ARTICLE pubs.acs.org/est
Influence of 150 Years of Land Use on Anthropogenic and Natural Carbon Stocks in Emilia-Romagna Region (Italy) Riccardo Scalenghe,† Francesco Malucelli,*,‡ Fabrizio Ungaro,§ Luca Perazzone,|| Nicola Filippi,‡ and Anthony C. Edwards^ †
Universita di Palermo, Italy Regione Emilia-Romagna, Italy § Consiglio Nazionale delle Ricerche, Sesto Fiorentino, Italy Studio Corona, Torino, Italy ^ Scottish Agricultural College, Craibstone, Aberdeen, United Kingdom
)
‡
bS Supporting Information ABSTRACT: Changes in land use/cover that are commonly associated with urbanization can dramatically influence the amount, chemical form, and spatial distribution of carbon (C) stocks. Measured values and relative literature for composition of natural and anthropogenic materials have been compiled. These data are used in conjunction with land cover statistics and expert assessment of building design to calculate C stocks associated with 150 years of land use change and development for an area of the Po River Valley, Northern Italy. Using 4 time periods (1853, 1954, 1976, and 2003), we demonstrate that the C stocks within this densely populated area have undergone considerable modification. A 52% increase in population density has been associated with an increase in the proportion of total C stored in anthropogenic stocks from 0.2% to 6%; this has been accompanied by a one order of magnitude increase in the carbon emission per capita per unit area. These changes have also been accompanied by a major shift in stocks from organic to inorganic forms of C.
’ INTRODUCTION Land use change into food production or urbanization has been a constant feature of human development, and is usually associated with major changes in ecosystem functions.1 For example, one-quarter of the annual global loss of terrestrial carbon (C) (∼0.1 Eg C, or 104 Mg C)2 can be attributed to three large geopolitical regions that together cover ten million square kilometers: the coterminous United States (0.30.6 Mg C ha1 year1),3 China (0.20.3 Mg C ha1 year1),4 and Europe (0.10.2 Mg C ha1 year1).5 The European landscape has been heavily modified through urbanization,6 which has involved a physical “sealing”, resulting in a decreased capacity for the soil to act as a C sink.7 In North America, on the other hand, researchers have demonstrated that the C storage capacity of urban areas is of a magnitude similar to that of forests. However, this point very much depends upon the density of development, with values for urban, periurban, and rural land covers being quoted as up to 420, 160, and 250 Mg C ha1, respectively.8 Here we collate information on soil and land use change, together with population statistics (cadastral and historical census), for one area of Northern Italy, and combine it with appropriate values of C contents in order to provide estimates for the change in C stocks. The objective is to develop and present an integrated approach to describing the evolution of both natural and r 2011 American Chemical Society
anthropogenic C (organic and inorganic) stocks over time. In addition, the effects of physical factors on CO2 emissions, such as soil sealing, are estimated. The approach adopted is based on data on soil and land use change, and on cadastral and historical census records starting in the 19th century and covering a period of 150 years of rapid development. The time periods selected are considered to be representative of four key phases of change which have taken place during the last 150 years. In addition to the information that acts as a source of reference, the novelty of the methodology also enables the identification of areas of uncertainty and gaps in measuring the phenomenon of sprawl and associated soil sealing. Study Area. The study region is situated within the 71 000 km2 of the Po River watershed, which flows in an easterly direction and ultimately drains one-quarter of the Italian national territory and is currently home to about 18 million inhabitants, with an average population density of 300 people km2. Today, the entire Po region produces 40% of the Italian gross domestic product, one-third of the national industrial and arable production, Received: December 2, 2010 Accepted: May 16, 2011 Revised: May 11, 2011 Published: May 24, 2011 5112
dx.doi.org/10.1021/es1039437 | Environ. Sci. Technol. 2011, 45, 5112–5117
Environmental Science & Technology
ARTICLE
Table 1. Synopsis of Carbon Stocks in the Po Plain of the Emilia-Romagna region (∼11 600 km2; data in Tg of C)a 1853
1954
1976
2003
0.98þþ/##
1.03þþ/##
1.09þþ/###
þþ/##
1.01þþ/###
þþþ/##
2.80þþþ/###
þ/#
1.30þþþ/####
þ/#
2.60þ/#
þþ/##
above soil 0.12þþ/##
forest
þ/# b
orchards
0.40
0.61
þ/##
urban (concrete)
þþ/##
0.27
roads (asphalt)
1.09
þ/#
n.a.
roads, infrastructures (concrete) industrial (concrete)
þþ/## b
c,d
0.51
þ/#
n.a. ∼0
1.01 þ/#
þþ/##
1.25 2.23 1.04 2.07
0.36
1.43
2.36þþ/##
70.5þþ/### 69.9þþ/###
70.5þþ/### 68.7þþ/###
67.6þþþþ/### 65.5þþþ/###
soil organic carbon (TOC) e carbonates (TIC) f
71.0þþ/### 70.5þþ/###
per capita carbon g concrete þ asphalt (Mg)
170 and >60 head km2, respectively, for pigs and cows), and in the southeast (densities being >4600 and >180 head km2, respectively, for poultry and rabbit husbandry) (source Italian National Institute of Statistics, ISTAT). Industry, along the Via Emilia, is concentrated in Parma, Modena, and Bologna (food and mechanical).
The ceramic sector is based in Faenza and Sassuolo. Small enterprises are widely distributed.
’ EXPERIMENTAL METHODS The C stocks have been calculated for (i) the time when these territories, which were previously dominated by the Papal State and other smaller Duchies, became part of Kingdom of Italy, and were populated by less than two million inhabitants (year 1853); (ii) the post-WWII reconstruction period (year 1954); (iii) between the first (1973) and the second (1979) energy crisis periods during the 1970s (year 1976); (iv) the most recent decade (year 2003), when the population has doubled, when compared to the first Italian census made in 1861. Carbon Stock Assessment. The assessment of the individual C pools considered in this work was based on the quality and availability of data. We collated C stocks for 11 600 raster cells, 1 km2 each by size;11 then C density (Mg ha1) of natural and anthropogenic terrestrial ecosystems were superimposed for each cell; their sum estimates natural (organic and inorganic) and anthropogenic C stocks. We started estimating current SOIL organic and inorganic (carbonates) C stocks for a reference depth of 30 cm. We then distinguished a TREE layer12,13 accounting C stocks in both forest and orchards (including rhizosphere), an ASPHALT layer (measuring C in materials of high-molecular weight hydrocarbons obtained by refining petroleum), and three CEMENT layers, specifically in urban buildings, in industrial buildings, and in roads and infrastructures (measuring C in ordinary portland cement, OPC, obtained by heating limestone with clay).14 The remarkable regional variability forced us to exclude additional details such as a WOOD layer: C storage in domestic products should include at least 25 kg C for each wooden window, 5 kg C m2 of wooden flooring, and house furniture and its contents for 1230 Mg C household1. The average lifetime of materials was not included, due to the complexities of calculating these figures for the large study area, although this 5113
dx.doi.org/10.1021/es1039437 |Environ. Sci. Technol. 2011, 45, 5112–5117
Environmental Science & Technology
ARTICLE
Figure 1. Ratio between anthropogenic stocks over soil carbon stocks in the Po plain of the Emilia-Romagna Region (∼11 600 km2) and the SW neighborhood of the city of Ferrara (∼1200 km2). Histograms indicate hectares of floodplain sorted by ratio classes, where 1 indicates the same amount of the two C pools inside each 1 ha cell (i.e., carbon is equally distributed between soil [Co] and cement [Cc]): hectares are on the ordinate (log scale), [Cc]/[Co] in abscissa. The ratio between anthropogenic over soil C stocks can be used as a proxy of the urban sprawl that characterizes the EmiliaRomagna plain. The enlarged area is the vicinity of Ferrara (less than 200 thousands inhabitants, centroid in 44500 N, 11370 E). A corresponding GoogleEarth file is available at http://geo.regione.emilia-romagna.it/gstatico/documenti/cartpedo/cc_vs_co/cc_vs_co.kml.
might make an interesting additional future comparison, as wood-based products, such as newspapers, can last at least some months, while magazines can last half a year, books and furniture can last decades, and part of buildings can last more than half a century.12 Soil and Vegetation. Topsoil (030 cm) C (organic and inorganic) densities (Mg ha1) were assessed at a regional scale for the whole Emilia-Romagna plain. We used the regional soil database, compiled between 1981 and 2005 by the Regional Soil Survey. The database contains nearly 19 000 entries; average sampling density is about 1.6 observations per km2, but this figure does vary. For each entry, the following analytical data are available (030 cm): sand, silt, and clay fraction (percent, USDA limits), coarse fragment content (percent), soil organic carbon (SOC) content, and total carbonates (percent). A Scorpan
Kriging approach15 was implemented, which combines the trend component of soil properties, as derived from the 1:50 000 soil map, with geostatistical modeling of the locally varying but spatially correlated stochastic component. The trend component is described in terms of varying local means, calculated taking into account soil type and dominant land use. The resulting concentration of SOC and contents of sand, silt, and clay were used in the calculation of topsoil SOC stocks, using a set of locally calibrated pedotransfer functions to estimate bulk density (Mg m3).11 The same approach was used to calculate the soil C stored as carbonates. The estimated C stocks (Table 1) refer to the year 2003, and the figures have been adjusted to account for the present degree of soil sealing; SOC stocks were also assessed to provide a continuous map coverage of the area; this was then corrected considering the soil/nonsoil ratio associated with each 5114
dx.doi.org/10.1021/es1039437 |Environ. Sci. Technol. 2011, 45, 5112–5117
Environmental Science & Technology
ARTICLE
Table 2. Estimated CO2 Emission in Atmosphere Due to Mineralization of Soil Organic Matter and Concrete Production in Po Plain (Emilia-Romagna Region, 11 600 km2)a post unificatione
post-WWII reconstruction f
toward globalizationg
whole periodh
soil sealing (Tg CO2-equiv)
1.84
∼0
10.64
12.48
concrete production (Tg CO2-equivb) total net emissions (soil þ concrete) (Tg CO2-equiv)
25.57 27.40
38.18 38.18
23.70 34.34
87.45 99.92
net emissions per year (Tg CO2-equiv year1)
0.27
1.74
1.27
0.66
population increase per year (person year1)
9,019
14,409
5,926
9,250
individual emission ratio (ΔCO2/Δpopulationc Mg person1) average annual per capita emissions d (Mg CO2 year1 person1)
29
120
215
71
0.16
0.70
0.49
0.30
a
Figures are referred to time intervals of different length, and are based on data presented in Table 1 (CO2-equivalents are calculated using a conversion factor of 3.67). b Every tonne of clinker concrete induces an emission of CO2, ranging between 0.80 and 1.04 tonnes (the value of 0.92 tonnes was used for calculations).11 c Δpopulation: 19541853, 42%; 19761954, 13%; 20031976, 6%; 20031853, 52%; the individual emission ratio is calculated as net emission per year divided by population increase per year. d Average annual per capita emissions are calculated assuming linear relationships between each couple of points in time. e Post unification (19541853), population of year 1853 is 1.29 million.9 f Post-WWII reconstruction (19761954), population of year 1954 is 2.22 million.9 g Toward globalization (20031976), population of year 1976 is 2.54 million.9 h Whole period (20031853), population of year 2003 is 2.70 million.9
land use class. The C stocks for the previous time steps (1853, 1954, and 1976) were calculated considering the changes in land use in terms of soil sealing development, as determined by the change in area of structures associated with human settlements. We started from the assumption that sealing resulted in the complete displacement and loss of the surface 030 cm of soil, while any of the organic C stored was ultimately lost to the atmosphere as CO2. These soils have permanently lost their capacity to act as a C sink. The inorganic C stocked as carbonates is assumed to be lost to the wider environment as we cannot discriminate if the soil was just buried onsite or physically removed from the previous location. Above ground, C stocks in trees and orchards were estimated using land use information at the four time steps. For orchards, it was assumed an average figure of 5 Mg ha1, although in northern latitudes the figure is different, >7 Mg ha1; for forests and urban forests an average figure of 80 Mg ha1was used.16,17 The below-ground biomass stock of C has been taken into account only in the “bulk” stock of the tree, where we consider the epigean and hypogean stocks together. Regarding soil accounting, we have considered only the humified organic matter (according to the Soil Organic Matter definition praxis). We assumed that the plant root system of annual (and perennial) crops will be rapidly decomposed and incorporated in the SOM estimate.18 Concrete and Asphalt C Stocks. Data from land use maps (1:25 000 scale) for the years 1853, 1954, 1976, and 2003 and cadastral data9 were used to infer building volumes. For each cadastral unit, the built-up areas at any time steps and the average number of floors are available, allowing us to make an estimation of the built-up volumes at any time step for each cadastral unit. The age of construction was given, and this allowed for a different composition of the building materials (i.e., concrete, brick, and other material), which were assigned to the 7 age classes considered in the cadastral database (1991). A different mass of concrete for unit of built volume was assumed for each class: pre-1945, 55 kg m3; 19461961, 65; 19621971, 69; 19721981, 71; 19821991, 72; 19921999, 80; and post 2004, 81 kg m3. Carbon stocks in concrete and other human infrastructures relies on the following set of simplifying basic assumptions. (1) Buildings are calculated volumetrically from the floor level (level 0), i.e., 0 m above ground level. (2) Volumes of
the built-up areas were assessed using an average number of floors (i.e., 2.5) and an average height for each floor (i.e., 3.0 m), obtained from cadastral data. (3) The volume of the structures (i.e., walls, floor, etc.) and the percentage of concrete used in house building was calculated according to the cadastral data and the local building techniques. Similarly, the content of concrete in structures has been defined according to the local building techniques using an average percentage (in volume) of 25% in walls and 40% in slabs and floors. (4) An average C content for mass unit of concrete (0.079 kg C kg1) was calculated after sampling and measuring the C contents of different building typologies. We calculated the cubic meters of asphalt using the development of the road network over time and used measurements of C contents of asphalt samples (average 0.169 kg C kg1). For each road typology (in terms of width and thickness of the paved layer), standard sizes were assumed. Carbon stored in paved roads is likely to be underestimated, due to the resolution constraints of the available land use maps. In the past decade, 33.5% of the concrete production was used in infrastructures and 66.5% was used in housing and public buildings.19,20 We assigned an equivalent quota of cement to road infrastructures, which was evenly distributed over the whole RER roads network. Total carbon was measured by combustion, using an NA 2100 Protein Nitrogen Analyzer (Thermoquest CE Instruments, Milan, Italy) on 25 samples of masonry, cement and bricks, and asphalt from private driveways, freeways, two-lane highway, and high-quality dual carriageways (100 km radius around 44 N, 10 E) covering a time span from 1860 to 2005. Average percent C concentrations ((S) by dry weight are: asphalt 10.2 ( 2.7 (after 2000, 7 ( 0.5), roads in cement 7.8 ( 1.0, structures in cement 5.3 ( 3.0, masonry 3.6 ( 1.1 (before 1900
