Elsevier

Landscape and Urban Planning

Volume 120, December 2013, Pages 70-84
Landscape and Urban Planning

Research paper
A new approach to quantify and map carbon stored, sequestered and emissions avoided by urban forests

https://doi.org/10.1016/j.landurbplan.2013.08.005Get rights and content

Highlights

  • Species composition, basal area and development patterns influence carbon.

  • Carbon storage was high in Sacramento due to native oaks with high wood density.

  • Avoided emissions were high in Los Angeles due to trees shading multiple buildings.

  • California's urban forests account for 2 percent of C stored by forests statewide.

  • California's urban forests account for 12 percent of C sequestered by forests statewide.

Abstract

This paper describes the use of field surveys, biometric information for urban tree species and remote sensing to quantify and map carbon (C) storage, sequestration and avoided emissions from energy savings. Its primary contribution is methodological; the derivation and application of urban tree canopy (UTC) based transfer functions (t C ha−1 UTC). Findings for Los Angeles and Sacramento illustrate the complex role of regional and local determinants. Although average tree density and size were substantially greater in Los Angeles, the mean C storage density (8.15 t ha−1) was 53 percent of Sacramento's (15.4 t ha−1). In Sacramento, native oaks with very high wood densities (815 kg m−3) accounted for 30 percent of total basal area. In Los Angeles, the most dominant taxa had relatively low wood densities (350-550 kg m−3). The inclusion of relatively more wooded land in the Sacramento study area may partially explain higher C storage levels. In Los Angeles, where development is relatively dense, 14 percent of all trees surveyed shaded more than one building compared to only 2 percent in Sacramento. Consequently, the transfer function for avoided emissions in Los Angeles (2.77 t ha−1 UTC yr−1) exceeded Sacramento (2.72 t ha−1 UTC yr−1). The approach described here improves C estimates and increases the resolution at which C can be mapped across a region. It can be used to map baseline C storage levels for climate action planning, identify conservation areas where UTC densities are highest and determine where opportunities for expanding UTC are greatest.

Introduction

Carbon stored in urban forests is not typically included in national, statewide and regional inventories of greenhouse gas (GHG) emissions and sinks, perhaps because cities make up a small fraction of total land area and intensive management of trees can release large amounts of GHGs (Ryan et al., 2010). Although urbanized areas account for 3 percent of total land area and 81 percent of total population in the US (Cox, 2012), Heath, Smith, Skog, Nowak, and Woodall (2011) found that trees in US cities sequester about 14 percent of the amount of carbon (C) sequestered by US forests. Although relatively small in stature, urban forests store substantial amounts of carbon. Accurate quantification and mapping of these stocks is fundamental to the inclusion of urban forestry in local climate action plans and carbon offset markets. This study combines field surveys, biometric information for urban tree species, Geographic Information System (GIS) data sets and remote sensing of urban tree canopy (UTC) to quantify and map C stored, sequestered and emissions avoided in two urban forests. By incorporating age-related differences among census block groups that influence tree species composition and stand structure, this approach improves C estimates and increases the resolution at which C can be mapped across a region.

In California, a cap and trade program began in 2012 to drive investment in cleaner fuels and more efficient energy use. The Compliance Offset Protocol for Urban Forest Projects is one of four protocols that will guide carbon offset projects (California Air Resources Board, 2011). The current protocol focuses on offsets from tree planting projects, but the Climate Action Reserve (2010) is exploring development of a second protocol based on field surveys that are linked with urban tree canopy (UTC) mapped with remote sensing. Urban tree canopy (UTC), defined as the layer of leaves, branches and stems that cover the ground when viewed from above, is a metric used to quantify the areal extent of the urban forest (Raciti et al., 2006). Using UTC as a metric for C storage has the advantage of being readily measured and tracked over time at high spatial resolution and with increasing accuracy as technologies develop. Mapping the spatial distribution of existing and additional stocks can provide the basis for quantifying and reporting changes in C storage, as well as for planning and managing urban forests to increase C stocks. For example, maps can be used to locate areas with the greatest potential for increasing tree canopy through tree planting, as well as areas where the foremost need is to preserve C stored in existing canopy (Escobedo et al., 2010, Myeong et al., 2006).

Broadly speaking, transfer function is a term used to describe the transfer of data for a particular “study site” to a “policy site” for which little or no data exist (Brookshire and Neill, 1992, Downing and Ozuna, 1996). In this study, transfer functions are defined as field plot-based measures of C per hectare UTC (t ha−1 tree canopy cover) that are aggregated and applied to a region by land use class.

Two studies have derived and applied UTC-based transfer functions. Nowak and Greenfield (2010b) calculated mean UTC storage (333.7 t ha−1 UTC) and sequestration (11.0 t ha−1 UTC) densities using 2001 NLCD imagery and field sampling in 17 US cities. Their values may be relatively high because the classification process was found to significantly underestimate UTC (Nowak & Greenfield, 2010a). Strohbach and Haase (2012) used high resolution orthophotos to classify UTC and intensive UFORE field sampling to estimate aboveground carbon and UTC densities for 19 land use classes. They compared C estimates derived using UTC and land use class alone and found that UTC-based estimates provided higher accuracy, greater precision and improved spatial detail. The UTC-based approach eliminated variation in UTC within land use classes, an important source of error. Because field plot sampling did not fully capture the extent of UTC for each land use class, land use based C storage estimates had large standard errors in areas where UTC was highly heterogeneous, such as town centers.

To derive UTC-based transfer functions, C storage, sequestration and avoided emission values are calculated for trees in each plot and divided by the plot's UTC. Plot data are aggregated by land use class and descriptive statistics are applied to determine sample means and standard errors. Different values reflect different stand structures and dynamics that influence C. For instance, the C storage transfer function for a hectare of UTC in an old residential neighborhood will be relatively high when the stand consists of mature oaks (Quercus sp.) and a lush understory. In contrast, the transfer function for a hectare of UTC in a new residential area will be lower when the stand is characterized by juvenile pear (Pyrus sp.) trees with a sparse understory.

The value of a transfer function reflects species composition and attributes of stand structure. Stand attributes, such as the vertical layering of woody biomass in strata, tree density and basal area influence the amount of biomass per hectare UTC and the resulting value of a transfer function. Species is important because of its influence on the tree's biomass and partitioning into roots, bole, branches, stems and foliage. Also, the amount of biomass converted into carbon depends on the species’ wood density, which can vary widely among species.

The transfer function for each land use class is transferred to the UTC delineated from imagery for the corresponding land use. Using GIS capabilities, C values are mapped and summed based on the amount of UTC in each land use class. These maps provide spatially explicit information on the distribution of urban forest C for planning and management purposes.

The magnitude of C stored, sequestered and avoided emissions by urban forests depends on regional and local determinants. Regional context influences climate, soil, potential vegetation and urban morphology (Nowak et al., 1996, Sanders, 1984). Desert cities can have lower overall C storage densities than cities in temperate climates because unmanaged open space in deserts contains less biomass than in forest biomes such as Atlanta and Baltimore (Yesilonis & Pouyat, 2012). Older, densely developed cities can contain less growing space for urban vegetation that stores less C than sprawling cities (Zipperer, Sisinni, Pouyat, & Foresman, 1997).

Local determinants of urban forest C storage include species composition, age structure, stand density and tree management practices, as well as neighborhood age and land use (Zhao, Kong, Escobedo, & Gao, 2010). Carbon storage typically decreases as impervious surfaces and land development intensity increases. In Leipzig for example, land uses with the highest C storage densities were cemeteries, parks and single-family residential areas, while commercial, industrial and multifamily residential had lower C densities (Strohbach & Haase, 2012).

Each urban forest can be viewed as a mosaic of neighborhood stands with structural features that reflect historic changes in species preferences, planting practices, land development patterns, and tree conservation and planting policies (Berland, 2012, Conway et al., 2011, Palmer, 1984). The importance of neighborhood age on urban forest stand density, species composition and structure has been well demonstrated. Lowry, Baker, and Ramsey (2012) found that among several physical factors, neighborhood age was the most influential factor explaining tree canopy abundance. In Baltimore, Maryland the abundance of neighborhood tree canopy increased with neighborhood age to about 45–50 years, then decreased (Grove et al., 2006). This suggests that forest and housing stocks followed a parallel inverted U relationship that traced periods of appreciation and depreciation. However, in desert cities the abundance and diversity of vegetation was found to decrease as neighborhoods aged (Hope et al., 2003, Martin and Stabler, 2004). In US cities, the majority of trees and potential tree planting sites are in low density residential land uses, so their age-related stand structure and dynamics are especially important (McPherson & Rowntree, 1993).

This study extends research cited above by incorporating age-related differences that influence species composition, stand structure and C storage in low density residential areas. The derivation and application of transfer functions is demonstrated for urban forests in Los Angeles and Sacramento, California. Relations between differences in transfer functions and causal factors such as species composition, basal area and tree densities are discussed. Sources of error and uncertainty are estimated, and the relevance of this approach for California's emerging cap and trade market is discussed.

Terms used in this study are defined as follows:

  • Carbon storage: C accumulated in the aboveground and belowground biomass of trees over many years (excludes C stored in the soil).

  • Carbon sequestration: change in C storage in aboveground and belowground biomass that result from tree growth during a single growing season.

  • Avoided carbon emissions: C equivalent emission reductions from electricity generation and natural gas combustion from urban forest effects on annual building energy use for space heating and cooling.

  • Carbon density: C values divided by respective total land areas. Because the area of each region differs, comparisons between cities and land uses are made based on C density values for storage (t ha−1), sequestration (t ha−1 yr−1) and avoided emissions (t ha−1 yr−1).

  • UTC density: C values divided by respective total UTC areas. These transfer functions are rendered comparable in terms of UTC C density values for storage (t ha−1 UTC), sequestration (t ha−1 UTC yr−1) and avoided emissions (t ha−1 UTC yr−1).

Section snippets

Study areas

The two study areas cover 1022 km2 in the City of Los Angeles, CA and 1732 km2 in the Sacramento, CA metropolitan area (Fig. 1). The City of Los Angeles (latitude: 34°06′36″ N, longitude: 118°24′40″ W) lies within one of the largest metropolitan areas in the United States and has a land area of 1225 km2. The city's population is 3.8 million and there are 15 council districts and 86 neighborhood councils. Topographic gradients are small in the coastal areas and inland valleys; however, within the

Urban tree canopy

Overall classification accuracy for Los Angeles was 88.6 percent based on a pixel by pixel comparison. The accuracy for classifying existing UTC was 74.3 percent. Not surprisingly, TCC was most often misclassified as irrigated grass (13 percent), and vice versa (17 percent). Factors that affected the mapping accuracy included the treatment of the shadowed area and minimum mapping units during digitizing. For the Sacramento study area, there was 90 percent agreement between direct interpretation

Comparison between Los Angeles and Sacramento

Differences in the amounts of C stored, sequestered and avoided emissions by the urban forests of Los Angeles and Sacramento illustrate the complex influence of regional and local determinants. For example, average tree density (49 vs. 40 ha−1) and size (24.4 vs. 19.8 cm dbh) was greater in Los Angeles, but mean C storage density was 53 percent of Sacramento's. On average, Sacramento's trees stored (329 vs. 178 kg per tree) and sequestered (23.0 vs. 9.6 kg per tree annually) nearly twice the amount

Conclusions

By sequestering C city trees mitigate climate change and by modifying urban climate and conserving energy they are an effective adaptation strategy (Stone, 2012). Urban areas in California encompass 5 percent of the total land and support 95 percent of the population. Although the California Air Resources Board (2008) has adopted a tree planting project protocol for fungible offsets, its official statewide estimate of GHG emissions and sinks does not include urban forests. After applying the

Acknowledgements

We would like to thank the City of Los Angeles, Sacramento Metropolitan Air Quality Management District and the Sacramento Tree Foundation for their support, as well as the more than 90 volunteers and staff who contributed to the tree survey and other aspects of the study. David Nowak and colleagues with the U.S. Forest Service's Northern Research Station were very kind to share their information from the UFORE analyses, and Lorraine Weller (UC Riverside) and Antonio Davila (USFS) led field

References (79)

  • S. Myeong et al.

    A temporal analysis of urban forest carbon storage using remote sensing

    Remote Sensing of Environment

    (2006)
  • D.J. Nowak et al.

    Carbon storage and sequestration by urban trees in the USA

    Environmental Pollution

    (2002)
  • D.J. Nowak et al.

    Measuring and analyzing urban tree cover

    Landscape and Urban Planning

    (1996)
  • J.F. Palmer

    Neighborhoods as stands in the urban forest

    Urban Ecology

    (1984)
  • R.A. Sanders

    Some determinants of urban forest structure

    Urban Ecology

    (1984)
  • Y.-C. Wang et al.

    Air quality enhancement zones in Taiwan: A carbon reduction benefit assessment

    Forest Policy and Economics

    (2012)
  • J. Yang et al.

    The urban forest in Beijing and its role in air pollution reduction

    Urban Forestry and Urban Greening

    (2005)
  • M. Zhao et al.

    Impacts of urban forests on offsetting carbon emissions from industrial energy use in Hangzhou, China

    Journal of Environmental Management

    (2010)
  • E. Aguaron et al.

    Comparison of methods for estimating carbon dioxide storage by Sacramento's urban forest

  • A. Berland

    Long-term urbanization effects on tree canopy cover along an urban–rural gradient

    Urban Ecosystems

    (2012)
  • D.S. Brookshire et al.

    Benefit transfers: Conceptual and empirical issues

    Water Resources Research

    (1992)
  • M.A. Cairns et al.

    Root biomass allocation in the world's upland forests

    Oecologia

    (1997)
  • California Air Resources Board

    Climate change scoping plan

    (2008)
  • California Air Resources Board

    Compliance offset protocol: Urban forest projects

    (2011)
  • L. Chaparro et al.

    Ecological services of urban forest in Barcelona

    (2009)
  • Climate Action Reserve

    Urban forest project protocol. Version 1.1

    (2010)
  • W. Cox

    New US urban area data released

    (2012)
  • T.J. Gillespie et al.

    A time series of urban forestry in Los Angeles

    Urban Ecosystems

    (2012)
  • J.M. Grove et al.

    Data and methods comparing social structure and vegetation structure of urban neighborhoods in Baltimore, Maryland

    Society and Natural Resources

    (2006)
  • L.S. Heath et al.

    Managed forest carbon estimates for the US greenhouse gas inventory 1990–2008

    Journal of Forestry

    (2011)
  • D. Hope et al.

    Socio-economics drive urban plant diversity

    Proceedings of the National Academy of Sciences

    (2003)
  • B. Husch et al.

    Forest mensuration

    (1982)
  • J.C. Jenkins et al.

    Comprehensive database of diameter-based biomass regressions for North American tree species

    (2003)
  • J.C. Jenkins et al.

    National-scale biomass estimators for United States tree species

    Forest Science

    (2003)
  • R. Kohavi et al.

    Glossary of terms

    Machine Learning

    (1998)
  • M. Lefsky et al.

    Volume estimates of trees with complex architecture from terrestrial laser scanning

    Journal of Applied Remote Sensing

    (2008)
  • H. Lieth et al.

    Modeling the primary productivity of the world. Primary productivity of the biosphere

    (1975)
  • J.H. Lowry et al.

    Determinants of urban tree canopy in residential neighborhoods: Household characteristics, urban form, and the geophysical landscape

    Urban Ecosystems

    (2012)
  • L.J. Markwardt et al.

    Strength and related properties of woods grown in the United States

    (1935)
  • Cited by (0)

    View full text