Above-ground carbon storage by urban trees in Leipzig, Germany: Analysis of patterns in a European city

Abstract

Many aspects of global change, including carbon dioxide emissions, have been attributed to urban areas. On the other hand, cities have been found to provide valuable ecosystem services such as carbon storage. The aim of this study is to estimate the above-ground carbon storage in trees in the central European city of Leipzig and produce spatially explicit carbon storage maps. We used stratified random sampling across 19 land cover classes using 190 sample plots to measure carbon storage. In addition, we derived canopy cover from color-infrared orthophotos using an object-oriented approach and Random Forest machine learning. Finally, we apply an error assessment method that includes sampling error, but also uncertainty stemming from allometric equations, and that so far has only been applied to rural forests. The total above ground carbon stock of Leipzig was estimated using both land cover and canopy cover, which was more laborious than just using land cover but reduced the standard error. Canopy cover was approximately 19% of the city area. Leipzig's above-ground carbon storage was estimated to be 316,000 Mg C at 11 Mg C ha⁻¹. The distribution of carbon storage across the city showed the highest values at intermediate urbanization levels. Carbon storage in the city of Leipzig was in the lower range compared to cities in Europe, Asia and the USA, and our results indicate that great care should be taken when transferring values between cities. We provide spatially explicit and detailed maps of above-ground storage that can contribute to ecosystem services assessments.

Introduction

The perspective on urban areas is often shifted towards the negative impacts they have on ecosystems, from local to global scale, and in fact, many aspects of global change have their origins there (Grimm et al., 2008). For example, a high proportion of the greenhouse gas carbon dioxide is emitted from urban areas (Svirejeva-Hopkins, Schellnhuber, & Pomaz, 2004), with most emissions being related to the activities of urban dwellers that require fossil fuel, like industrial production, traffic, heating, or cement production (Solomon et al., 2007), but also to the disturbance and alteration of soils and vegetation through urbanization (Churkina, 2008, Imhoff et al., 2004, Solomon et al., 2007). Cities traditionally have been viewed in opposition to adjectives like “natural”, “pristine”, or “wilderness” (Benton-Short & Short, 2008).

Although cities might appear as merely lifeless concrete deserts, they can contain rich and diverse ecosystems (Benton-Short & Short, 2008). Even highly modified ecosystems such as parks generate a variety of ecosystem services for the benefit of urban dwellers (Bolund and Hunhammar, 1999, Niemela et al., 2010, Snep and Opdam, 2010). Keeping in mind that most humans today live in urban areas (United Nations Population Division, 2008) and that urban areas are “hot spots of global change” (Grimm et al., 2008), the importance of a better understanding of these systems becomes obvious. Studying ecological functions and how they are modified by humans, allows for the evaluation of the ecological performance of urban areas that differ in building density and design, or amount and type of green space (Alberti, 1999, Pauleit and Duhme, 2000, Whitford et al., 2001). Attributing value to urban ecosystem functions using ecosystem services, can help planners to conserve and (re)create urban areas that are more sustainable and promote human well-being (Bolund and Hunhammar, 1999, Niemela et al., 2010, Snep and Opdam, 2010, Tratalos et al., 2007). However, detailed information on ecosystem services in cities is still scarce (Niemela et al., 2010).

Trees provide important ecosystem services, one of which is carbon storage. Urban forests, defined as the “sum of all woody and associated vegetation in and around dense human settlements” (Miller, 1997 in Konijnendijk, Ricard, Kenney, & Randrup, 2006), have been found to store significant amounts of carbon in cities in the USA (Churkina et al., 2010, Hutyra et al., 2011, Nowak, 1994, Nowak and Crane, 2002, Rowntree and Nowak, 1991), Germany (Kändler, Adler, & Hellbach, 2011), the UK (Davies, Edmondson, Heinemeyer, Leake, & Gaston, 2011), Spain (Chaparro & Terradas, 2009), Korea (Jo, 2002), China (Zhao, Kong, Escobedo, & Gao, 2010) and Australia (Brack, 2002). Humans are largely responsible for altering and designing the urban forest and their complex interaction with biophysical processes creates very distinct patterns of carbon storage. For example, in arid regions, urbanization can increase carbon stocks: the city has a denser urban forest than the wildlands due to urban dwellers efforts in planting and irrigating shade trees. In humid areas, urbanization often involves clearing forests and therefore reduces carbon stocks (Golubiewski, 2006, Imhoff et al., 2004).

Factors such as age (Rowntree & Nowak, 1991), composition (Nowak, Stevens, Sisinni, & Luley, 2002), and history (Brack, 2002, Nowak, 1993) of the urban forest also influence carbon storage and can differ between and within cities. Increasing human population size and density of cities has been shown to result in decreasing green space density (Fuller & Gaston, 2009). Alberti and Hutyra (2009) hypothesize that carbon storage increases with distance from the urban core, for the humid and densely forested Seattle region in the pacific northwest of the USA. Davies et al. (2011) found that above-ground carbon storage in trees in Leicester, UK, exceeded current national estimates. However, considering the limited body of literature, the above-ground carbon storage of central European cities remains largely undocumented.

In this paper we studied the above-ground carbon stored in trees in the central European city of Leipzig, Germany. Our main objectives were: (1) Estimating the above-ground tree carbon storage in Leipzig and comparing it to existing published studies. (2) Mapping the carbon storage and documenting the variability across different land covers and along different measures of urbanization intensity. (3) Evaluating the advantages and disadvantages of land cover- vs. canopy cover-based estimation. (4) Finally, we discuss how our results can be integrated into an ecosystem service assessment.