Soil organic carbon stocks in verges and urban areas of Flanders, Belgium I. Mestdagh*, S. Sleutel†, P. Lootens*, O. Van Cleemput‡ and L. Carlier* *Department of Crop Husbandry and Ecophysiology, Agricultural Research Centre, Merelbeke, Belgium, †Laboratory of Soil Management and Soil Care, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium, and ‡Laboratory of Applied Analytical and Physical Chemistry, Faculty of Bioscience Engineering, Ghent University, Gent, Belgium
Abstract Soil organic carbon (SOC) stocks in different terrestrial ecosystems have already been calculated. However, SOC stocks of grassy verges and grass-covered urban areas have never been reported. In this study the total grass-covered area of verges and urban areas in different agro-pedological regions of Flanders, Belgium was calculated and SOC stocks estimated. The total grassy area of verges along roads, waterways and railways was calculated to be 18 027 ha, 9530 ha for gardens and parks and 2360 ha for recreations areas. Total SOC stock for grass-covered verges, extrapolated to a depth of 1 m, was estimated to be 3520 kt SOC. Total SOC stock for grassy vegetation in urban areas was 1738 kt SOC. These total SOC stocks equate, respectively, to 0Æ10 and 0Æ05 of the total SOC stock in Flemish grassland. Keywords: verges, roads, waterways, railways, urban area, soil organic carbon stocks, Flanders
Introduction Under the Kyoto Protocol, some parties may choose to calculate total soil organic carbon (SOC) stocks for their different terrestrial ecosystems (UNFCCC, 1997). However, SOC stocks in verges along roads, waterways and railways, and in urban areas, have not previously been reported. Nevertheless, as most verges and vegetation in urban areas are covered with grass, they could store considerable carbon.
Correspondence to: I. Mestdagh, Agricultural Research Centre, Department of Crop Husbandry and Ecophysiology, Burg. Van Gansberghelaan 109, B-9820 Merelbeke, Belgium. E-mail:
[email protected] Received 7 September 2004; revised 5 January 2005
Verges in Flanders (Belgium) have a high nature conservation value. Large biodiversity of flora and fauna are represented (Zwaenepoel, 1993) because they form a refuge for many species from typical grassland and forest ecosystems and form a unique habitat. However, they are also influenced by human activities, such as disturbance of the soil profile for maintenance of roads and public facilities, agricultural practices in adjacent fields, passing traffic and management practices on the vegetation. In the past, much research has been carried out on the influence of management of verges on biodiversity but, so far, no estimate has been made of the total surface of verges in Flanders as well as of the total SOC stock. Approximately 0Æ25 of the land area of Flanders is classified as urban area of which a small proportion is ‘green’ area (De Bruyn and Peymen, 2003). This ‘green’ area is dominated by grasslands and shrub vegetation. Although city parks generally have a lower nature conservation value than forests and grasslands, they are still a major form of habitat for wildlife within a city. Urban areas differ from natural environments in several aspects such as temperature, rainfall (precipitation), humidity and air quality (De Bruyn, 2001). In addition to parks and gardens, sport fields, camping sites and playgrounds also represent a considerable area of grassland and it is important to determine their SOC stocks. The aims of this study were to: (i) obtain a realistic estimate of the area of grassy verges along roads, waterways and railways, and (ii) calculate total SOC stocks for the different types of grass-covered verges and for the grassy vegetation in urban areas, namely parks, gardens and recreation fields.
Materials and methods Climate and soil type Flanders is situated in the north of Belgium and has a temperate wet climate. The different agro-pedological regions in Flanders, given in parentheses, represent
Ó 2005 Blackwell Publishing Ltd. Grass and Forage Science, 60, 151–156 151
152 I. Mestdagh et al.
different soil types: polders (clay), dunes (sand), campines (sand), a sandy loam region (sandy loam), a silt region (silt), a sandy region (sandy) and the pasture area of Liege (silt). All these agro-pedological regions are found in lower and central Belgium, where soils are allochthonous without gravel or rocks.
Verges along roads, waterways and railways Collection of the SOC data Soil samples were taken in different types of verges in the different agro-pedological regions. Five random verges were sampled per region with twenty-five samples taken on each verge from different depths (0– 10, 10–30 and 30–60 cm). The soil samples were analysed for C by the method of Walkley and Black (1934) and adjusted by 1Æ33 to compensate for incomplete oxidation (Batjes, 1996). For the dunes, no samples were taken. The SOC data from the campines were also used for the dunes. Both regions have a very similar sandy soil texture.
Bulk density Bulk densities for the different depths were used according to Mestdagh et al. (2004). For the depth of 60–100 cm, the bulk density for the depth of 30–60 cm was used.
Extrapolation of the SOC data to a depth of 1 m Soil samples could only be taken to a depth of 60 cm due to the presence of public facilities (gas, water, electricity). However, to compare with SOC stocks of other terrestrial ecosystems, calculation to a depth of 1 m is necessary. The SOC content to 60 cm was calculated with the classical layer-based method whereby the SOC contents for the depths 0–60 cm is calculated using the following equation: Total SOC (060 cm)ðt C ha1 Þ ¼ ½ðð%SOC=100Þ BD010 10Þ þ ðð%SOC=100Þ BD1030 20Þ þ ðð%SOC=100Þ BD3060 30Þ 100
ð1Þ
where %SOC is the percentage soil organic carbon, BD is the bulk density (g dry soil cm)3) and 10, 20 and 30 are the depths (cm) which were sampled at, respectively, 0–10, 10–30 and 30–60 cm. However, because no data were available for the layer 60–100 cm, it is still necessary to extrapolate data for the layer 60–100 cm. For the determination of the parameter k, required for the extrapolation from 60 cm to 1 m depth, the equation:
CðzÞ ¼ C b þ ðC 0 C b Þekz
ð2Þ
from Hilinski (2001) was used, where C(z) represents the SOC content (in g C g)1 dry soil) at depth z (cm), Cb the SOC content at the bottom of the profile, C0 at the upper layer of the profile, z the depth (cm) and k (cm)1) is the parameter representing the exponential decrease of SOC content with depth. The parameter k was calculated by non-linear regression from a series of soil profiles of grass-covered verges. For the extrapolation of the C data of the layer (30–60 cm) to the total SOC content to 1 m, an integration equation (Sleutel et al., 2003) was used: Total SOC ð60--100 cmÞðg C cm2 Þ ¼ qð1 e40k Þk1 ðC0 Cb Þ þ 40qC b
ð3Þ
where q is the bulk density (g dry soil cm)3) of the profile, k (cm)1) is the parameter which gives the decrease of SOC with depth, and C0 and Cb (g C g)1 dry soil) are, respectively, the SOC content in the upper and bottom layer of the profile. For the SOC content of the ‘upper’ layer of the profile, the mean SOC data from the 30–60-cm layer of the verges sampled in the different agro-pedological regions were used. For the SOC content at the bottom of the profile, the SOC content for the interval 80–100 cm from the ‘Aardewerk’ database (Van Orshoven et al., 1988) was used, assuming that the SOC content at that depth had remained the same since the 1950s (Degryze et al., 2004). The total SOC content to 100 cm was then obtained by summing the contents for the different layers. To calculate the total SOC stock to a depth of 1 m in a specific region, the SOC content to 1 m (in t C ha)1) was multiplied by the area (ha) considered.
Calculation of the total area The length of roads, railways and waterways was determined by overlaying the Flemish Road Atlas (ground for Geographical Information System, Catholic University Leuven) the Flemish Hydrographical Atlas with the administrative map (communities) for Flanders (source: Flemish Land Agency). For each community, it was possible to determine the length for each category of roads and waterways and the length of railways. The lengths of the different categories in Flanders are shown in Table 1. The extent of the verges covered with grass along roadsides was determined in the field by measuring the width of the different categories of roads (Table 1). For almost 0Æ07 of the total road length, the width was measured. The width of verges on each side along railways was, according to the National Railway Company of Belgium (NMBS), on average, 5Æ5 m, but a reduction factor of 0Æ50 was used to compensate for parts overgrown with trees and
Ó 2005 Blackwell Publishing Ltd, Grass and Forage Science, 60, 151–156
Soil organic stocks in verges and urban areas 153
Table 1 Length (km) and width (m) with standard deviation of mean (s.d.) for the different categories of roads, waterways (Source: Institute for Nature Conservation, Brussels) and railways (Source: National Railway Company of Belgium) in Flanders.
Category Motorways Roads like motorways High roads Secondary roads Connecting roads Important local roads Local roads Access roads Other roads Navigable waterways Not navigable waterways (category 1) Not navigable waterways (category 2) Not navigable waterways (category 3) Not classified waterways Railways
Length (km) for Flanders
Width at one side (m)
s.d. width
860 570 2100 2550 7770 1800 23 570 19 850 410 1580 1200 6280 6360 3790 2040
2Æ5 0Æ5 0Æ5 0Æ4 0Æ6 0Æ05 0Æ75 0Æ5 0Æ5 7Æ5 4Æ5 3 1Æ5 0Æ75 2Æ75
2Æ3 – 0Æ58 0Æ38 0Æ26 0Æ05 0Æ33 – – – – – – – –
shrubs (based on visual observations on different rail sections). The widths of the different categories of waterways (Source: Institute for Nature Conservation, Brussels) can also be seen in Table 1.
texture. For the pasture area of Liege (only one community in Flanders), the mean SOC content of the silt region was used. Both regions have a similar silt soil texture.
Management
Bulk density
In Flanders, verges have a specific management prescribed by law (De Wilde and Hermy, 2000). No biocides are allowed, cut grass has to be removed within 10 d of mowing, and the first cut is not allowed before 15 June and the second not before 15 September. No damage of below-ground plant parts or woody plants is allowed. Most of the verges are mown once or twice a year. Some of the verges along rivers and canals are grazed.
Bulk density was measured using the same procedure as for the verges.
‘Gardens and parks’ and ‘recreation areas’ SOC data To calculate the SOC stocks in gardens and parks as well as in recreation areas, the mean SOC contents of temporary agricultural grasslands (grassland younger than 5 years; Scottish Executive, 2004) of each agropedological region were used. Soil samples were taken at three depths (0–10, 10–30 and 30–60 cm) in temporary grassland in the different agro-pedological regions. The soil samples were analysed for C by the method of Walkley and Black (1934) and adjusted by 1Æ33 to compensate for incomplete oxidation (Batjes, 1996). No soil samples were taken from the dunes and in the pasture area of Liege. For the dunes, the mean SOC contents of the campines were used because both regions have a very similar sandy soil
Extrapolation of the data Agricultural grasslands were sampled to a depth of 60 cm. The SOC content for the layer 0–60 cm was calculated with the classical layer-based method according to Equation 1. Therefore, the layer 60–100 cm also needs to be extrapolated according to Equation 2. This was done using the mean SOC data from the 30–60-cm layer of the temporary grasslands in the different agropedological regions and the SOC content for the interval 80–100 cm from the ‘Aardewerk’ database (Van Orshoven et al., 1988).
Calculation of the total area Data for the categories ‘gardens and parks’ and ‘recreation areas’ (camping sites, sport fields, playing grounds, swimming pools and racecourses) are reported yearly by the National Institute for Statistics (NIS, 2004). A reduction factor of 0Æ50 was used because not all the surface is covered with grass. This factor is an estimate because it is impossible to retrieve how much surface of those two categories is covered with grass and how much of it is covered with trees, shrubs, buildings or concrete.
Ó 2005 Blackwell Publishing Ltd, Grass and Forage Science, 60, 151–156
154 I. Mestdagh et al.
Results and discussion Area of grass verges The total area of verges along roads, waterways and railways in Flanders was estimated to be 18 027 ha. The total area of grass-covered verges along roads was 7413 ha, 9542 ha along waterways and 1072 ha along railways. The largest grass area is on the verges along waterways, whereas railways only have a small grasscovered area. The value calculated here for roads was lower than the estimate made by De Wilde and Hermy (2000). They calculated an area of between 15 200 and 19 200 ha for grass-covered verges along roads only, based on their results and on those of Van Dale (1993). Their value was more than twice the area calculated along roads in this study. The reason for this difference comes from the width of the verges used, which is very important for the calculation of the total area. On average, De Wilde and Hermy (2000) used a verge width of between 1Æ4 and 1Æ8 m (one side), whereas here a mean width of only 0Æ75 m (one side) was used. Each year, the mean width of verges declines because of maintenance works, widening of roads, irreparable damage by the passing traffic and because most of the verges along local roads are ploughed along with the adjacent croplands. An average width of 1Æ5 m is quite large because only motorways still have a considerably larger width. However, all other roads, which form the main part of the road matrix in Flanders, have an average width of about 0Æ5 m. Along waterways, De Wilde and Hermy (2000) estimated the surface of verges along navigable waterways at 1625 ha and at 1080 ha for waterways of category 1 (see Table 1); whereas we calculated, respectively, 2370 and 10 800 ha. Again the difference can be explained by the difference in width used. De Wilde and Hermy (2000) used a width for the navigable waterways of 5Æ13 m (one side) whereas, according to the Institute for Nature Conservation, it is, on average, 7Æ5 m (one side). In this study, the error on the length for roads, waterways and railways in Flanders is very small due to the high accuracy of the GIS
data. The standard deviations of the widths along roads measured here are large. This indicates that the variation within the widths is large, for example, sometimes there is no grassy vegetation for some kilometres, whereas for the following kilometres there is a width of 6 m of grassy vegetation.
Area of ‘gardens and parks’ and ‘recreation areas’ The total surface covered with grass for the category ‘gardens and parks’ was estimated at 9530 ha and for the category ‘recreation areas’ was 2360 ha.
Soil organic carbon stocks To make an extrapolation for the layer 60–100 cm, values for the parameter k were estimated for grasscovered verges in Flanders (Table 2). The mean SOC contents for grassy verges to a depth of 1 m are shown in Table 3. Using these SOC contents, it was possible to calculate total SOC stocks (to a depth of 1 m) for the different types of grassy verges. The total SOC stock to 1 m was estimated to be for grass-covered verges along roads 1438 kt SOC, for waterways 1873 kt SOC and for railways 209 kt SOC. The total SOC stock (to a depth of 1 m) for the three types of grass-covered verges was estimated to be 3520 kt SOC. The calculated total grassy surface of Flemish verges in this study comes to almost 0Æ08 of the total grassland area in Flanders and the total SOC stock to 0Æ10 of the total SOC stock in grasslands. The potential for additional C sequestration and mitigating atmospheric CO2 for verges along roads is likely to be small due to (i) the increasing pressure of traffic on the verges; (ii) road maintenance and maintenance for public facilities (gas, water and electricity) and (iii) the limited possibilities in changing their management. Along railways, the possibilities for a changing management are also very limited because half of the total surface consists of woody vegetation and the grassy parts are mown as often as necessary to assure a clear view for train drivers. However, the above-mentioned problems do not often occur in verges
Temporary grassland
Grass-covered verges
Agro-pedological region
k-value
s.e.
k-value
s.e.
Campines Silt region Polders Sandy region Sandy loam region Pasture area of Liege
0Æ01597 0Æ01642 0Æ01486 0Æ01604 0Æ01612 0Æ01642
0Æ00123 0Æ00361 0Æ00198 0Æ00188 0Æ00098 0Æ00361
0Æ01597 0Æ01657 0Æ01579 0Æ01593 0Æ01660 0Æ01626
0Æ00438 0Æ00532 0Æ00429 0Æ00435 0Æ00626 0Æ00454
Table 2 Values for the parameter k (cm)1) with standard error of mean (s.e.) for the layer of 60–100 cm for grasscovered verges and for temporary agricultural grassland for the different agro-pedological regions in Flanders.
Ó 2005 Blackwell Publishing Ltd, Grass and Forage Science, 60, 151–156
Soil organic stocks in verges and urban areas 155
Table 3 Mean and standard deviation of mean (s.d.) of soil organic carbon (SOC) contents (t C ha)1) for temporary grasslands and grass-covered verges in Flanders to a depth of 1 m for the different agropedological regions.
Temporary grassland Agro-pedological region Campines Silt Region Polders Sandy region Sandy loam region Pasture area of Liege
along waterways and therefore, any potential for C sequestration should be focused on these verges. A possible measure to enhance C sequestration in these verges is to increase the frequency of grazing. Only a very small proportion of the verges along waterways are grazed, the rest are mown. In general, grazing enhances the SOC stock (Hassink and Neeteson, 1991; Schuman et al., 1999). Planting trees in all three types of verges could also lead to an additional C sequestration mainly in the above-ground biomass but less in the soil. Table 2 shows the values of the parameter k for temporary grassland in ‘gardens and parks’ and ‘recreation areas’ for the extrapolation for the layer 60–100 cm. The mean SOC contents to 1 m, used to calculate the total SOC stocks, are shown in Table 3. A total SOC stock of 1392 kt SOC was calculated for the grassy parts of the category ‘gardens and parks’ and 346 kt SOC for the category ‘recreation areas’. For the grassy urban area, the total surface equates to 0Æ05 of the total grassland area and the total SOC stock equates to almost 0Æ05 of the total SOC stock for Flemish grasslands. Management measures to enhance C sequestration in the urban area are quite difficult because (i) gardens are private property and cannot easily be subjected to laws and (ii) parks and recreation fields keep their grasses very short. However, the IPCC (2001) mentions a management strategy for urban and peri-urban land. Tree planting, improved waste management and wood production could lead to a net annual rate of change in C stocks of 0Æ3 t C ha)1 year)1.
Conclusion The total SOC stock (to a depth of 1 m) for the three types of grass-covered verges was estimated at 3520 kt SOC. For the grassy parts of the category ‘gardens and parks’ a total SOC stock of 1392 kt SOC was calculated and 346 kt SOC for the category ‘recreation areas’. Although the potential for additional C sequestration by both verges and urban areas is relatively small, they still have a total SOC stock which is almost 0Æ15 of the total SOC stock in Flemish grasslands.
Grass-covered verges
Mean
s.d.
Mean
s.d.
176 106 135 162 129 –
70 33 15 66 50 –
190 163 183 215 186 262
57 52 50 29 58 29
Acknowledgments This study is part of the CASTEC research project (Carbon Sequestration potential in Terrestrial Ecosystems), funded by the Belgian Federal Science Policy Office, Brussels. The authors thank Marcel Voet (Institute for Nature Conservation) for helping with the determination of the length of roads, waterways and railways.
References Batjes N.H. (1996) Total carbon and nitrogen in soils of the world. European Journal of Soil Science, 47, 151–163. De Bruyn L. (2001) The urban area (in Dutch). In: Kuijken E., Boeye D., De Bruyn L., De Roo K., Dumortier M., Peymen J., Schneiders A., van Straaten D. and Weyemberh G. (eds) Nature Report 2001. Condition of Nature in Flanders: Figures for Policy. Communications of the Institute for Nature Conservation No. 18, Brussels, pp. 106– 110. Brussels, Belgium: Institute for Nature Conservation. De Bruyn L. and Peymen J. (2003) The urban area (in Dutch). In: Dumortier M., De Bruyn L., Peymen J., Schneiders A., Van Daele T., Weyemberh G., van Straaten D. and Kuijken E. (eds) Nature Report 2003. Condition of Nature in Flanders: Figures for Policy. Communications of the Institute for Nature Conservation No. 21, Brussels, pp. 130–134. Brussels, Belgium: Institute for Nature Conservation. De Wilde M. and Hermy M. (2000) Nature Technical Assimilation of Hay from Verges (in Dutch). Brussels, Belgium: Ministry of the Flemish Government, Department of General Environmental and Nature Policy. Degryze S., Six J., Paustian K., Morris S.J., Paul E.A. and Merckx R. (2004) Soil organic carbon pool changes following land-use conversions. Global Change Biology, 10, 1120–1132. Hassink J. and Neeteson J.J. (1991) Effect of grassland management on the amounts of soil organic N and C. Netherlands Journal of Agricultural Science, 39, 225–236. Hilinski T.E. (2001) Implementation of exponential depth distribution of organic carbon in the Century Model. http://www.nrel.colostate.edu/projects/century5/
Ó 2005 Blackwell Publishing Ltd, Grass and Forage Science, 60, 151–156
156 I. Mestdagh et al.
reference/html/century/exp-c-distrib.htm (date of last access 29 March 2005). IPCC (2001) Land Use, Land-use Change and Forestry. Cambridge, UK: Cambridge University Press. Mestdagh I., Lootens P. and Carlier L. (2004) Soil organic carbon stocks in Flemish grasslands: how accurate are they? Grass and Forage Science, 59, 310–317. NIS (2004) Soil Use in 2003 (In Dutch). Brussels: National Institute of Statistics. Schuman G.E., Reeder J.D., Manley J.T., Hart R.H. and Manley W.A. (1999) Impact of grazing management on the carbon and nitrogen balance of a mixed-grass rangeland. Ecological Applications, 9, 65–71. Scottish Executive (2004) Cap Reform: Cross Compliance (Good Agricultural and Environmental Condition). Edinburgh, UK: Scottish Executive. Sleutel S., Deneve S. and Hofman G. (2003) Estimates of carbon stock changes in Belgian cropland. Soil Use and Management, 19, 166–171.
UNFCCC (1997) Kyoto Protocol to the United Nations Framework Convention on Climate Change. http://Unfccc.Int/ Resource/Docs/Convkp/Kpeng.Html (date of last access 27 February 2005). Van Dale G. (1993) Feasibility Study on the Assimilation Possibilities of Hay from Verges on Community Roads (in Dutch). Committee Jean Pain/Ovam, Brussels: Public Waste Agency of Flanders. Van Orshoven J., Maes J., Vereecken H., Feyen J. and Dudal R. (1988) A structural database of Belgian soil profile data. Pedologie, 38, 191–206. Walkley A. and Black I.A. (1934) An examination of the Djegtareff method for determining soil organic matter and a proposed modification of the chromic acid titration method. Soil Science, 37, 29–38. Zwaenepoel A. (1993) Management and typology of the vegetation on verges in Flanders (in Dutch). Ph.D. thesis, University Ghent, Ghent.
Ó 2005 Blackwell Publishing Ltd, Grass and Forage Science, 60, 151–156
