...
Archived data set | Data set | Data types |
Astronomical data | Observed time series | NetCDF-CF timeseries |
Observed series | Observed time series | NetCDF-CF timeseries |
Surge forecasts | External Forecast time series | NetCDF-CF timeseries |
Forecaster Notes on various level | Messages | text |
Delft-FEWS configuration changes | Configuration | zip (FEWS-configuration) |
Rating Curves | Rating Curves | xml (FEWS PI-rating) |
NWP forecasts | External Forecast time series | NetCDF-CF timeseries |
| Simulations (forecast) | Simulated timeseries | NetCDF-CF timeseries Model States Modifiers Reports |
| Products as handed over to downstream users | Products | Products in original format |
| Local datastore | Snapshot | Local datastore |
Note: External data which is archived elsewhere should in general not be archived within the context of the Delft-FEWS application.
Events
Within Forecasting organisations, events are situations that can be distinguished from business as ususal as something special or interesting is happening, e.g. a flood, drought or calamity. Generally the operations during an event is hectic and no one wants to spent time on data archiving. however, after the event the archived data may be crucial for legal inquiries, performance assessment, or model calibration or to use as input for training. Therefore, the concept of Event has been introduced to label a period of interest for a certain area, such that one easily can find this data back.
Many of the important data sets stored in a Deltares Archive associated with a real time system can be organized in events, this is a special type of metadata overarching one or multiple data sets. The event type, area and period covered by the event determine which data sets belong to an event. Each event type will include one or multiple data sets. Typical event types within a real time application are:
- Historic Event: This event type contains observed time series that are stored in the archive. When a new historic event is made the data include in the historic event period will become available as historic event in the Delft-FEWS operator client and can be viewed together with the current observed or simulated data.
- Review Event: This event type contains may contain observed time series, external forecast time series, simulated time series, modifiers and messages. When a new review event type is made the data sets covered will be tagged and can be downloaded to a Delft-FEWS Stand Alone for post event review purposes.
- Flood Watch Event: This event type contains observed time series and external forecast time series. When a new flood watch event type is made the data sets covered will be tagged and can be downloaded to a Delft-FEWS Stand Alone for skill perfomance analysis
- Watercoach/Training Event: This event type contains observed time series, external forecast time series, simulated time series, modifiers, model states and messages. When a new water coach event type is made the data sets covered will be tagged and can be downloaded to a Delft-FEWS Stand Alone. The data downloaded can then be used for making a new Delft-FEWS Water Coach script for training purposes.
Event types can be defined as appropriate, but need to be registered on the archiver server side.
Additional to tagging and an event in the Deltares archive it is also possible to have attachments uploaded and linked to an event in the archive. Creating an event, downloading events and data sets, and attaching attachments to events can be done with the Delft-FEWS Archive display.
Finally, events can play a crucial role in the data management of the archive. Data labelled for a specific type of event may be help longer in the archive compared to data not labelled for an event. E.g. Numerical Weather Predictions may be removed quickly, except for those external forecast that are part of a Flood Watch or training event.
Archive display
All data that is archived can be accessed and downloaded through the Delft-FEWS archive display. The archive display can be started from the Delft-FEWS Explorer from the Tools menu.
...
To search for data in the archive; take the following steps:1
- Select the area
...
- Set the search start and end times.
2 Press the search button on the right part of the display to check if there is any data in the archive for the selected area and period.
...
- Select the type of dataset
- Select the source of the data (if available, from 2015.01 onwards only)
- Press the search button
- Select one or multiple data
...
- sets from the list
...
- for download
- Press the download and import data button to download the data from the archive server and import into the local datastore. If do not want the data imported, you can also press the download (only) button.
to the local computer.
Figure Delft-FEWS Archive display; search in archive and download data
...
Importing archived data into Delft-FEWS stand alone application
Once If you did not import the data is downloaded from the archive (through events or data download); the data can be imported directly as part of the download, you can import the data in the Delft-FEWS Stand Alone system with use of the Manual Forecast display. To import downloaded data; take the following steps:
...
All data that is downloaded and available in the ArchiveDownload folder will be imported in the localdatastore of the Stand Alone system. As the imported data will be imported in the localDatastore of the Deft-FEWS Stand Alone it is advised to delete or rename the existing localdatastore folder and all data in the ArchiveDownload folder before starting a new Archive download and import session. Note that this import activity will bring back the data using the original data administration.
If you want to use the data within a different context, you may also copy the data to the normal Import folders, and setup a normal timeSeriesImport run module with idmapping to bring the data into the Delft-FEWS stand alone application registered under your timeSeriesImport run module.