Studying the Urumea River

The article you are about to read was published at Environmental Expert:

A flood study of the Urumea River in The Basque Country

Source: DHI Water & Environment

The Urumea river catchment area which is located in the Basque Country in the northern part of Spain has historically been seriously affected by flooding. The districts most affected by the frequent inundations are the municipalities of San Sebastian, Astigarraga and Hernani. As recently as October 2008 a minor flood was experienced in Hernani. In 2006 and 2004 major flood events with significant damage occurred.

To improve this situation, the Department of Environmental Affairs of the Basque government will in the coming years invest 60 million € in construction of 18 preventive measures. The overall objective of the project is to 'Let the river breathe' and consequently reduce the frequency of flooding of the Urumea River.

The preventive measures include substitution of a number of exciting bridges, which at the moment are obstructing the river flow, establishment of by-pass channels, enhancements of the riverbed in defined sections, and creation of public recreational areas. All actions amid at increasing the flow capacity of the river, while respecting the environment and natural flow path of the river.

In order to investigate in detail these 18 predefined interventions, DHI was contracted to set up a 2D model (MIKE FLOOD) of the 15 km stretch from Hernani in the south to San Sebastian in the north.

In the First phase of the study a model of the actual situation was established, incorporating all relevant bridges and hydraulic structures. The model was calibrated against observed data from the flood event that occurred in the winter of 2004 and it was verified against the smaller flood event in 2006.

Following the calibration/verification the model was executed for the three return periods T=10, 100, and 500 years.

In the second phase the 18 corrective measures was incorporated into the model, and simulations of the T=10, 100, and 500 years return periods was carried out.

The projects design return period was set to T=500 year, which means that the planed flood protection measures should be capable of withstanding a flood that statistically should happen once every 500 years. (Which could be tomorrow!)

The result of the 500 year event simulation is shown below: The green color shows the extent of the flood in the actual situation, while the blue color represents the future situation with the improvement in place.

Comparison of the simulation results of the actual situation and the future situation when all the improvements are in place.

It is obvious that the flood prevention measures will have the desired effect; - the simulated 500 year flooding will not extend beyond the defined floodplain (yellow line).

The third phase of the study consisted in visualization the results as video animation, which was used for the Basque government press releases and information material. The videos were generated partly by incorporating the results in Google Earth using the DHI GE plug-in and partly by using the Result viewer.

Notes regarding the model configuration.
The bathymetry/topography of the floodplains was defined using available LIDAR data (1x1m), while the description of the riverbed was based on HEC-RAS cross section data. The HEC RAS data was first converted into a MIKE11 model, which then was used to create a bathymetry specifically for the riverbed. This riverbed bathymetry was then “burnt” into the floodplain bathymetry, and thereby generating a bathymetry that integrates the LIDAR data with the detailed data from the HEC RAS model.

Fifteen bridges were incorporated into the model. The description of the flow through (and over) the bridges was accomplished either by directly incorporating the bridge pillars into the bathymetry or by modeling the bridges in MIKE11 and then linking then dynamically to the 2D model via MIKEFLOOD. The decision on which one of the two methods to apply was taken based on the physical dimensions and the design of each bridge.

The model was calibrated against observed data from the flood event that occurred in the winter of 2004 and it was verified against the smaller flood event in 2006.

The model displayed an excellent correspondence with the observed data, except in the lower part of the river. This inconsistency was suspected to derivate from erosion of riverbed that will occur in the real life situation. Meaning that, the concrete erosion of the riverbed material will increase the rivers flow capacity and thereby function as natural flood prevention measure.

The phenomenon was investigated a bit further by use of a simple relation between the calculated maximum water velocity and the erosion depth. The figure below shows a specific river cross-section without erosion (the brown line) and with erosion (grey line). The red line represents the flow velocity.

The river cross sections with sedimentation and without sedimentation.

The results of the two simulations for the 2004 event are shown below, - without erosion to the left and with erosion to the right-. As it can be observed, the situation without considering erosion exhibits a much larger inundation than the situation where the erosion is taken into consideration.

If we compare the simulated flood maps with the actual flood event of 2004, the simulation with erosion (right) correspond almost perfectly to what was observed during the flood, while the simulation without, overestimates the floods.

Results of the simulations of the 2004 event: Left: without erosion, right: with erosion.

The conclusion of the analysis was that the suspicion is largely supported by the modeling results, but considering the very simple erosion description applied, the phenomenon should be investigated more thoroughly using real sediment data and a more sophisticated model 3 dimensional model (MIKE 3 FM).

.... ... .