For testing the hydraulic response to the test- and reference rains in MIKE 21 (see sec-tion 6.4), a generic model of an urban catchment was constructed in ArcMap. The model was designed as a square catchment with size 8 km x 8 km (in total 64 km2), with the outlet in one corner, and a main drainage path diagonally through the square, see Figure 8. Nested square subcatchments were constructed in the model, with sizes of 2 km x 2 km, 4 km x 4 km and 6 km x 6 km, giving four catchments of sizes of 4, 16, 36 and 64 km2 respectively, including the complete model, all with the same square shape. The nested catchments in the urban model is referred to as follows: 2 km x 2 km: A, 4 km x 4 km: B, 6 km x 6 km: C, and the entire model of 8 km x 8 km: D. According to Tusher (2019a), 95 % of all urban catchments in Sweden which are not cut through by major watercourses, are smaller than 5 km2. The largest urban catchment according to the same definition was 33 km2, situated in Stockholm (Tusher, 2019b). The catchment model hence represents sizes of Swedish urban catchments with a well taken margin. Us-ing catchment sizes less than 2 km x 2 km was regarded inappropriate, since that was the resolution of the precipitation radar used for constructing the test rains.
Part 2: Modelling of hydraulic response
Figure 7: Map showing the general structure and the topography of the urban catchment model. The four nested square catchments of sizes 2 km x 2 km, 4 km x 4 km, 6 km x 6 km and 8 km x 8 km are shown delimited with dark gray lines, where the entire model corresponds to the largest catchment.
The principal drainage paths are shown in blue lines, with the main drainage path diagonally through the model. The background colour is showing the topography, with blue colours representing the lowest terrain and red colours representing the highest terrain. The total height difference within the model is set to 45 m.
The topography of the catchment model was set in order to obtain waterflow in accor-dance with the three nested subcatchments, see Figure 8. The complete topography was created in ArcMap as a spline interpolation between a number of points with defined height. The total relative height difference within the model was set to 45 m, with the lowest point at the outlet, and the highest points where the subcatchment water dividers intersect with the model border. This yielded a slope of 1.4 ‰ along the main drainage path, and 2 ‰ along the auxiliary drainage paths in each subcatchment.
A total relative height difference of 45 m over an area of 64 km2 is relatively low in comparison to most Swedish cities. It is comparable to the conditions in cities located on clay plains, like Malmö, and to some degree Uppsala. Stockholm and Göteborg are located in joint valley landscapes (Swedish: sprickdalslandskap) with somewhat higher relative height differences. Especially the small scale topography in such a landscape is much more pronounced than in the model. The steep slopes that frequently occur in such landscapes are however concentrated to higher terrain with rock and moraine, while the low terrain, where the main water flow paths are located, are constituted by flat valleys with clay, with slopes comparable to those in the model. The model topography was set
Part 2: Modelling of hydraulic response
in accordance, since the main water flow paths were considered more important than the higher terrain, where less water flows. Hence, the model topography can be considered representative for most larger cities in Sweden, except those located in hilly terrain with higher relative height differences, like Jönköping and Sundsvall.
The urban catchment model was constructed separating between four kinds of features:
Impervious surfaces (representing hardened surfaces like pavement and asphalt), pervious surfaces (representing non hardened surfaces like lawns and nature), buildings and low points (low terrain without topographical outlets, in this case pervious). The resolution was set to 10 m, and roads were built one grid wide. This is a rather coarse resolution, but since the aim of the study is not where water flows in a high detail manor but rather the hydraulic responses over larger areas, the correct direction of water flow on a larger scale is sufficient.
Figure 8: Illustration showing how the entire urban catchment model, to the right, is constructed by arrangements of uniform blocks of 400 m x 400 m, shown to the left. An unobstructed road, constituting the main drainage path, passes diagonally from the top right to the down left corner.
The model was constructed by uniform square basic blocks of size 400 m x 400 m merged together, making the model quasi-uniform on larger scales, as shown in Figure 9. The basic block was constructed in order to represent a generic Swedish urban catchment as a whole. Different parts of the block were constructed to represent city centers, apartment areas, commerce-/ and industrial areas and detached house areas respectively, with green belts in between, and roads along two sides. Two low points were also inserted into the basic block. The basic block was constructed almost symmetrically mirrored along the diagonal. The side roads were only built on one of the mirrored sides though, completing the square road grid along the block borders first when the blocks were merged together.
The basic blocks were also rotated when put together in order to make the city center areas located in two of the corners of the basic blocks to converge into clusters of 4, in order to imitate the general structure of extended urban areas, with accumulation of
Part 2: Modelling of hydraulic response
closed building structures in suburban nodes. A 20 m wide hardened path was added diagonally through the entire catchment model along the main drainage path, in order to allow a generous water flow. The overall proportion of hardened surfaces (including buildings) was yielded as 38 %, which can be considered representative for a Swedish city (Svenskt vatten, 2011).