Assessing urban tree carbon storage and sequestration in Bolzano, Italy

Abstract

Recent climate change, environmental design, and ecological conservation policies require new and existing urban developments to mitigate and offset carbon dioxide emissions and for cities to become carbon neutral. Some North American models and tools are available and can be used to quantify the carbon offset function of urban trees. But, little information on urban tree carbon storage and sequestration exists from the European Southern Alps. Also, the use of these North American models in Europe has never been assessed. This study developed a protocol to quantify aboveground carbon (C) storage and sequestration using a subsample of urban trees in Bolzano, Italy, and assessed two existing and available C estimation models. Carbon storage and sequestration were estimated using city-specific dendrometrics and allometric biomass equations primarily from Europe and two other United States models; the UFORE (Urban Forest Effects Model) and the CUFR Tree Carbon Calculator (CTCC). The UFORE model carbon storage estimates were the lowest while the CUFR Tree Carbon Calculator (CTCC) C sequestration estimates were the highest. Results from this study can be used to plan, design, and manage urban forests in northern Italy to maximize C offset potential, provide ecosystem services, and for developing carbon neutral policies. Findings can also be used to predict greenhouse gas emissions from tree maintenance operations as well as estimating green waste yield from landscape maintenance activities and its use as biofuel and compost. Managers need to be aware that available models and methods can produce statistically different C storage and sequestration estimates.

Keywords:

• i-Tree Eco
• growth rates
• allometric equations
• ecosystem services
• CTCC model

Introduction

Climate change is one of the most important and pressing environmental, economic, and security issues our world faces today (Barnett 2003; IPCC 2007; Karagiannis & Soldatos 2010). Urban areas are steadily growing throughout the world (Grimm et al. 2008) and by 2030 it is expected that 60% of the world’s population will be living in cities (Rydin et al. 2012). Thus, as urban environments become more important as living space for humans, they are an increasing source of carbon emissions.

Several studies in North America, China, and Australia (Brack 2002; Nowak & Crane 2002; Zhao et al. 2010; Dobbs et al. 2011; Martin et al. 2012; Roy et al. 2012), and more recently in the United Kingdom and Germany (Davies et al. 2011; Strohbach & Haase 2012; Strohbach et al. 2012), have shown that trees in urban environments remove carbon dioxide from the atmosphere through growth and photosynthesis, and store excess carbon as biomass in roots, stems, and branches. Indirectly, urban trees also reduce building energy used for cooling through their shade and climate amelioration effects, thereby reducing CO₂ emissions from decreased energy production (Akbari et al. 2001).

The estimation of tree carbon sequestration depends on species types and their mortality and growth characteristics as well as their overall condition (Escobedo et al. 2010; Lawrence et al. 2012). Urban tree mortality can be influenced by site and tree characteristics such as land use, natural disturbance (e.g. pests, fire, and drought), human activities and urbanization effects (Iakovoglou et al. 2002; Nowak et al. 2004; Lawrence et al. 2012), and stewardship (Sklar & Ames 1985). Similarly, tree growth is influenced by genetics, climate, soil, moisture, light, and competition (Peper & McPherson 1998; Bühler et al. 2007). These effects on tree growth and mortality are well-known in European forests, but the majority of studies of urban trees growth rates have been conducted in the United States of America (USA) (Jo & McPherson 1995; Iakovoglou et al. 2002; Lawrence et al. 2012; Roy et al. 2012). We know only a few studies on urban tree mortality and growth in northern Italy (Sanesi et al. 2007; Semenzato et al. 2011; Marziliano et al. 2013), but no studies providing information specific to urban tree growth rates in South Tyrol, Italy, are known to us.

Several European cities have begun to formulate CO₂ mitigation policies and this is exemplified by the city of Bolzano, Italy, which decided to become carbon neutral by 2030 (Sparber et al. 2010). This carbon sequestration mitigation potential of urban trees is being considered a regulating ecosystem service (Escobedo et al. 2011) and according to the EU Biodiversity Strategy 2020, by 2014 all European member states should map and assess the state of ecosystem services in their national territory (Maes et al. 2012). However, with the exception of some studies in Germany and the United Kingdom (Davies et al. 2011; Strohbach & Haase 2012; Strohbach et al. 2012) and assessments in Spain, Switzerland, and England using the i-Tree Eco/Urban Forest Effects (UFORE) model developed in the USA (i-Tree 2012a) we know of no studies of carbon storage and sequestration by urban trees in Europe and particularly in South Tyrol in peer-reviewed literature.

The United States models and modeling approaches are currently the basis for tools that are become increasingly applied not only in North America and China, but in Europe as well (e.g. UFORE, i-Tree Eco, i-Tree Streets, CUFR Tree Carbon Calculator (CTCC)) (i-Tree 2012b). Aguaron and McPherson (2012) have compared the Eco/UFORE and other North American C storage estimation models with tree data from a United States city. In addition, Escobedo et al. (2013) destructively measured aboveground tree biomass and C storage for urban Quercus spp. and compared their C storage results to tree-level C storage estimates from the Eco/UFORE and CTCC models. The authors found that Eco consistently underestimated C storage by 15% and CTCC overestimated by 2%, on average. To our knowledge, the appropriateness of these models for European trees has not been assessed. According to Ferrini and Fini (2011) errors of modeled carbon estimates can be substantial.

As such, C storage and sequestration methods that are developed using local or regional allometric equations and site-specific growth rates and dendrometrics should provide more reliable, consistent, and context-specific information. Therefore, the two specific objectives of this study were to: (1) estimate aboveground carbon storage and sequestration for a subsample of public trees in Bolzano using Italian and European allometric equations and local growth rates obtained from re-measurements, and (2) assess the performance of this method against carbon storage and sequestration models from the USA that are increasingly being applied in Europe. The role of biomass removal for maintenance and its related carbon emissions will also be discussed.

Material and methods

Study area

The study area was the City of Bolzano, Italy, located in the autonomous region of Trentino-Alto Adige/South Tyrol in northern Italy (46° 29′ 28″ N, 11° 21′ 15″ E). Bolzano is the capital of the province of Alto Adige/South Tyrol in the Southern Alps and its 2011 census showed a population of about 100,000 people (Comune di Bolzano 2012). The city of Bolzano covers an area of over 50 km² with approximately 12,000 public urban trees (Comune di Bolzano 2010, personal communication). Multi-year climatic data (1981–2011) shows that the annual average temperature in Bolzano is 12.6°C and average annual rainfall is 701 mm (Provincia Autonoma di Bolzano 2012). The coldest month of the year is January with a minimum of −4.1°C, a maximum of 6.6°C, and an average of 1.2°C. The warmest month is July with a minimum of 16.3°C, a maximum of 30.3°C, and an average of 23.3°C. The extreme records range from −17°C to +40°C.

Allometric equations and carbon storage

According to the principle of allometry, a tree allometric equation relates biomass, volume, or several tree components to stem diameter at breast height (DBH), tree height, and/or other dendrometric variables (Henry et al. 2013).

The use of group (i.e. composite) allometric equations, or the processes used to assign both species and non-species-specific allometric equations to sampled urban trees to estimate biomass and subsequent C storage, is an internationally accepted approach (Jo & McPherson 1995; Strohbach & Haase 2012). Indeed, this approach is the basis of models such as the Eco/UFORE and CTCC (Nowak et al. 2008; Aguaron & McPherson 2012; Yoon et al. 2013). The majority of these allometric equations are derived primarily from non-urban, forest-grown trees that are destructively sampled (i.e. felled and weighed on site; Basuki et al. 2009; Yoon et al. 2013). However, due to local regulations, liability, and public perceptions and safety, destructive sampling is expensive and difficult in an urban environment. Although McHale et al. (2009) found that some of these allometric equations for forest-grown trees yield similar biomass estimates for urban-grown trees; these allometric equations can produce very different results when applied to sites outside the region where the equations were originally developed (Zapata-Cuartas et al. 2012).

Therefore, since the Eco/UFORE and CTCC models use North American equations and this study was conducted in Europe, we used site-specific tree species, tree dendrometrics such as stem circumference (subsequently converted to diameter), and height data and applied mostly European-specific allometric equations derived from the literature (Appendix 1) to better approximate local urban tree C storage estimates. The equations in Appendix 1 were used specifically to calculate dry weight aboveground biomass of each measured tree and not total dry weight biomass due to the complexity in estimating the belowground portion as reported by Hutyra et al. (2011) and Strohbach and Haase (2012). Dry weight aboveground biomass, obtained from equations in Appendix 1 was multiplied by 0.5 to obtain C storage.

Field sampling

Our study used existing tree inventory data from the City of Bolzano, Italy (Servizio Giardineria 2013). As is the case for most cities, Bolzano’s urban tree inventory was developed to assess tree condition, hazards and risks, and overall public safety. Therefore, the data had to be supplemented with field measurements. We used ArcGIS (Version 10) and obtained a subsample of trees from the inventory using a stratified random sample – according to land cover classes – and selected individual trees in the inventory’s spatial database for field measurement (Piano Urbanistico Comunale 2012).

During June 2011, we measured the selected trees and collected data for 475 trees. Specific measurements included: tree species, total and crown base height (m), crown width in two directions (m), percent crown dieback, percent missing canopy, and crown light exposure (Figure 1). Specific field methods are outlined in Nowak et al. (2008). Since Bolzano’s urban tree inventory only provided girth measurements data at 1 meter above the surface (Russo 2013), we also re-measured tree circumference (cm) at 1 meter and additionally at 1.37 meter above the surface. The circumference was then converted to diameter (DH) and DBH by dividing by π.

Figure 1. (color online) Urban trees parameters sampled in Bolzano, N Italy: Sp = species, DH = diameter at 1 m, DBH = diameter at breast height (1.37 m), Cb = crown base height, Ht = total height, Cw = crown width, CLE = crown light exposure, PCM = percent canopy missing, D = crown dieback (Nowak et al. 2002, 2008).

The DBH data were used in our European allometric equations, and DBH and other data were used in the UFORE model to quantify carbon storage and sequestration. The CTCC requires only information on tree species, DBH, and an overall characterization of Bolzano’s climate.

Estimated height increments and growth rates

Several allometric equations in Appendix 1 require continuous data on tree height (m) in addition to DBH. However, Bolzano’s tree inventory provided only tree height classes (Russo 2013). To obtain the necessary tree height increment data we used our 2011 subsample data to develop an Ordinary Least Squares predictive regression model h = f (DBH) based on the 2011 subsample’s measured tree height (h; m) and DBH (cm) to estimate the function parameters for the statistical relationship of DBH – h (Table 1; (Pretzsch 2009)). The model was developed using the PROC REG procedure in the Statistical Analysis Software (Version 9.2).

Table 1. Tree height (y, in meters) prediction using DBH (x, in centimeters) models for urban tree genera in Bolzano, Italy.

Genus	Models	R^2
Abies, Pinus, Picea	y = 6.8788 ln(x) − 10.131	0.54
Acer	y = 5.2586 ln(x) − 5.1651	0.81
Alnus, Carpinus, Ostrya	y = 0.4717x + 2.5591	0.63
Betula, Fagus	y = 0.3059x + 3.4955	0.64
Cupressus	y = −0.004x^2 + 0.5878x − 0.5975	0.75
Fraxinus	y = 4.732 ln(x) − 3.621	0.53
Prunus	y = 0.0038x^2 + 0.1054x + 4.8598	0.58
Quercus	y = 0.0045x^2 + 0.0715x + 4.9053	0.60
Robinia	y = 5.0266 ln(x) − 4.4342	0.65
Salix, Populus	y = 11.024 ln(x) − 21.16	0.96
Tilia	y = 2.1438x^0.5301	0.63
Ulmus, Zelkova	y = 12.837 ln(x) − 30.193	0.97

We then used the tree diameter–height models from Table 1 to estimate height in 2011 (H_est1) and height in year 2012 (H_est2) by using measured 2011 DBH and the estimated growth rates reported in the following results section. The mean annual tree height increment () was then calculated as the difference between the estimated height at year 2011 and the estimated height at year 2012 using Equation (1):

(1)

where is the mean annual tree height increment (m/year), H_est2 is the estimated height (m) at year 2012 (m), and H_est1 is the estimated height (m) at year 2011. The tree height in 2012 (H₂) was then derived using Equation (2) using the mean annual tree height increment (H_i; m/year) multiplied by the number of years (n) added to the height measured in 2011 (H_1m):

(2)

Finally, diameter growth of the individual trees was calculated using diameter at 1 meter above the surface (DH), as the difference between the DH measured at the beginning and the end of a given time period (Laar & van Akça 2007).

Specifically in this study, the annual growth rate (AGR; cm/year) was calculated using Equation (3) (Stoffberg et al. 2008):

(3)

where, AGR is the annual growth rate (cm/year), DHY₁ is the DH at a given year i.e. different DHs were measured during different years for different trees thus years change with different locations, DHY₂ is the DH in 2011 and t is the time period (months) between measurements. To increase sample sizes for individual tree species, AGR and mean were averaged at the taxonomic order and division level (USDA 2013). Trees that had a DHY₂ less than DHY₁ were excluded from the analyses.

Carbon sequestration and biomass removals from pruning operations

Annual carbon sequestration was the estimated amount of carbon assimilated to biomass by a tree stem and branches during 1 year of growth. Thus, in this study, annual gross carbon sequestration (kg/year) was estimated as the difference of C stored between year x (2011) and year x + 1 (2012) (Liu & Li 2012) and was determined using an individual tree’s annual growth rate and predicted height increment as explained in the previous section.

A report on municipal waste (ISPRA 2012) shows that the green waste biomass from urban vegetation mowing and urban tree pruning operations in the Trentino-Alto Adige/South Tyrol region was 15,705 t in 2009. Hence, the amount of biomass waste from pruning operations can and should be accounted for when estimating net carbon sequestration effects from urban trees (Sajdak & Velazquez-Marti 2012). Thus, to better estimate a portion of the net annual carbon sequestration, we estimated the amount of annual biomass removals to account for maintenance-related indirect C emissions associated with Bolzano’s tree population. According to Bolzano’s Gardens Department (Personal communication, 2012), trees in parks are primarily pruned for health reasons. If there are no particular problems, the trees are not pruned and the only two trees that are subject to periodic and systematic pruning during the analysis period were Sophora japonica L., which are pruned every 2 years, and Platanus hybrida Brot., which are pruned every 7–10 years. However, the amount of biomass that is removed by tree pruning operations in Bolzano has never been measured. So, to account for maintenance-related C emissions for biomass removal, we calculated the green waste biomass removal (y) obtained from pruning Sophora japonica L. using the following linear Equation (4) derived from Sajdak and Velazquez-Marti (2012) data:

(4)

where y is the dry weight biomass obtained from annual pruning operations (y; kg) and x is the DBH (cm). This assumes one single pruning intervention in 2012 for all Sophora japonica L. trees in our subsample.

UFORE and CTCC data input methods

The UFORE model is used to estimate the benefits and costs of urban trees and was developed in the late 1990s by the United States Department of Agriculture’s Forest Service to quantify urban forest structure, function, and value (Nowak & Crane 2000). A recent and updated user interface version is available for use and is referred to as i-Tree Eco. Using field measurements, study area characteristics, hourly annual meteorological data, and hourly annual pollution concentrations data the model quantifies several ecosystem services and disservices (Nowak et al. 2008).

The UFORE/i-Tree Eco model (referred to as UFORE hereafter) calculates urban forest and individual total tree (aboveground and belowground) carbon storage using forest-grown tree allometric biomass equations (Nowak 1994; Nowak et al. 2002). Dry weight biomass estimates for open-grown street trees are multiplied by a factor of 0.8 (Nowak et al. 2002) since these trees tend to have less aboveground biomass than predicted by these forest-derived biomass equations for trees of the same DBH (Nowak 1994; Nowak & Crane 2000). Total tree biomass estimates are then multiplied by 0.5 to obtain total stored carbon. Specific details can be found in Nowak et al. (2008).

Annual gross C sequestration is also estimated by the UFORE model as the difference in estimates of changes in carbon storage between year x and year x + 1 (Nowak et al. 2002). Once C storage is obtained for year x, a growth rate is used for each tree to obtain a DBH and subsequent C storage at year x + 1 for the same tree. The model uses average DBH growth rates obtained from a few sites in the USA (Nowak et al. 2002; Lawrence et al. 2012). For example, for trees in forest stands, the model uses an annual growth rate of 0.38 cm/year (Smith & Shirley 1984; Nowak et al. 2002), for park-like structure the model uses 0.61 cm/year (Nowak et al. 2002). Average height growth is calculated based on formulas obtained from Fleming (1988) as reported by Nowak et al. (2002) and the specific DBH growth factor used for the tree. According to Nowak et al. (2002) and Nowak et al. (2008), growth rates are then adjusted based on the tree condition (i.e. no adjustment for trees in fair to excellent condition, trees in poor condition are multiplied by 0.76, critical trees by 0.42, dying trees by 0.15, and dead trees by 0). Adjustment factors are based on percent crown dieback and the assumption that less than 25% crown dieback had a limited effect on DBH growth rates. The more recent Eco version also adjusts the base growth rate based on the study area’s average annual plant growing period (i-Tree 2012a).

Another available model is the Center for Urban Forest Research’s (CUFR) Tree Carbon Calculator (CTCC; Urban Forest Project Reporting Protocol 2008) that was developed by the USDA Forest Service, Pacific Southwest Research Station. The CTCC is a Microsoft Excel spreadsheet that estimates urban tree carbon dioxide sequestration and building heating/cooling energy savings. The model estimates CO₂ sequestration for individual trees located in one of the sixteen climate zones of the USA (Aguaron & McPherson 2012). The CTCC requires climate zone, species, and DBH or age input data to calculate individual tree CO₂ sequestration (kg/tree/year), total CO₂ stored (kg/tree), and aboveground biomass (dry weight) (kg/tree). Tree size and growth data were developed from samples of about 1000 urban trees and approximately 20 predominant species in each of the 16 United States reference climate zone cities (Aguaron & McPherson 2012). Many of the biomass equations used to derive total CO₂ stored and sequestered are derived from open-grown city trees according to Aguaron and McPherson (2012).

Using tree data from our Bolzano tree inventory 2011 subsample, we adapted the input variables for use in the UFORE (Version ACE 6.5) model’s complete tree inventory option based on methods outlined in Nowak et al. (2002, 2008). We also formatted our subsample data for the use in the CTCC model. According to McPherson (2010) and McPherson and Peper (2012), the use of the CTCC and i-Tree Streets (formerly STRATUM) models is dependent on selecting an appropriate reference city in the United States. Therefore, Bolzano’s trees were matched to existing CTCC tree species and climates using similarities in tree taxonomy, growth forms, and overall tree structure. Specific CTCC inputs for Bolzano are presented in (Appendix 2) and are based on climate information from Bonatti (2008).

Finally, to better assess our European allometric-based urban tree C storage and sequestration methods to the UFORE and CTCC models, we converted UFORE estimated total tree C estimates into aboveground C by subtracting the belowground portion using a root-to-shoot ratio of 0.26 as reported in Nowak et al. (2002) and Cairns et al. (1997). The CTCC model was adjusted by dividing the total biomass by 1.28 since total biomass is 1.28 times the aboveground biomass (Aguaron & McPherson 2012).

Model assessment

To assess the performance of the UFORE and CTCC model against our allometric equation-based approach, we tested for significant differences (P < 0.05) between these three methods using the PROC TTEST procedures in SAS version 9.2. Specifically, we used a paired t-test to test the null hypothesis that there were no differences in carbon storage and sequestration between the means from each of the three methods. Additionally, data were checked for normality using Q–Q plots and the Kolmogorov–Smirnov test. Data for the three model outputs were then fitted to a linear regression and comparison made between variables (e.g. Allometric equations vs. CTCC, Allometric equations vs. UFORE, and CTCC vs. UFORE) using a PROC GLM procedure in SAS and tested (P < 0.05) to determine whether the slope differed from 1.0.

Results

Forest structure

The mean AGR in the subsample are presented in Table 3 according to taxonomic ‘division’ and ‘order’. For example, the order Fabales includes the following species: Cercis siliquastrum L., Gleditsia triacanthos L., Gymnocladus dioicus (L.) K. Koch, Robinia pseudoacacia L., Sophora japonica L., and Wisteria sinensis (Sims) DC. Overall, the order Rosales had the greatest mean AGR (1.02 cm/year) while the order Magnoliales had the lowest mean AGR (0.57 cm/year). Table 4 presents the mean annual height increments in m/year and shows that Populus spp. and Salix spp. had the greatest height increments (0.63 m/year), while Tilia spp., Robinia pseudoacacia L., Gleditsia triacanthos L., and Sophora japonica L. had the lowest height increments (0.13 m/year).

Table 2. The ten most common public tree species in Bolzano, Italy, and the number sampled in 2011 (n), mean diameter at breast height (DBH) and height in meters (m). SE is standard error.

Tree species	n	DBH		Height
		Mean (cm)	SE	Mean (m)	SE
Quercus pubescens Willd.	46	21.8	1.12	9.2	0.52
Cedrus deodara (Roxb.) G. Don	22	66.1	4.44	22.4	1.38
Platanus hybrida Brot.	22	64.5	4.18	21.8	0.94
Acer platanoides L.	20	31.8	4.20	12.0	0.84
Acer pseudoplatanus L.	20	35.3	3.68	13.2	0.88
Sophora japonica L.	19	39.0	3.54	13.4	0.82
Betula pendula Roth	18	29.6	4.08	12.0	1.13
Aesculus hippocastanum L.	13	34.1	5.57	12.1	1.26
Cupressus sempervirens L.	12	29.7	3.79	12.4	1.27
Tilia americana L.	12	49.8	3.89	18.1	0.80

Table 3. Mean annual growth rate (AGR) of urban trees in the city of Bolzano, Italy.

Order	Sample size # of trees	Mean AGR (cm/year)	Standard error
Fabales	20	0.73	0.11
Fagales	50	0.77	0.08
Ginkgoales	6	0.80	0.27
Hamamelidales	30	0.89	0.10
Magnoliales	13	0.57	0.11
Malvales	27	0.62	0.10
Pinales	59	0.72	0.08
Rosales	17	1.02	0.14
Salicales	10	0.99	0.23
Sapindales	57	0.63	0.07
Scrophulariales	12	0.82	0.22
Urticales	11	0.85	0.26
Division
Magnoliophyta	279	0.78	0.03

Table 4. Predicted mean annual height increments (m/year) of urban tree species in the city of Bolzano, Italy.

Species	n	Mean (m/year)
Abies spp., Picea spp., Pinus spp.	23	0.20
Cupressus spp.	13	0.24
Acer spp.	52	0.15
Alnus spp., Carpinus spp., Ostrya spp.	16	0.36
Fagus spp., Betula spp.	21	0.24
Fraxinus spp., Olea europea L.	16	0.22
Populus spp., Salix spp.	9	0.63
Prunus spp., Pyrus spp.	25	0.23
Robinia pseudoacacia L., Gleditsia triacanthos L., Sophora japonica L.	22	0.13
Quercus spp.	67	0.24
Tilia spp.	30	0.13
Ulmus spp., Zelkova carpinifolia (Pall.) Dippel	9	0.25

Note: n: number of trees sampled.

Comparison of storage estimations

Using our allometric equation method we estimated that the total tree carbon stored in our subsample was 179.14 Mg. Meanwhile, using our field measurement data as model inputs, we estimate 140.15 Mg of C storage using the CTCC model and 134.89 Mg using the UFORE model (Figure 2). The amount of carbon stored in the five most frequent tree species using the three different methods are also presented in Figure 3.

Figure 2. Total carbon stored (Mg) by 475 trees in Bolzano calculated using three different methods.

Figure 3. Average carbon storage (kg) estimates for the five most common tree species calculated using three different methods. Error bars represent ± one standard error of the mean.

The paired t-test shows that predictions from our allometric equations are significantly higher than the CTCC (t = 4, P < 0.0001) and UFORE (t = 8.43, P < 0.0001) models. But, there was no significant difference in predictions between the CTCC and UFORE (t = −0.82, P = 0.413). Additionally, the regression slope between our allometric equations and the CTCC model was significantly different from 1 (P = 0.003), which suggests that predictions from two methods are also different. Similarly, the slope between our allometric equations, UFORE (P =< 0.0001), and CTCC and UFORE (P =< 0.0001) were also significantly different from 1 (P =< 0.0001); therefore, we can say that predictions were statistically different.

Comparison of C sequestration estimates

The total gross annual carbon sequestration for trees in our subsample was 5.73 Mg/year using the allometric equations and Bolzano’s growth rates and/or height increment predictions. However 8.27 Mg/year of annual carbon sequestration were estimated using the CTCC model and 5.82 Mg/year using the UFORE model (Figure 4). The amount of carbon sequestered for the five most frequent tree species using the different methods is shown in Figure 5.

Figure 4. Annual carbon sequestration (Mg/year) by 475 trees in Bolzano calculated using three different methods.

Figure 5. Average carbon sequestration (kg/year) estimates for the five most common tree species calculated using three different methods. Error bars represent ± one standard error of the mean.

A paired t-test showed that predictions from our allometric equations were significantly lower than the CTCC model (t = −7.71, P < 0.0001). Also, there was no significant difference in estimates between the allometric equations and the UFORE model (t = −0.60, P = 0.54). However, estimates from the CTCC model were significantly higher than the UFORE model (t = 7.30, P < 0.0001) and the regression slope between the allometic equations and the CTCC model was significantly different from 1 (P < 0.0001). This suggests that the predictions from these two methods are also different. Similarly, the slope between the allometric equations and UFORE model (P =< 0.0001) and the CTCC and UFORE models (P =< 0.0001) were also significantly different from 1 (P =< 0.0001); thus model predictions are also different.

The green waste biomass from annual pruning operations of Sophora japonica L. was estimated at 678 kg/year. Assuming this biomass is burned and emitted as carbon, in the same year the trees were pruned, a potential 339 kg can be emitted per year. Since the gross annual C sequestration from trees in our subsample was 5710 kg, our net annual C sequestration (i.e. gross C sequestration minus C emitted from maintenance) is 5371 kg/year. Additionally, the 678 kg/year of dry weight biomass could be used as biofuel or as compost and thus acts as a carbon sink.

Discussion and conclusion

Our study provides a quantification of the C stored and sequestered by urban trees in an Italian city in the Southern Alps. As opposed to studies that estimate urban tree C storage and sequestration using North American models, we present an approach and protocol that primarily uses European allometric equations, city-specific measured growth rates and predicted tree height increments. In addition the study developed a protocol and compiled a list of biomass equations that can be used to estimate C storage, mean annual growth rates, and predict height increment for northern Italian urban trees at the order, division, and genera level, respectively. Finally, the study assessed the performance of two USA urban tree C models against our allometric equation approach.

Allometric equations and growth rates

The allometric equations used in our dry weight and biomass C storage estimates were developed primarily for European, forest-grown trees and were applied to 60.3% of trees in our subsample. More specifically, Italian-specific equations were applied to 51.5% of the trees in our subsample, equations from Spain and the UK were applied to 0.2% and 8.6% trees in our subsample (Bunce 1968; Zianis et al. 2005; Muukkonen & Mäkipää 2006; Tabacchi et al. 2011a, 2011b; Ruiz-Peinado et al. 2012). Due to the presence of non-native trees and lack of European-specific equations for certain species, the remaining equations were from China (4% of subsampled trees (Liu & Li 2012)) and North America (35.7% of subsampled trees; (Jenkins et al. 2003)).

Overall, our growth rates are different from those reported by Jo and McPherson (1995), Iakovoglou et al. (2002), and Lawrence et al. (2012) for trees in the United States. The order Fagales, for example, had an AGR estimated at 0.77 cm/year which was lower than the 0.85 cm/year (average growth rates of subtropical Q. laurifolia, Q. nigra, Q. virginiana, O. virginiana) reported by Lawrence et al. (2012). Also, our growth rates for hardwood trees estimated at 0.78 cm/year (Magnoliophyta) was lower than the 1.09 cm/year reported by Jo and McPherson (1995), but greater for softwood trees 0.72 cm/year (Pinales) instead of 0.51 cm/year (Jo & McPherson 1995). Our results also differ from those reported in Strohbach et al. (2012) in Leipzig, Germany, and in Bühler et al. (2007) in Copenhagen, Denmark. These differences could be due to our inability to report species-specific rates and several factors such as soil and climatic conditions as well as maintenance characteristics such as irrigation and light availability.

Carbon storage and sequestration

The C storage and sequestration results from this study are difficult to assess in terms of accuracy and to compare with other studies because of the use of different estimation methodologies, climatic condition, different species composition, and urban forest structures (Jo & McPherson 1995; Strohbach & Haase 2012). Our estimates are different from those reported in Table 5. In particular, the comparison between our C estimates with the UFORE model and other European studies that have used the UFORE/i-Tree Eco model in Europe show that the average carbon storage and sequestration per tree was higher in our study than estimates reported by Wälchli (2012) in Zurich in Switzerland (about 235 km from Bolzano) and Paoletti et al. (2011) in Florence, Italy (about 300 km from Bolzano). The difference is likely due to Bolzano’s species composition, tree size, and health (Martin et al. 2012).

Table 5. Reported average per tree carbon storage and sequestration and estimation methods for case studies in Europe.

Study area	# of trees	Average C storage (kg)	Average C sequestration (kg/year)	Method	References
Bolzano, Italy	475	377.14	12.06	Aboveground C in urban trees, European allometric equations, and field data	This study
Bolzano, Italy	475	295.06	17.41	Aboveground C in urban trees, CUFR Tree Carbon Calculator (CTCC), and field data	This study
Bolzano, Italy	475	283.98	12.26	Aboveground C in urban trees, UFORE model, and field data	This study
Florence, Italy	885	354.60	9.79	Aboveground and belowground C in trees, UFORE model, and field data	Paoletti et al. (2011)
Leicester, United Kingdom	267,647	206.61	na	Aboveground C in public trees, stratified random sampling across land cover and land ownership	Davies et al. (2011)
Lisbon, Portugal	41,247	509.86	43.06	Above- and below-ground C* in trees, STRATUM model and field data	Soares et al. (2011)
Padua, Italy	219	138.62	12.84	Aboveground and belowground C* in trees, STRATUM model, and field data	Crema (2008)
Padua, Italy	219	260.36	na	Aboveground and belowground C in trees, N. American equations	Crema (2008)
Zurich, Switzerland	130	348.88	12.97	Aboveground and belowground C in trees, i-Tree Eco model, and field data	Wälchli (2012)
Zurich, Switzerland	130	375.46	30.69	Aboveground and belowground C* in trees, i-Tree Streets model	Wälchli (2012)

Note: *We converted CO₂ to carbon, na = not analyzed.

As discussed in Nowak et al. (2008), the UFORE model estimates gross C sequestration using over 20 required field measurements and a series of assumptions that include non-measured root-to-shoot ratios, non-city-specific growth rates, adjustments according to tree condition, light exposure, and land use, modeled removal and decomposition rates, and no inclusion of city-specific biomass removal for maintenance-related C emissions (Escobedo et al. 2013; Yoon et al. 2013). Thus, our gross and net C sequestration estimates based on annual re-measurement data, AGRs, and predicted height increments values for Bolzano – as well as accounting for some of the observable biomass removal for maintenance-related C emissions – presents an alternative protocol based on city-specific information and fewer assumptions and parameters derived from the United States.

Model assessment

According to Jo and McPherson (1995), McHale et al. (2009), and Yoon et al. (2013) the use of allometric biomass equations based on forest-grown trees can overestimate or underestimate urban tree biomass. For example an urban tree with the same DBH or height as a forest trees could have a different biomass due to the conditions of the urban environment relative to forest-grown trees (Jo & McPherson 1995).

The UFORE model reduces biomass estimates of open grown street trees by 20% based on a study of 30 street trees of nine different species in Chicago, USA (Nowak 1994; Yoon et al. 2013). However, in the case of Bolzano’s urban trees, we observed urban trees were often not open-grown, were in overall good condition, were regularly fertilized and irrigated relative to forest-grown trees. Therefore, given the uncertainty in UFORE’s assumption, we did not subtract 20% for open grown trees using our allometric equation method.

Overall, the UFORE model produced the lowest estimates (134.89 Mg) for carbon storage, and this might be because forest-based equations are used exclusively with application of the 0.8 multiplier to open-grown trees (Aguaron & McPherson 2012). The CUFR Tree Carbon Calculator (CTCC), however, produced an intermediate estimate of 140.15 Mg while our allometric equations produced larger estimates of 179.14 Mg. Accounting for Nowak’s (1994) and Peper and McPherson (1998) correction factor for open-grown urban trees, multiplying the carbon storage from our allometric equation method by a factor of 0.8 results in a carbon storage of 143.3 Mg that is more similar, yet still greater, than that estimated by the UFORE model.

The supporting evidence or verification of the application of UFORE correction factors are limited (Yoon et al. 2013). Yoon et al. (2013) quantify the variability these correction factors can have on actual Korean urban tree biomass estimates.

The CUFR Tree Carbon Calculator (CTCC) C sequestration estimates for our subsample was the greatest at 8.27 Mg/year, while the UFORE model (5.73 Mg/year) and our equations (5.82 Mg/year) produced similar estimates. These results are similar to Escobedo et al. (2013) and Aguaron and McPherson (2012), who found that the UFORE model (i-Tree Eco) produced the lowest carbon storage estimate, while the CUFR Tree Carbon Calculator (CTCC) produced the largest C sequestration estimates. There are however not only statistical but also practical differences in these three methods for the calculation of C storage and sequestration as well. Table 6 shows some overall strengths and weaknesses of the three carbon calculation approaches and models for Bolzano, Italy.

Table 6. Some general strengths and weaknesses for three carbon storage and sequestration estimation methods applied in Bolzano, Italy: Urban Forest Effects/Eco (UFORE), Allometric equations, and Center for Urban Forest Research Tree Carbon Calculator (CTCC).

	Allometric equations	CTCC	UFORE
Strengths	Local – regional equations	User-friendly formatted spread sheet	User-friendly (i-Tree Eco) model interface
	Species-specific	Freely available on internet	Freely available on internet
	Requires only species, DBH, and height	Requires only species, DBH or age	Species-specific
	Local growth rates	Urban tree based equations	Calculates additional ecosystem services
Weaknesses	Time-consuming literature review and compilation of equations	North American urban-based equations	North American biomass equations
	Forest-based tree biomass equations	Limited number of species	Forest tree biomass equations
	Equations, coefficients, measurements, and calculations need to be entered and calculated manually using spread sheet or other data-processing software	North American growth rates	Requires species, DBH, height, land use and tree condition field measurements
		US climate zones	North American land-use based growth rates
			More expensive as specialists needed for data collection of over 20 measurements (Natural England 2013)
			Professional field data costs were 1000 Euros per 100 trees in Bolzano

Note: DBH: Diameter at Breast Height; CTCC: Center for Urban Forest Research Tree Carbon Calculator; UFORE: Urban Forest Effects.

Recommendations and applications

Further research is needed for accurate C measurements and for developing urban tree allometric equations using destructive or non-destructive approaches (Yoon et al. 2013). For example, it could be possible to develop urban tree equations using destructive sampling of trees removed in new developments or reconstruction sites as done by Escobedo et al. (2013). The emissions of CO₂ from mechanical maintenance operations can be quite variable and substantial based on the type of equipment, frequency, and intensity of maintenance activities and would require additional survey work and data collection (Strohbach et al. 2012), therefore we obtained only biomass removal for maintenance-related C emissions. However, this is an opportune area of future research. Further research is needed to determine the true C footprint of urban trees in Europe.

Another limitation was the calculation of the annual height and diameter increments that are not based on felled tree measurements tree cores, or re-measured height. While tree ring and stem analysis of felled trees is the most accurate method it is time-consuming and expensive, and not reasonable for urban trees. On the other hand, re-measurement of height on the same trees can have a large measurement error relative to the actual height increment (Hasenauer & Monserud 1997). Therefore, a modern approach to estimate urban tree heights and increments is using Light Detection and Ranging (LiDAR) data (Shrestha & Wynne 2012; Ramdani 2013). LiDAR data can be also used for estimating urban tree volume that can then be converted to biomass using specific gravity values for individual tree species (McHale et al. 2009).

In conclusion, our methods, findings, and model assessment can be used for integrating and assessing urban landscapes and trees into environmental design, planning, and climate-change initiatives and policies. The use of regional and European-specific biomass equations and local annual growth estimates in Italian cities can provide an alternative and arguably more reliable and transparent carbon storage and sequestration estimates than those from commonly used USA models. Findings from this study on annual growth rates, annual height increments, and model assessments can be applied to existing tree inventories and used for the development of similar protocols, model, and tools for other Italian cities or other urban areas in the Southern Alps or Europe. Similarly, green and dry weight biomass from pruning operations can be estimated and used to predict green waste yield from urban landscape maintenance activities for use as biofuel and compost, and greenhouse gas emission information from maintenance operations can also be used in green space life cycle analyses. We propose that our protocol and results from this study can be used to plan, design, and manage cities to maximize the potential of urban trees to provide ecosystem services and for developing carbon neutral policies.

Acknowledgments

We would like to thank everyone who helped during the development of this paper, in particular Dr. Ulrike Buratti (Municipality of Bolzano), Dr. Mauro Tomasi, and Prof. Giustino Tonon; this paper could not have been written without your support. The authors would like to thank the anonymous reviewers for their valuable comments and suggestions to improve the quality of the paper.

Appendix 1

Table S1. Sources and geographical distribution of allometric equations that were used for biomass calculation in Bolzano, Italy.

Species sampled	Equation	Parameters	Region	Reference
Abies spp.	dw4 = b1 + b2d^2h + b3d	dw4 = total aboveground dry weight b1 = −2.1386 b2 = 1.8125 × 10^−2 b3 = 1.1089 h = total tree height d = diameter at breast height	IT	Abies alba Mill. (Tabacchi et al. 2011a, 2011b)
Acer spp.	dw4 = b1 + b2d^2h	dw4 = total aboveground dry weight b1 = 6.4595 b2 = 2.6368 × 10^−2 h = total tree height d = diameter at breast height	IT	Acer spp. (Tabacchi et al. 2011a, 2011b)
Aesculus spp., Catalpa spp., Celtis spp., Cercis spp., Cornus spp., Diospyros spp., Ginkg biloba, Gleditsia spp., Gymnocladus spp., Hibiscus spp., Juglans spp., Koelreuteria spp., Lagerstroemia spp., Laurus spp., Liquidambar spp., Liriodendron spp., Magnolia spp., Melia spp., Morus spp., Paulownia spp., Photinia spp., Platanus spp., Pterocarya spp., Tamarix spp., Toona spp., Wisteria spp.	bm = Exp(β0 + β1 In dbh)	bm = total aboveground biomass β0 = −2.4800 β1 = 2.4835 dbh = diameter at breast height	NA	Mixed hardwood (Jenkins et al. 2003)
Alnus spp.	dw4 = b1 + b2d^2h + b3d	dw4 = total aboveground dry weight b1 = −1.6747 × 10 b2=1.7930 × 10^−2 b3=2.6664 h = total tree height d = diameter at breast height	IT	Alnus spp. (Tabacchi et al. 2011a, 2011b)
Betula spp., Corylus spp.	loge y = a + b (loge x)	y = tree dry weight (trunk + branches) x = tree girth at 1.3 m a = ^−5.223864 b = 2.425436	UK	Birch combined (Bunce 1968)
Carpinus spp., Ostrya spp.	dw4 = b1 + b2d^2h	b1 = 3.2485 b2 = 3.0167 × 10^−2	IT	Carpinus spp., Ostrya spp. (Tabacchi et al. 2011a, 2011b)
Cedrus spp., Chamaecyparis spp., Cryptomeria spp., Metasequoia spp., Sequoiadendron spp., Taxodium spp.	bm = Exp(β0 + β1 In dbh)	bm = total aboveground biomass β0 = −2.0336 β1 = 2.2592 dbh = diameter at breast height	NA	Cedar/larch (Jenkins et al. 2003)
Cephalotaxus spp., Taxus spp.	bm = Exp(β0 + β1 In dbh)	bm = total aboveground biomass β0 = −2.5384 β1 = 2.4814 dbh = diameter at breast height	NA	True fir/hemlock (Jenkins et al. 2003)
Cupressus spp.	dw4 = b1 + b2d^2h + b3d	dw4 = total aboveground dry weight b1 = −4.1345 b2 = 2.4359 × 10^−2 b3 = 1.4156 h = total tree height d = diameter at breast height	IT	Cupressus spp. (Tabacchi et al. 2011a, 2011b)
Fagus spp.	dw4 = b1 + b2d^2h	dw4 = total aboveground dry weight b1 = 1.6409 b2 = 3.0775 × 10^−2 h = total tree height d = diameter at breast height	IT	Fagus sylvatica L. (Tabacchi et al. 2011a, 2011b)
Fraxinus spp.	dw4 = b1 + b2d^2h	dw4 = total aboveground dry weight b1 = 2.1893 b2 = 3.2949 × 10^−2 h = total tree height d = diameter at breast height	IT	Fraxinus spp. (Tabacchi et al. 2011a, 2011b)
Olea europaea L.	Ws = 0.0114 × d^2 × h Wb7 = 0.0108 × d^2 × h Wb2–7 = 1.672 × d Wb2 + l = 0.0354 × d^2 + 1.187 × h	Ws: biomass weight of the stem fraction (kg); Wb7: biomass weight of the thick branches fraction (diameter larger than 7 cm) (kg); Wb2–7: biomass weight of medium branches fraction (diameter between 2 and 7 cm) (kg); Wb2 + l: biomass weight of thin branches fraction (diameter smaller than 2 cm) with leaves (kg); Wr: biomass weight of the belowground fraction (kg); d: diameter at breast height (cm); h: tree height (m)	ES	Ruiz-Peinado et al. (2012)
Picea spp.	dw4 = b1 + b2d^2h + b3d	dw4 = total aboveground dry weight b1 = 1.4146 × 10^−1 b2 = 1.7620 × 10^−2 b3 = 5.6209 × 10^−1 h = total tree height d = diameter at breast height	IT	Picea abies (L.) Karst (Tabacchi et al. 2011a, 2011b)
Pinus halepensis Mill.	dw4 = b1 + b2d^2h + b3d	dw4 = total aboveground dry weight b1 = −8.1012 b2 = 2.1559 × 10^−2 b3 = 2.2591 h = total tree height d = diameter at breast height	IT	Pinus halepensis Mill. (Tabacchi et al. 2011a, 2011b)
Pinus nigra Arnold	ABW = a + b·D^2·H + c·D^2	ABW = total aboveground woody biomass a = −3.5712 b = 0.014429 c = 0.068047 H = height D = diameter	IT	Pinus nigra Arnold, Equation 739 (Muukkonen & Mäkipää 2006)
Pinus pinea L.	dw4 = b1 + b2d^2h	dw4 = total aboveground dry weight b1 = 4.5885 × 10^−1 b2 = 2.5176 × 10^−2 h = total tree height d = diameter at breast height	IT	Pinus pinea L. (Tabacchi et al. 2011a, 2011b)
Pinus strobus L.	dw4 = b1 + b2d^2h	dw4 = total aboveground dry weight b1 = 5.6156 b2 = 1.5939 × 10^−2 h = total tree height d = diameter at breast height	IT	Exotic pine group (Tabacchi et al. 2011a, 2011b)
Pinus sylvestris L.	dw4 = b1 + b2d^2h	dw4 = total aboveground dry weight b1 = 2.8848 b2 = 2.2080 × 10^−2 h = total tree height d = diameter at breast height	IT	Pinus sylvestris L. (Tabacchi et al. 2011a, 2011b)
Populus spp., Prunus spp., Pyrus spp., Tilia spp., Ulmus spp., Zelkova spp.	dw4 = b1 + b2d^2h + b3d	dw4 = total aboveground dry weight b1 = −1.2825 × 10 b2 = 1.1993 × 10^−2 b3=3.1553 h = total tree height d = diameter at breast height	IT	Other broadleaves group (Tabacchi et al. 2011a, 2011b)
Pseudotsuga spp.	bm = Exp(β0 + β1 In dbh)	bm = total aboveground biomass β0 = −2.2304 β1 = 2.4435 dbh = diameter at breast height	NA	Douglas – fir (Jenkins et al. 2003)
Quercus palustris Münchh., Quercus petraea (Mattuschka) Liebl., Quercus robur L., Quercus rubra L.	ln(ABW) = a + b·ln(D)	ABW = total aboveground woody biomass a = −2.4232 b = 2.4682	UK	Quercus spp., Equation n. 601 (Zianis et al. 2005)
Quercus pubescens Willd.	dw4 = b1 + b2d^2h + b3d	b1 = −7.1745 b2 = 3.3299 × 10^−2 b3 = 1.2623	IT	Quercus pubescens Willd. (Tabacchi et al. 2011a, 2011b)
Robinia pseudoacacia L.	dw4 = b1 + b2d^2h + b3d	b1 = −1.0114 × 10 b2 = 2.4042 × 10^−2 b3 = 2.2065	IT	Robinia pseudoacacia L. (Tabacchi et al. 2011a, 2011b)
Salix spp.	dw4 = b1 + b2d^2h	b1 = 9.0561 b2 = 2.1087 × 10^−2	IT	Salix spp. (Tabacchi et al. 2011a, 2011b)
Sophora japonica L.	Bs = 0.069 × D^2.54, Bb = 0.068 × D^1.89	Bs = stem, Bb = branch	CN	Liu and Li (2012)

Note: IT = Italy, ES = Spain, UK = United Kingdom, NA = North America, CN = China.

Appendix 2

Table S2. Tree species from Center for Urban Forest Research Tree Carbon Calculator (CTCC) and United States’ climate zones assigned to each species in the Bolzano, Italy inventory.

Bolzano species	CTCC climate zones	CTCC assigned species
Abies spp.	8 – Temperate Interior West	Pinus sylvestris L.
Acer negundo L.	12 – Midwest	Acer negundo L.
Acer platanoides L., Acer pseudoplatanus L.	9 – Pacific Northwest	Acer platanoides L.
Acer rubrum L.	9 – Pacific Northwest	Acer rubrum L.
Acer saccharinum L.	4 – Central Valley	Acer saccharinum L.
Aesculus spp., Toona sinensis (A. Juss.) M. Roem.	7 – Northeast	Aesculus hippocastanum L.
Alnus incana (L.) Moench, Betula pendula Roth, Corylus colurna L.	9 – Pacific Northwest	Betula pendula Roth
Carpinus betulus L., Ostrya carpinifolia Scop.	9 – Pacific Northwest	Carpinus betulus L. ‘Fastigiata’
Catalpa bignonioides Walter, Paulownia tomentosa (Thunb.) Siebold & Zucc. ex Steud.	8 – Temperate Interior West	Catalpa speciosa (Warder) Warder ex Engelm.
Cedrus spp.	2 – South Coast	Cedrus deodara (Roxb.) G. Don
Celtis australis L.	4 – Central Valley	Celtis sinensis Pers.
Cephalotaxus harringtonia (Knight ex Forbes) K. Koch	2 – South Coast	Podocarpus macrophyllus (Thunb.) Sweet
Cercis siliquastrum L.	13 – Lower Midwest	Cercis canadensis L.
Chamaecyparis lawsoniana (A. Murray) Parl., Cryptomeria japonica (L. f.) D. Don, Cupressus sempervirens L., Taxodium disticum spp.	9 – Pacific Northwest	Calocedrus decurrens (Torr.) Florin
Cornus mas L.	11 – Coastal Plain	Cornus florida L.
Diospyros kaki L. f.	4 – Central Valley	Pyrus kawakamii Hayata
Fagus spp.	9 – Pacific Northwest	Fagus sylvatica ‘atropunicea’
Fraxinus spp.	7 – Northeast	Fraxinus pennsylvanica Marshall
Ginkgo biloba L.	4 – Central Valley	Ginkgo biloba L.
Gleditsia triacanthos L.	4 – Central Valley	Gleditsia triacanthos L.
Gymnocladus dioicus (L.) K. Koch	6 – Mountains	Gymnocladus dioicus (L.) K. Koch
Hibiscus syriacus L., Tilia cordata Mill., Tilia × europaea L. (pro sp.) [cordata × platyphyllos]	9 – Pacific Northwest	Tilia cordata Mill.
Juglans nigra L.	8 – Temperate Interior West	Juglans nigra L.
Koelreuteria paniculata Laxm., Melia azedarach L.	4 – Central Valley	Koelreuteria paniculata Laxm.
Lagerstroemia indica L.	4 – Central Valley	Lagerstroemia indica L.
Laurus nobilis L.	6 – Mountains	Prunus sp.
Liriodendron tulipifera L.	3 – Inland Empire	Liriodendron tulipifera L.
Magnolia spp.	4 – Central Valley	Magnolia grandiflora L.
Metasequoia glyptostroboides Hu & W.C. Cheng, Sequoiadendron giganteum (Lindl.) J. Buchholz	1 – North and Central coast	Sequoia sempervirens (Lamb. ex D. Don) Endl.
Morus alba L.	9 – Pacific Northwest	Morus alba L.
Olea europaea L.	5 – Desert	Olea europaea L.
Photinia serrulata Lindley	9 – Pacific Northwest	Malus angustifolia (Aiton) Michx.
Picea spp.	13 – Lower Midwest	Picea pungens Engelm.
Pinus halepensis Mill.	5 – Desert	Pinus halepensis Mill.
Pinus nigra Arnold, Pinus pinea L.	13 – Lower Midwest	Pinus nigra Arnold
Pinus strobus L.	13 – Lower Midwest	Pinus strobus L.
Platanus hybrida Brot.	4 – Central Valley	Platanus hybrida Brot.
Populus spp., Salix spp.	9 – Pacific Northwest	Populus balsamifera ssp. Trichocarpa (Torr. & A. Gray ex Hook.)
Prunus avium (L.) L., Prunus laurocerasus L.	6 – Mountains	Prunus sp.
Prunus cerasifera Ehrh.	1 – North and Central coast	Prunus cerasifera Ehrh.
Pseudotsuga menziesii (Mirb.) Franco	9 – Pacific Northwest	Pseudotsuga menziesii (Mirb.) Franco
Pyrus communis L.	6 – Mountains	Pyrus sp.
Quercus palustris Münchh., Quercus petraea (Mattuschka) Liebl.	12 – Midwest	Quercus palustris Münchh.
Quercus robur L., Quercus rubra L.	12 – Midwest	Quercus rubra L.
Robinia pseudoacacia L., Sophora japonica L.	1 – North and Central coast	Robinia pseudoacacia L.
Styraciflua L.	4 – Central Valley	Liquidambar styraciflua L.
Taxus baccata L.	2 – South Coast	Podocarpus macrophyllus (Thunb.) Sweet
Tilia americana L.	9 – Pacific Northwest	Tilia americana L.
Tilia tomentosa Moench	7 – Northeast	Tilia tomentosa Moench
Ulmus spp.	6 – Mountains	Ulmus pumila L.
Wisteria sinensis (Sims) DC.	4 – Central Valley	Gleditsia triacanthos L.
Zelkova carpinifolia (Pall.) K. Koch	4 – Central Valley	Zelkova serrate (Thunb.) Makino