URBAN SMS Soil Management Strategy

URBAN SMS Soil Management Strategy Baseline scenario analysis Modeling future urban sprawl in pilot cities A. Łopatka, G. Siebielec, A. Żurek, M.Głuszynska, H. van Delden & T. Stuczynski December, 2010

Deliverable report Deliverable number: 6.2.1 BASELINE SCENARIO ANALYSIS - MODELING FUTURE URBAN SPRAWL IN PILOT CITIES Responsible for the deliverable: PP8 Institute of Soil Science and Plant Cultivation State Research Institute Research Institute for Knowledge Systems BV Authors: Artur Łopatka Grzegorz Siebielec Anna Żurek Magdalena Głuszynska Hedwig van Delden Tomasz Stuczynski Partners Involved: LP: City of Stuttgart PP2: City of Vienna PP3: Umweltbundesamt, Environment Agency Austria, Vienna/ Austria PP4: City of Milan, Executive Plans and Programs for Buildings Department- Office for Reclamation Plans, Italy PP9: Czech University of Life Sciences, Prague PP10: Soil Science and Conservation Research Institute, Bratislava, Slovakia PP11: District Authority Stuttgart, Germany December, 2010

CONTENT: 1. INTRODUCTION... 3 2. MATERIALS AND METHODS... 4 2.1. Data sources... 4 2.2. Approach for urban sprawl simulation cellular automata (CA)... 5 2.3. Land Suitability... 7 3. RESULTS FOR SIMULATION OF BASELINE SCENARIO... 9 4. SUMMARY... 17 5. REFERENCES... 17 2

1. INTRODUCTION During the last decade more emphasis was given to functions of landscapes and their sustainability as a response to a need to minimize the depletion of land resources and to reduce environmental and social impacts caused by land use changes. Changes are not limited to land cover only, but potentially lead to disturbance of landscape functions. Land use change-related pressures on ecosystems are inherent consequences of economic and social growth, and as such cannot be entirely avoided. However, it is an unavoidable necessity to balance between an increasing demand for goods and services and environmental quality. The soil framework strategy presented by the European Commission (COM 231, 2006) identifies number of threats to maintaining soil functions in Europe: erosion, decline of organic matter, local and diffuse contamination, sealing, compaction, decline in biodiversity, salinisation, floods and landslides. Sealing is one of main threats itself, additionally urbanization of agricultural land may accelerate the other degradation processes (Stuczynski, 2007). Modeling is becoming an important tool in context of conflicts between urbanization and landscape or soil protection, since urbanization driven degradation processes are often irreversible. Even if the prediction power of models is sometime limited, they can provide valuable insights into the development of trends caused by different policy scenarios or soil protection regulations. The idea of using models lies in their ability to detect possible conflicts which may arise as a result of existing or implementing given new policies affecting land use (Hilferink & Rietveld, 1999; Westhoek et al., 2006). Cellular Automata are a family of models that can be suitable for the simulation of urban growth and land use changes. We used Metronamica software, developed by Research Institute for Knowledge Systems (RIKS), which is an unique tool for planners and researchers to simulate future land use change, especially urbanization, in a spatial manner. The tool takes socioeconomic and biophysical aspects into account in modeling process. It allows also to test various scenarios, relative to different EU or national policies, when the quantitative description of the scenario is incorporated into a model. Our goal was to forecast urbanization sprawl in the URBAN SMS pilot cities up to year 2030 for the baseline scenario that assumes that there are no limitations for soil consumption as related to soil quality. 3

Furthermore, it was assessed how the predicted changes would affect the soil resources in the cities. The analysis was performed for Stuttgart, Milan, Prague, Bratislava, Vienna, Salzburg and Wroclaw. 2. MATERIALS AND METHODS 2.1. Data sources Two main spatial databases were applied for urban spread simulation within the pilot urban areas: land use classification and soil maps. Land use information layers were prepared through classification of satellite images as it was in detail described in the Urban SMS 6.1.2 deliverable report: Assessment of soil protection efficiency and land use change. Land use information was gathered for two periods: the initial status map for 1990-92 and the current status map for 2006-07. The land use change maps representing two periods enabled assessment of time-dependent land transformation rates under the as is soil management system. Land use maps contained 13 different land use classes. For the modeling purposes the original 13-class classification was reclassified to 4 classes representing 4 groups of urban land utilization: 0) agricultural and semi-natural 1) residential continuous and discontinuous, commercial and industrial, dump and mineral extraction sites, airports, transport facilities, sport and leisure facilities 2) forests, green recreation areas 3) water bodies. In the simulation process it was assumed that urbanization may take place only on agricultural and semi-natural areas (class 0), thus this group of land uses served as land pool available for the potential sealing. Generally use of forest areas is restricted in all cities, thus this class was excluded from allocation of new urban fabrics. Class 2 (forests and green recreation areas) and class 3 (water bodies) remained unchanged in the modeling their areas did not increase nor were not reduced. Soil maps were provided by the project partners responsible for the respective test cities. Due to diverse soil quality assessment systems, present in different Central European countries, and content of the soil maps the polygons on each map were grouped into 3 classes. They represented high, medium and low quality soils (either from perspective of production function, ecosystem function, buffering, retention etc.). In some cities the available soil maps did not cover the whole city area. In such cases it was assumed that unsealed areas without 4

soil information are represented by medium soil quality class. This enabled full coverage of city areas with simulated future urban sprawls. The land use maps and the soil maps were converted to raster data with grid resolution of 50 meters. It has been observed that new residential areas were rather clustered and did not exhibit elements smaller that 50 meters. 2.2. Approach for urban sprawl simulation cellular automata (CA) In the analysis we used the Cellular Automata-based Metronamica model. The software was developed and provided by the Research Institute from Knowledge Systems (RIKS) from Maastricht, The Netherlands. The software utilizes cellular automata model to spatially distribute areas of particular land use classes. Cellular automata is a discrete model which uses regular grid of cells, each classified within finite number of states. In case of Metronamica land use change modeling, these states refer to land use classes. Land use types are classified in the model according to their behavior into the following categories: feature states (fixed land uses that do not change dynamically), function states (they change dynamically as the result of the local and the regional dynamics) and vacant states (they change dynamically due to the local dynamics only). In our simulations: Class 0) agricultural and semi-natural uses are vacant states Class 1) residential continuous and discontinuous fabrics, commercial and industrial, dump and mineral extraction sites, airports, transport facilities, sport and leisure facilities are function states Class 2) forests, green recreation areas are feature states Class 3) water bodies are feature states. The neighborhood of a cell (surrounding cells) influences the transition of this cell into other class in the next time step. The cells located further away have a smaller effect than cells closer to the centre cell. The transition rules are the core of the CA and determine if, and how, the state of each cell in the next time step changes. The neighborhood effect in this analysis is defined as: the attraction or repulsion effect of surrounding cells which eventually causes a change in cell status (type of land use) of the centre cell. For each land use function, a set of rules determines the degree to which it is attracted to, or repelled by, the other functions present in the neighborhood. In our simulation, based on the neighborhood land use interaction and land suitability, described in the subsequent sub-section, the model calculates the value of transition potential 5

(Pk) for each cell and land use function and for every simulation time step. All cells are ranked according to their transition potential, and cell transitions begin with the highest ranked cell and the transition process proceeds downward. The number of cells required for urban function in the subsequent years is determined by simple linear relationship between historical data (land use map for 1990/92) and current data (land use map 2006/2007) (Figure 1). 140 120 residential area [km2] 100 80 60 40 20 0 1985 1990 1995 2000 2005 2010 2015 2020 2025 2030 2035 year Figure 1. Exemplary linear trend for prediction of residential area in 2030 Areas converted to urban land use class for all test cities from Urban SMS project are collected in Table 1. In the second column the yearly area converted is expressed as % of existing urbanized area. This enables comparison of urban growth rate between cities of different size. Table 1. Statistics for conversion of agricultural and semi-natural areas (class 0) into urban functions based on historivcal data and predicted consumption based on the observed trend City yearly urbanization of agricultural and seminatural areas area converted to 2030 [ha] [ha] [% urbanized area] Bratislava 29.4 0.279 675 Milan 21.5 0.188 515 Prague 50.7 0.221 1217 Salzburg 8.4 0.141 192 Stuttgart 9.9 0.081 237 Vienna 16.4 0.067 377 Wroclaw 34.6 0.301 830 6

2.3. Land Suitability The term suitability is used here to describe the degree to which a cell is able to support a particular land use function. It is a composite measure, prepared in a Geographical Information System (GIS), on the basis of factor maps determining the physical, ecological and environmental appropriateness of cells. For transition into residential areas the Land Suitability (LS) includes four factors: terrain suitability S (slope), road accessibility S (distance to road), urban potential S (density of urbanized cells) and soil suitability S (soil). It is also assumed that these factors work independently so that the final land suitability (LS) is a product of them: LS = S( slope) S( dist. to road ) S( urban potential ) S( soil) Values for first three factors were derived from share of new urbanized cells that appeared in the selected groups (percent slope or distance to road or density of residential area in a neighborhood of cell) between 1990/92 and 2006/2007 (Figure 2). 100 100 90 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0-50 50-100 100-150 150-200 200-250 250-300 300-500 500-1000 > 1000 0 0-5 5-10 10-15 15-20 > 20 2a) suitability for distance to road classes 2b) suitability for slope classes 100 100 90 80 90 80 Scenario 1 Scenario 2 Scenario 3 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0-0,2 0,2-0,4 0,4-0,6 0,6-0,8 0,8-1 0 low quality soil medium quality soil high quality soil 2c) suitability for urban potential classes 2d) suitability for soil quality classes Figure 2. Suitability values for four components of the Land Suitability (LS). 7

All three partial suitability layers were normalized to reach values within range from 0 to 100 (0 means that a given cell is not useful for residential area, 100 means that there are no limitations to allocate urban function in this cell). Soil suitability values for urban function were established according to soil quality maps and were dependent on the analyzed soil protection scenario. In the baseline scenario, covered by this report, it is assumed that all three soil quality classes have the same soil suitability value, equal 100. Exemplary maps of suitability components and the final Land Suitability for Bratislava are presented on Figure 3. 3a) Road accessibility. 3b) Terrain suitability. 3c) Urban potential suitability. 3d) Soil suitability. 8

3e) Land suitability for urban land use function. Figure 3. Maps of suitability land suitability components (3a-3d) and final Land Suitability (3e) for Bratislava 3. RESULTS FOR SIMULATION OF BASELINE SCENARIO Maps generated by Monte Carlo method (100 runs) display probabilities for transformation of a given cell of agricultural or semi-natural area into urban function. In order to forecast the impact of the baseline scenario on soil cover, the probabilities are combined with soil quality information. 9

Figure 4. Historical urban sprawl in Bratislava between1992 and 2007. Figure 5. Simulated probability of urbanization in Bratislava between 2007 and 2030. 10

Figure 6. Historical urban sprawl in Milan between1991 and 2006. Figure 7. Simulated probability of urbanization in Milan between 2006 and 2030. 11

Figure 8. Historical urban sprawl in Prague between1990 and 2006. Figure 9. Simulated probability of urbanization in Prague between 2006 and 2030. 12

Figure 10. Historical urban sprawl in Salzburg between1991 and 2007. Figure 11. Simulated probability of urbanization in Salzburg between 2007 and 2030. 13

Figure 12. Historical urban sprawl in Stuttgart between1992 and 2006. Figure 13. Simulated probability of urbanization in Stuttgart between 2006 and 2030. 14

Figure 14. Historical urban sprawl in Wroclaw between1991 and 2006. Figure 15. Simulated probability of urbanization in Wroclaw between 2006 and 2030. 15

Figure 16. Historical urban sprawl in Vienna between1991 and 2007. Figure 17. Simulated probability of urbanization in Vienna between 2007 and 2030. 16

4. SUMMARY The baseline scenario forecast of urbanization was performed for 7 urban pilot areas by using Cellular Automata Metronamica model. The forecast utilizes, so called, transition potential of a given piece of land (cell in spatial information layer) which combines effect of surrounding land uses and land suitability for the urbanization. Land suitability combines such factors as slope, road density, urban fabrics density and soil suitability. Baseline scenario assumed no limitations, related to soil quality, for transformation of agricultural or seminatural lands into urban functions. For some countries, with no clear urban soil protection system, this scenario would be approximate to the as is scenario. Under the analyzed scenario the transformation of land into urban purposes is driven only by socio-economic factors and some physical conditions such as slope or distance to transport network. For each pilot city the potential spatial distribution of new urban objects was produced up to year 2030. This information was superimposed on soil quality maps in order to learn what would be quality of soils lost under the baseline scenario. The baseline scenario will be in detail compared to alternative soil protection scenarios in the subsequent Deliverable 6.2.2 report ( Urban sprawl under alternative soil protection scenarios ). 5. REFERENCES COM231. 2006. Communication from the Commission to The Council and European Parliament, The European Economic and Social Committee of The Regions. Thematic Strategy for Soil Protection Hilferink M., P. Rietveld. 1999. LAND USE SCANNER: An integrated GIS based model for long of land use in urban and rural areas. Journal of Geographical Systems, 1:155-177 RIKS BV. 2005. Metronamica - A Dynamic Spatial Land Use Model, RIKS, Maastricht, The Netherlands Stuczynski T. 2007. Assessment an modeling of land use change in Europe in the context of soil protection. Pulawy, Poland Westhoek H.J, M. van den Berg, J.A. Bakkes. 2006. Scenario development to explore the future of Europe s rural areas. Agriculture, Ecosystems and Environment 114: 7-20 17

URBAN SMS Soil Management Strategy This paper belongs to the following section of URBAN SMS work plan: WP6 Acceptance and awareness / 6.2 Protection scenario modelling / 6.2.1 Baseline scenario analysis www.urban-sms.eu Contact details of project partner commissioning / responsible for this paper: Mr Grzegorz Siebielec, Institute of Soil Science and Plant Cultivation, ul. Czartoryskich 8, 24-100 Pulawy, PL, gs@iung.pulawy.pl Other URBAN SMS Partners contributing to this paper: Petra Blümlein, City of Stuttgart, Department for Environmental Protection, DE Isabel Wieshofer, City of Vienna, Environmental Protection, AT Sigbert Huber, Umweltbundesamt, Environment Agency Austria, Vienna, AT Marco Parolin, City of Milan, Executive Plans and Programs for Buildings Department- Office for Reclamation Plans, I Petra Vokurková, Czech University of Life Sciences Prague, CZ Jaroslava Sobocká, Soil Science and Conservation Research Institute, Bratislava, SK Siegmar Jaensch, District Authority Stuttgart, DE This project is implemented through the CENTRAL EUROPE Programme co-financed by the ERDF. The paper in hand reflects the author s views and the Managing Authority of the INTERREG IV B CENTRAL Programme is not liable for any use that may be made of the information contained therein.