IS-ENES 2 coding sprint 20-22 June 2016, KNMI

IS-ENES2 organized a code sprint to work on the climate4impact portal.
The sprint took place from 20 till 22 June, 2016-06-20/2016-06-22 at KNMI, the Netherlands.


  • Christian Page (CERFACS, France)
  • Manuel Vega Saldarriaga (University of Cantabria, Spain)
  • Fokke de Jong (Alterra, the Netherlands)
  • Ronald Hutjes (Alterra, the Netherlands)
  • Wim Som de Cerff (KNMI, the Netherlands)
  • Ernst de Vreede (KNMI, the Netherlands)
  • Maarten Plieger (KNMI, the Netherlands)


  • Day 1: Set the items to work on (see white board picture) and defined the multi model time series graph use case. This is a use case derived from discussions with users of the portal. Started working on the issues. Also work was identified to integrate the C4I search into the Cantabria downscaling portal. Documentation (Howto sections, Use cases, homepage) were also updated.
  • Day 2: Coding sprint; worked on the issues identified.
  • Day 3: Coding sprint, wrap-up & closing.

Results can be seen on the climate4impact portal and on the whiteboard picture below (read check-mark).

Use case: Create a multi-model time series graph

Jack wants to create a times eries for a certain region and period from five models for temperature at air surface. He would do this first for one scenario: RCP8.5. The result for the use case would be both an image and a CSV file. The use case needs to be reproducible. One ensemble model per model is suitable for this case.

The search could for example be:

This returns monthly data, it would need to be averaged to yearly data.

He would need to do time selection, spatial subsetting, averaging and visualization to create the multimodel timeseries graph.

This use case can be executed with the following WPS components:

  • [X] Data check and filter - Input is search query parameters.
    • Does time_subsetting: Filter the filelist
    • Result from this process is a set of filelists. Each set will become a separate process. Each filelist resembles one "stream", e.g only concatinating over time is needed. Before starting the process, the user should be informed about what is going to be calculated.
The following processes accept one file or a list of files split over time:
  1. [OK] WPS time_spatial_subsetting : Extract a region specified by a bbox from a bigger grid
  2. [OK] WPS yearly_average : Calculate average and aggregate files to singles
  3. [X] WPS spatial_average
  4. [X] WPS multitimeseriesaverage
  5. [X] WPS visualize

usecase_multimodeltimeseries_20160622.jpg View (165 KB) Maarten Plieger, 06/22/2016 08:57 AM

codingsprint20160622_whiteboard.jpg View (416 KB) Maarten Plieger, 06/22/2016 03:37 PM