This Master Configuration File, describes the RISC-KIT Coastal application specific content:

Delft-FEWS Coastal

Based on Delft-FEWS, with the expansions developed for RISC-KIT, a Delft-FEWS Coastal application has been set up for coastal and flash flood applications. This example configuration consists of the model train for the North Sea. To give an impression of the application, we show some screenshots of the application. All screenshots can be clicked upon to enlarge.

Figure: Delft-FEWS Coastal – overview of the North Sea grid and locations

Figure: 2D spectra from Wavewatch III (Global scale) used as forcing for the Delft3D FLOW/WAVE model

Figure: Meteo forecast from HIRLAM meteo (KNMI) (European scale)

Figure: Water level forecast from Delft3D FLOW/WAVE model (Regional scale)with HIRLAM meteo used as forcing 

Figure: Water level (gridded data) from the Xbeach model for Egmond (Local scale), with water level (scalar) forcing from Delft3D FLOW/WAVE model

Model train

The master configuration file consists of the model train used for this example case study site, including the import of all necessary data, set-up of the different models making use of model adapters and visualization of the results. Furthermore, we have added a first version of the configuration of the 11 case study sites. By incorporating this preliminary work, the master configuration gives an overview of all the case study sites, their model grids and their inputs and model outputs. However, these configurations are still very much work in progress, and the final version will be made in Task 5.1 of WP5 (Month 29).

Figure: Model train for the North Sea

Use of Delft-FEWS Coastal

Running workflows

The application is ready to be used in an operational environment, which means all tasks (i.e. Workflows) are run automatically, from importing the necessary (meteo) data to running the complete model train every 6 hours. However, within the scope of the RISC-KIT project, the application is used in a so-called stand-alone mode. This means that every task is started manually. The user starts a task via the Manual Forecast Display, by selecting the proper Workflow (see Figure 5.2). The workflow are displayed in the order they need to be executed. In this example we start with the import of the Nomads meteo, which is used as a forcing for the WavewatchIII workflow. This run supplies the 2D spectra boundary conditions that are used in the Delft3D FLOW/WAVE model that is run next, but not before we have also imported the local Dutch weather forecast HIRLAM v7.2. The model train is completed with an XBeach run, either in hindcast or in forecast mode.

Figure: Select a Workflow in the Manual Forecast Display to run a model

Displaying results

The data can be displayed as Timeseries Data or Spatial Data, depending on the model output. To view Timeseries Data, one can open the Data Viewer and select at least one source, location and parameter. This way, both 1D time series as 2D spectra can be displayed. Gridded data is displayed in the Spatial Data display. In this display, the pre-configured plots are arranged per case study site. Different spatial displays are shown on this wiki, all have been configured for the North Sea model train.

Figure: Select at least one source, location and parameter in the Data Viewer (enlarged in the figure) to display in the Timeseries Data display (on the right)

Configuration Training Material

The following presentation material was used during the Delft-FEWS configuration training:

Master Configuration File - Technical documentation

We have prepared a basic Risc-Kit configuration (Delft-FEWS Coastal), which run some of the most used models / model trains. This basic configuration can be used as a starting point. For each partner, we have linked their model train to existing example xml files. Please note some specifics about this configuration mentioned below before you start. This comments will only make sense after you followed the basic configuration course, with which we will start the first week of Configuration Training.

Naming

This basic configuration is as generic as possible, with no or little model specific information. This means we use generic names where possible (e.g. moduleConfigFiles, locations and locationSets). Also, the displays (e.g. Filter, DisplayGroups, GridDisplay) are relatively straightforward and not specific. This approach works best if the naming convention throughout the configuration is consistent.

Template ModuleConfigFiles

For the ModuleConfigFiles we have made use of the so called Template files, in order to save you work. The template files make use of keywords or $PROPERTIES$. These "variables" are filled in by FEWS on the fly based on the information in the corresponding workflowFile. This means the moduleInstance file is generated automatically each time the workflow is run. This file therefore does not exist as a separate file in the configuration, but still has to be declared in the ModuleInstanceDescriptors.xml. See below for some examples. This approach is only worth your while if the moduleInstance will be multiple times. Therefore, many of the import moduleInstances do not make use of this approach.

Locations and attributes in CSV files

The locations and locationSets are generated as much as possble based on csv files (see MapLayerFiles). This method is explained on the wiki:http://publicwiki.deltares.nl/display/FEWSDOC/22+Locations+and+attributes+defined+in+CSV+files%2C+Shape-DBF+files+or+external+tables

Model Adapters

The model adapters are part of the configuration (Modules/bin/*_adapter/). The adapters are described on several wiki pages.

ModuleDataSetFiles

We keep a copy of the ModuleDataSetFiles in the folder ModuleDataSetFilesUnzip, in order to be able to do file version management. When you need to change (or add) a ModuleDataSetFile, first make a / go to the folder in ModuleDataSetFilesUnzip/ModelX, make your changes and zip this folder to Config/ModuleDataSetFiles, which will be used by FEWS.

  • No labels

Disclaimer