surface SQ (1 km2) of the square containing the sample plot. As the variables are measured in sample plots with three different sizes (as described in Chap. 3), the expansion factors are 106/12.566371, for the variables measured in each of the AdS2, 106/50.265482, for the variables measured in the AdS4, and 106/530.929159 for those measured in the AdS13. For example, each tree with 4.5 ≤ DBH < 9.5 cm measured in the AdS4 represents 1 × 106/50.265482 per square kilometre. In the same way, the value of each variable of that tree (e.g., its basal area, volume or biomass) must be multiplied by 106/50.265482 to get the amount per square kilometre.

Among the quantities surveyed, the number of trees and the basal area of the stands are easily computed. The same is also true for the carbon stocks either in the living ecosystem components (trees, small trees and shrubs, removed trees) or the dead ones (dead trees, deadwood lying on the ground, stumps) after their biomass has been estimated, because conversion was obtained assuming 50% carbon content of woody biomass (IPCC, 2006; Matthews, 1993; Woodall et al., 2012). For this reason, the descriptions of methods below will be limited to the biomass computation steps. Timber volume, tree biomass, annual volume increment and removals required more complex preliminary processing. They entailed construction of models that used data widely representative of the conditions across the entire country so that they might be used beyond NFI computations.

This chapter describes the data analysis and the models used for estimating the quantitative variables measured in the sample plots by INFC2015. Some of the models described were developed by previous or connected research projects, in view of the INFC needs. The conversion factors needed to estimate the biomass of deadwood, saplings and shrubs were obtained through an additional field campaign of the second Italian NFI (INFC2005) and the following laboratory analyses.

6.2 Tree Height

Total tree height (h) is an explanatory variable used in the model to estimate the volume and the biomass of trees. Due to the high costs, NFIs generally limit the tree height measurement on sample trees that are a subsample of those callipered (Gschwantner et al., 2016), which involves estimating the total tree height through models. INFC2015 increased the sample trees per plot to ten, as compared to the past assessment, which ranged from five to ten (Bosela et al., 2016; Gasparini & Di Cosmo, 2016; Gasparini et al., 2010). For the sample trees, five were randomly selected, and five were selected based on representative criteria; the collected data were used to build height-diameter models. The equations described were obtained re-calibrating the model developed in INFC2005; the new calibration allowed including the wider range of growing conditions revealed by the data we collected.

The model adopted draws from the Chapman-Richards function (Liu & Li, 2003; Ratkowsky, 1990) and is shown in Eq. (6.1):

$$\hat{h}_{ij} = \left( {a_1 + a_2 \cdot H_{domj} } \right)\left[ {1 - \exp \left( {\left( {a_3 + a_4 \cdot H_{domj} } \right)d_{ij} } \right)} \right]^{\left( {a_5 + a_6 H_{domj} } \right)}$$
(6.1)

where:

ĥij is the total tree height estimated of tree i of plot j, in m;

dij is the DBH of tree i of plot j, in cm;

Hdomj is the dominant height (Hdom) of plot j, in m, obtained as the arithmetical mean of the total heights of the three widest trees in DBH.

We graphically evaluated the affinity of the height-diameter relationship for species with small samples, to create groups with higher number of data and obtain more stable estimates of the numerical coefficients. Table 6.1 shows the 52 equations obtained and their relative estimated coefficients. They concern the main species or species groups of the Italian forests and allowed for an estimation of the total tree-heigh to almost all of the undamaged trees. For the species not provided with an equation, we evaluated which could be best used through graphical and analytical analyses. A limited number of sample-plots could not be provided with the Hdom. For the species included, we developed models based solely on the DBH as the explanatory variable. Finally, for trees of species callipered only a few times but provided with a relevant percentage of tree height data, we adopted by convention the mean measured heights, by DBH class in case of visible trends.

Table 6.1 Tree height model: regression coefficients for Eq. (6.1) / Modello delle altezze: coefficienti di regressione per l’Eq. (6.1)

Table 6.2 shows the number of observations by species or group of species used for models’ calibrations and the range of values for DBH and Hdom. This information is useful in establishing awareness beyond the INFC computations.

Table 6.2 Number of observations and extreme values of diameter at breast height (DBH) and dominant height (Hdom) used for calibrating the height-diameter equations, by species / Numero di osservazioni e valori estremi di diametro (d1.30) e altezza dominante (Hdom) utilizzate per la calibrazione delle equazioni di previsione dell’altezza totale degli alberi, per specie

6.3 Growing Stock and Biomass

6.3.1 Living Trees

Distinct procedures were adopted to estimate the volume and the biomass of unbroken or broken (truncated) trees. The volume and biomass of unbroken trees were estimated through double entry equations available in Tabacchi et al., (2011a, b). Those equations were built before and in view of the second Italian NFI (INFC2005) and partially updated afterwards. The data were collected from tree samples across wide areas of the country. In addition to being measured, the trees were cut, and samples of wood were analysed in the laboratory to determine the dry weight. At present, there are prediction equations available for 26 species or group of species; a set of 5 equations is given for each species/group whose form is shown in Eq. (6.2):

$$\hat{y}_i = b_1 + b_2 \cdot DBH_i^2 \cdot h_i + b_3 \cdot DBH_i$$
(6.2)

where:

ŷi is the volume (v in dm3) (stem plus large branches and treetop up to 5 cm cross section diameter), the biomass (in kg) for the same component (w1), the biomass of the small branches and the treetop (w2), the stump biomass (w3) or the whole aboveground biomass (w4) of tree i;

DBHi is the stem diameter at 1.30 m from the ground level of tree i, in cm;

hi is the total tree height of tree i, in m.

w4 can also be obtained by the sum of its three components (w1, w2 and w3) as the model assures additivity. Table 6.3 shows the coefficients for estimating the volume of the species or groups of species.

Table 6.3 Tree volume functions: regression coefficients for Eq. (6.2) / Funzioni di cubatura: coefficienti per l’Eq. (6.2)

Table 6.4 shows the coefficients for estimating the four biomass components.

Table 6.4 Tree biomass functions: regression coefficients for Eq. (6.2) / Funzioni di stima della fitomassa: coefficienti per l’Eq. (6.2)

For the species not provided with a function, we adopted one of the functions available, i.e., the function that was thought to be the most suitable, based on the morphologic affinity of the species, or we adopted the prediction equation for the heterogeneous group of the other broadleaved species. In the inventory survey, the size interval of the measured DBH was for some species wider than the set of sample trees used to develop the prediction equations. In a small number of trees with DBH near the calipering threshold, volume estimates showed negative values. Their volume was estimated using two equations, one for conifers and the other for broadleaves, built with a subset of the data previously used by Tabacchi et al., (2011a, b) for calibrating equations, i.e., the trees with DBH < 13 cm. The two equations, available in Tomter et al. (2012), are as follows (Eqs. 6.3 and 6.4):

$$\hat{y}_i = 1.2849 + 3.9579 \cdot 10^{ - 2} \cdot DBH_i^2 \, h_i \quad \quad ({\text{conifers}})$$
(6.3)
$$\hat{y}_i = 0.5997 + 3.9619 \cdot 10^{ - 2} \cdot DBH_i^2 \, h_i \quad \quad ({\text{broadleaves}})$$
(6.4)

where:

ŷi is the volume (v in dm3) (stem plus large branches and treetop up to 5 cm cross section diameter) of tree i;

DBHi is the stem diameter at 1.30 m from the ground level of tree i, in cm;

hi is the total tree height of tree i, in m.

For a small number of trees with DBH near the calipering threshold, the prediction equation for one or two aboveground biomass components returned negative values. These values were conventionally set to zero, with a consequent underestimation of the contribution by these trees to the total biomass of the sample unit. When all three aboveground tree components returned negative values, the biomass of the stem and large branches was estimated by multiplying the previously estimated volume of this component by the basal density value (dry weight per unit fresh volume). The basal density was found for the 26 species or group using the data of all the unbroken trees in the dataset for which no null or negative biomass estimates had occurred. Finally, the value of the other two components was conventionally set to zero. Last, to avoid estimates from extrapolations far beyond the observation limit of the maximum DBH of the sample trees used for calibrating the prediction equations, the volume of excess size trees was prudently estimated assuming a conventional maximum DBH.

The volume of truncated trees was estimated as half the cylinder volume obtained by multiplying the basal area and the tree height at breakage. The biomass of each tree was obtained multiplying its volume by the basal density of the species or group calculated as explained above.

After estimating the volume or the biomass of all the trees growing in a sample-plot, the overall values in each, by surface unit, was obtained by summing up the tree values per km2, using the expansion factors indicated in Sect. 6.1.

6.3.2 Standing Dead Trees

The volume of dead, still standing, and unbroken trees was estimated by the same procedure described for the unbroken living trees. In the same way, the volume of dead broken trees was estimated adopting the cylinder as the reference shape and using the DBH and the tree height measured in the field. The volume was then converted to biomass using the conversion factors shown in Table 6.5. Those were determined in an addressed additional field survey campaign of INFC2005, in 2008–2009, by measuring and analysing woody samples either in the field or in the laboratory (Gasparini et al., 2013). The conversion factors were obtained by category (standing dead trees, deadwood lying on the ground and stumps), group of species (conifers and broadleaves) and decay class (Di Cosmo et al., 2013).

Table 6.5 Basic density (Mg m−3) of deadwood biomass components (standing dead trees—SDT, stump, deadwood lying on the ground—DWL) by group of species (conifers and broadleaves) and class of decay (from Di Cosmo et al., 2013, modified) / Valori di densità basale (Mg m−3) del legno morto (alberi morti in piedi—SDT, ceppaie—Stump, legno morto grosso a terra—DWL) per classi di decadimento e gruppo di specie (da Di Cosmo et al., 2013, modificato)

As done for the living trees, after estimating the volume or the biomass of all the trees in a sample plot, the overall values in each by surface unit was obtained by summing up the tree values per km2, using the expansion factors indicated in Sect. 6.1.

6.4 Average Annual Growth

The estimate of the annual volume increment of trees was based on the increments measured on the cores taken from the sample trees in the sample plots. More specifically, the cores allowed an estimation of the annual increment of stem radius as the average of the five outermost ring width (not including the one from the survey growing season).

The estimation procedure went through four computational steps for obtaining: (a) the percent volume increment of any sample tree; (b) the plot level volume mean percent increment; (c) the volume increment of any callipered tree; (d) the overall volume increment of all the trees callipered in the plot.

(a) The percent annual volume increment of a sample tree was obtained by Eq. (6.5) (Hellrigl, 1969, 1986):

$$pv_{zj} = 100\left( {2\Delta d_{zj} /d^{\prime}_{zj} } \right) + \left( {\Delta h_{zj} /h^{\prime}_{zj} } \right)$$
(6.5)

where:

Δdzj is the annual DBH increment of sample tree z of plot j;

dzj is the DBH of sample tree z of sample plot j during the field survey;

d’zj is the DBH of sample tree z of sample plot j one year prior to the field survey;

Δhzj is the annual total tree height increment of sample tree z of plot j;

hzj is the total tree height of sample tree z of plot j during the field survey;

h’zj is the total tree height of sample tree z of plot j one year prior to the field survey.

Δhzj was calculated using differences between the height in the measurement year and that corresponding to the tree DBH one year before the survey, both estimated using the height-diameter model described in Sect. 6.2.

(b) The mean percent annual volume increment for the entire sample plot was calculated as the average of the percent annual increment values of all the sample trees, each weighted with the volume of the corresponding tree, as shown in Eq. (6.6):

$$pv_j = \mathop \sum \limits_{z = 1}^m \left( {pv_{zj} \cdot Vol_{zj} } \right)/\mathop \sum \limits_{z = 1}^m Vol_{zj}$$
(6.6)

where:

Pvj is the average percent annual volume increment for plot j;

m is the number of sample trees of plot j;

pvzj is the percent annual volume increment of sample tree z of plot j;

Volzj is the volume of sample tree z of plot j.

The volume of the sample trees was estimated using the volume functions of Eq. (6.2) while the total tree height was known because recorded in the field.

(c) The annual volume increment for any callipered tree in the plot was obtained by multiplying its volume by the average weighted percent increment, as shown in Eq. (6.7):

$$\Delta v_{ij} = Vol_{ij} \cdot pv_j$$
(6.7)

where:

Δvij is the annual volume increment of tree i of plot j;

Volij is the volume of tree i of plot j;

pvj is the average percent annual volume increment for plot j.

(d) The sample plot annual volume increment was obtained by summing the annual increment values of all n trees in the plot, as shown in Eq. (6.8)

$$\Delta V_j = \mathop \sum \limits_{i = 1}^n v_{ij} \cdot f_i$$
(6.8)

where:

ΔVj is the annual volume increment for plot j;

vij is the annual volume increment of tree i of plot j;

fij is the expansion factor of tree i, that varies with the DBH threshold value of 9.5 cm.

In a limited number of sample plots, with few cores available, an alternative procedure was used. This unfavourable condition was limited in INFC2015 by the choice to sample a fixed number of ten trees in each sample plot. The annual volume increment in plots with less than four cores available (278 in total) was estimated using a model developed by Gasparini et al., (2017) from the data of INFC2005. The model predicts the five-year periodic volume increment according to Eq. (6.9):

$$\ln \left( {PAI} \right) = a_0 + b_1 \ln \left( {GSV} \right) + b_2 \ln (N)$$
(6.9)

where:

ln(PAI) is the natural logarithm of the periodic increment (PAI, period = 5 years) (m3 ha−1 5 years−1);

ln(GSV) is the is the natural logarithm of the growing stock volume (GSV, m3 ha−1);

ln(N) is the natural logarithm of the number of trees per hectare.

Coefficient a0 is dependent on forest type, as shown in Table 6.6. The antilogarithm of ln(PAI) so estimated must be multiplied by a correction factor equal to 1.125, before being divided by 5, to get the mean annual increment. This prevents bias in log-transformed allometric equations (Sprugel, 1983).

Table 6.6 Volume increment stand-level model: regression coefficients for Eq. (6.9) (from Gasparini et al., 2017, modified) / Modello per l’incremento periodico di volume: coefficienti di regressione per l’Eq. (6.9) (da Gasparini et al., 2017, modificato)

Conversion and expansion of volume increment to the total of aboveground biomass annual production was obtained for each tree in the sample plot, multiplying its volume increment by the ratio w4/v, computed for the same tree.

6.5 Volume and Biomass of Stumps and Lying Deadwood

6.5.1 Stumps

In order to estimate the volume of the stumps, they were considered like cylinders. The volume of each stump was easily calculated by the mean diameter of the cut section and the mean height. Conversion of stump volume to biomass was obtained by the basal density factors shown in Table 6.5.

The overall stump volume or biomass for the entire sample unit was obtained by summing up the volumes of all the stumps in the sample plot and referring to the surface unit according to the expansion factor for AdS13.

6.5.2 Deadwood Lying on the Ground

The volume of the deadwood lying on the ground (which includes suspended deadwood) was estimated by assimilating each fragment to a truncated cone. The variables required by the formula (the diameters of the end sections and length of the fragment) were known from the field survey. The biomass was then calculated through the conversion factors showed in Table 6.5, which were based on the decay class as recorded by the surveyors.

The overall volume of deadwood lying on the ground for a sample plot was obtained by summing the volumes of all the fragments and, similarly to earlier descriptions, the value obtained was referred to the surface unit according to the expansion factor for AdS13.

Deadwood lying on the ground is one of the three components of deadwood surveyed by INFC, along with the standing dead trees and stumps. Together, these are referred to as coarse woody debris (e.g., Russell et al., 2015), although they have been defined in many ways and differences exist, especially in the size required (e.g., Enrong et al., 2006). The overall coarse woody debris volume or biomass was estimated at plot level following the usual procedure of summing up the individual values, previously referred to the surface unit according to the appropriate expansion factors.

6.6 Small Trees and Shrubs

Recording the species of the woody subjects under the minimum callipering threshold (4.5 cm) allowed estimating the variables of interest explicitly for small trees (seedlings and saplings) and shrubs in the survey. For both, the total number, the biomass, and the carbon stock were estimated, by size class. Biomass was estimated by multiplying the number of subjects recorded in the field by the unit dry weight of the correspondent category (small tree or shrub) and size class (three size classes), as shown in Table 6.7. The biomass values in Table 6.7 were obtained after the integrative survey of the second Italian NFI (INFC2005).

Table 6.7 Biomass (dry weight—kg) of small woody individuals (trees and shrubs under the callipering DBH threshold value) in the three size classes (from Di Cosmo & Gasparini, 2013) / Fitomassa (peso secco unitario—kg) degli individui di rinnovazione e arbusti nelle tre classi dimensionali adottate nel rilievo inventariale (da Di Cosmo & Gasparini, 2013)

As usual, the values per unit surface were obtained using the appropriate expansion factor, as discussed in Sect. 6.1, and the total per plot by summing up the values of all measured items.

6.7 Removed Trees

Tree volume and biomass removed in the year preceding the field survey were estimated based on the data recorded for the stumps with cutting section diameter ≥ 9.5 cm, within the sample plot AdS13. The procedure for estimating the volume and the biomass of each tree is identical with that adopted for unbroken trees, both living and standing dead, previously described in this chapter. As it is based on the knowledge of DBH, the procedure started by reconstructing this value, based on the stump diameter and height data. For this purpose, we used the prediction equations developed by Di Cosmo & Gasparini (2020). Those equations allow to estime the DBH for 16 tree species and the broader groups of conifers and broadleaves; their form is as showed in Eq. (6.10):

$$\widehat{DBH} = b_0 + b_1 \cdot D_{stump} + b_2 \cdot D_{stump}^2 + b_3 \cdot H_{stump}$$
(6.10)

where:

\(\widehat{DBH}\) is the diameter at 1.30 m above the ground line, in cm;

Dstump is the diameter of the stump cut section, in cm;

Hstump is the height of the stump, in m.

Table 6.8 shows the estimated parameters needed for applying Eq. (6.10).

Table 6.8 Coefficients of Eq. (6.8) for estimating the diameter at breast height (DBH) of removed trees by stump diameter and height (from Di Cosmo & Gasparini, 2020, modified) / Coefficienti dell’Eq. (6.8) per la stima del diametro degli alberi a 1.30 m da terra attraverso i valori di diametro e altezza della ceppaia (da Di Cosmo & Gasparini, 2020, modificato)

As usual, the volume estimated (v) is that of the stem plus large branches. The biomass removed is that of the same component (w1) plus that of the small branches and treetop (w2).

The volume or biomass removed from the plot was obtained by summing up all the tree level values.