The Climatemeter

A map of depressions is generated using a digital terrain model. This identifies depressions in the ground as potential floodplains. The calculation of areas at risk of flooding is made on the assumption that all surfaces are completely impermeable and there is no possibility of seepage or runoff via drains or sewers. The maps thus show risk areas where water may potentially collect and cause flooding. A GIS (Geographic Information System) view has the option to show only depressions of a certain volume, to avoid 'noise' in the map material as shown in the figure on the right. The simplest method does not contain any information on the volume of precipitation required to fill the depressions. Filling of the depressions can in most cases be described as coinciding with extreme precipitation at a very long return period. The measurement of the catchment area for each depression can be made using the terrain model, and waterways in the terrain can also be seen. The catchment size and degree of building development in it are used to calculate how many millimetres of rain are required to fill the depression, see figure on the right. This may however be misleading, as a large depression can be critical, e.g. even when half full. It is important to verify the quality of the terrain model data before it is used as a basis for flood analysis. As bridges often have the effect of barriers in raw elevation data, these areas should be examined and where necessary, underpasses incorporated. Often GIS registration information on major bridges in a municipality is available. This can either be directly incorporated in the terrain model or used as a template for a manual review of the areas. Similarly, watercourses should be covered, as stretches with piping will also act as a barrier. All areas adjacent to houses should also be quality-assured.

The simplest method - using the Map of depressions, can only be used to calculate the depth of depressions, catchment size and surface water flow paths. Photographic analysis and/or GIS analysis of base data, for example, can be supplemented with degree of building cover, which will then provide actual hydraulic surface calculations of flooding for current rainfall events. The calculations can be made as 1D or 2D surface calculations, that is to say by calculating catchment area runoff as one-dimensional channels or two-dimensional surfaces. The latter method is significantly more complicated in terms of calculations. These surface calculations provide a good estimate of the relationship between catchment area and depression volume, but do not take account of the contribution from or capacity of the sewerage network. This generally results in an overestimation of flooding upstream from the pipe system, and underestimation of flooding downstream from the drainage network. The figure on the right shows the difference in flood extents between a model with an impermeable surface and a model where the surface hydraulic properties are included.

A hydrodynamic drainage model (for example MIKE URBAN) calculates locations in the drainage system at which bottlenecks may arise after large volumes of rain and can thus provide a preliminary overview of initial water accumulation on land. This can be combined with information on the way water runs on the surface, and where water can collect in depressions. If water rises at several locations close to a risk area shown on a depression map analysis, further calculations on the area should be made. A hydrodynamic model does not describe the flow on the surface, and the interaction between the surface and the drainage system is not included. It can therefore only determine the intensity of rainfall events that the drainage system's capacity can carry. Where there is water on land, the results of the model become unreliable.

Flooding in urban areas often occurs where water collects in depressions on the surface because the drainage system capacity is insufficient. Water drains away only after space in the pipes is freed up. Considerable accuracy in mapping the risk areas can be achieved by combining a surface model with a drainage model. A digital analysis of flow patterns on the surface allows the surface to be reduced to a series of depressions and channels, suited to calculation methods used in dynamic one-dimensional flow models. In this way, the surface model can be directly combined with the drainage model so that a 1D-1D system is achieved, see figure on the right. The complexity of the generated 1D surface model depends to a high degree on the minimum size of depressions included in the model. The number of depressions can be reduced by combining depressions, for example using a volume limit value for each depression. The level of detail is greatest where limit values are low, but calculation time is increased correspondingly.

A hydrodynamic drainage model which describes water flow in the drainage system, for example MIKE URBAN, can be linked with a 2D surface model, for example MIKE Flood, to obtain a more precise picture of the dynamic of the water on the surface for an extreme rainfall event. A 2D surface model can more precisely describe the runoff on the surface than a 1D model but at the same time requires significantly more calculations. Elevation models are often used as input data for the surface model. These are derived from airborne laser altimetry or more detailed elevation models with a grid size of between 1 and 4 m. The surface model features polygons for houses, roads, etc., so that they are physically correctly represented in the model. Quality assurance of the model is important as selected drainage wells from MIKE URBAN are incorporated in the terrain model. As the placing of the drainage wells on the terrain model is of significance for the accuracy of the model, their siting should be checked so that the exchange of water between terrain and drainage systems is realistically represented. Furthermore the model should be calibrated with respect to the degree of building cover on the surface model, possible water paths on the surface that may be distorted by the subsequent addition of houses, etc, and the placing of and water exchange in included drainage wells. Calculation time is long and a large amount of high-quality data is required.

An overall combined model for drainage, surface, ground water and sea will be a tool for precise calculation of the dynamic with respect to a raised water table and sea level as well as of the consequences of extreme rainfall events. The model may consist of a combination of various numerical models (for example MIKE URBAN and MIKE SHE). Changes in the water table, for example, can result in increased flow of groundwater to water courses and sewers, which can result in more rapid flooding of the terrain; in this situation minor rainfall events can cause flooding more easily than before. Selected drainage wells are linked to the surface model. As the placing of the drainage wells on the terrain model is of significance for the accuracy of the model, their siting should be checked so that the exchange of water between terrain and drainage systems is realistically represented. MIKE SHE models the exchange of water between individual components of the hydrological cycle, including groundwater. The model should be calibrated in respect of the degree of building cover on the surface model, possible water paths on the surface that may be distorted by the subsequent addition of houses, etc, and the placing of and water exchange in included drainage wells. The model requires considerable input of high quality and a high spatial resolution. This makes quality assurance time-consuming, and calculation time is also significant.

An analysis of a digital terrain model will easily predict locations in the municipality at which a given sea water level, if applicable, can cause water to flow over the land and fill relevant water level contours. The flood map calculates water depth on terrain as the difference between water level at sea and the local terrain contour. Under this model, all areas of terrain lower than sea level will be flooded even if they are not situated close to the sea. The contour map in practice uses a colour key for the specific contours for visualisation (for example 2 m water level). Results from this model are naturally very uncertain as the method assumes flooding of areas that are not in direct contact with the sea. This method is most useful for coastal areas with significant slopes in the terrain, meaning that only areas close to the coast are flooded. The method is limited in that physical processes are ignored - for example the time it takes for water to flow over terrain. The method provides the picture of maximum flooding and is particularly good for quick screening of areas that are in a danger zone with regard to flooding. The raw terrain model often omits water course underpasses and structures under bridges. The water flow may thus appear to be obstructed by the topography, where in fact there is an underpass. This may mean that flood areas that in reality are connected are shown as separate. Quality assurance of the terrain model, where water course underpasses under roads and similar are included, gives the correct representation of potential flood areas.

A "cost-distance" analysis allows identification of below sea-level areas for specific sea levels, which are thus flood-prone. The result is a connection map, see figure on the right. The method is a significant advance on the simple contour map in that the significance of dykes and other barriers is factored in. However there are still several aspects that mean that the cost-distance method does not accurately portray the extent of flooding: 1. Water courses that are not wide enough to be represented in the terrain model (if the water course is 1 m wide and the DTM resolution is 1.6 x1.6m) may result in less serious flooding in the model than in reality. 2. Water course underpasses that are not open (no correction made) in the terrain model (for example under roads and small bridges) may result in less serious flooding in the model than in reality. 3. Flood defences that are not represented because the terrain model was generated before the building of the flood defence. 4. Bridges are removed in the terrain model and there are thus artificial 'gateways' in to the town (where for example there is in fact a flood defence dam under a bridge that closes at high water). This would result in artificial flooding of the town. 5. Rainwater systems and sewerage systems with overflow to the sea are not incorporated in the terrain model. This is generally the case, as the terrain model does not provide information on sub-surface systems. The model will potentially show less serious flooding than might occur, unless an assumption can be made that all pipes and overflow structures are equipped with backflow precautions. If the above limitations are incorporated in the digital elevation model used in the calculation, a stronger tool results, capable of predicting worst-case flooding. The scenario is worst-case, as the calculation does not take account of the fact that it takes time to flood land areas. The model assumes that all areas with contours lower than the water level will flood, but in reality a smaller area will be affected.

A hydrodynamic drainage model (for example MIKE URBAN) describes the flow of water in the drainage system pipes. For the calculation of the areas at risk of flooding from the sea, stretches of pipe that discharge or overflow to the sea and where water can freely rise up in the pipes (without backflow precautions or similar) must be analysed. Running the model highlights drainage wells at which sea water can rise through pipes to the ground surface. The hydrodynamic drainage model only identifies the locations at which water can arise on land and does not therefore predict the potential extent of the flood.

A "cost-distance" analysis allows identification of areas below sea level at specific sea elevations and that are connected to the sea and thus at risk of flooding. This is also called a connection map. The method is a significant advance on the simple contour map in that the significance of flood defences and other barriers is factored in. The digital height model can be supplemented with flood defences, sluices, bridges, water course widths and watercourse underpasses (as described under the method for the connection map). This is important in order to take into account the flow of water towards smaller water courses, under bridges, etc. Drainage systems that discharge into the sea can be incorporated in the height model by 'digging channels' where there are pipes. The result should be viewed with a critical eye, as the water in pipes can in fact only reach the terrain at drainage wells in the system. Instead the model shows flooding in all areas where the terrain contour adjacent to the "channel" is lower than sea level. Areas that are flooded in the model are maximum extent areas in that the flow dynamic on the surface and in channels is not taken into account. The speed of water flow will in fact mean that the entirety of a stated area will not in reality be flooded for a given maximum water level, because it takes time for the water to flow over the terrain and up through pipes.

A digital analysis of flow patterns on the surface allows a reduction of the DTM surface so that it consists of a series of depressions, channels and overflows suited to calculation methods used in dynamic 1-D flow models. In this way, the surface model can be directly combined with the drainage model so that a 1D-1D system is achieved, see figure on the right. Water from the sea is channelled via the runoff system and the terrain via reservoirs and pipes. The model does not give flow speed in terrain reservoirs. The complexity of the generated 1D surface model depends to a high degree on the minimum size of depression included in the model. The number of depressions can be reduced by combining depressions, for example using a volume-limit value for each depression. The level of detail is greatest where limit values are low, but calculation time is increased correspondingly.

This method uses a 2D calculation of the physical water flow on the terrain combined with a digital terrain model. The method, based on physical characteristics, describes in detail the flow and time delay involved when water flows through low-lying areas with a complex topology. The description of flow on terrain uses input from a digital terrain model which describes the topography in the model area. Sea level is the marginal condition which defines the variation of the water level and the subsequent potential flood. Quality assurance of the terrain model must be carried out for the model area, so flood defences, water courses, etc. are correctly included in the terrain model. The calculations result in time-specific, dynamic variations of water depths on the terrain - together with flooded areas. Additionally, information on flow speed can be plotted, for example for an analysis of the impact of water force. The impact in the form of floods from sea levels at various times can be calculated.

A 2D surface model showing flooding on terrain can be combined with a 1D drainage model, thus linking information on sea flooding over terrain with flooding via the drainage system. This is important where pipe systems traverse an area with low terrain contours. The advantage of using this model is that areas at risk of flooding can be mapped that might otherwise be falsely identified as risk-free. The drainage model can be reduced to comprise areas where the system has the potential to contribute to terrain flooding. This implies the omission of areas that lie significantly higher than the expected maximum sea water level in the linking between the two models. The calculations result in time-specific, dynamic variations of water depths on the terrain - together with flooded areas. Information on flow speed is included, for example for an analysis of the impact of water force, as well as information on water channelling and water levels in the drainage system.