Documentation of uploaded files, analysis of output data, protocol changes

Posted on Aug. 24, 2012

1. Identified problems with uploaded data

First of all, we are delighted to see that by now 13 modeling groups have already submitted output data to our data pool at DKRZ.

Along with the delivered files we have discovered some issues regarding several aspects of the files integrity among the majority of models. Most often headers (i.e. netCDF meta data) are not consistent with the specifications and the example header given in our simulation protocol. Other files are not split up and sorted correctly into the designated folders.

In order to have an overview of the issues we are aware of, we have created an online spreadsheet, which we’d like to share with you. This document will be kept as up-to-date as possible and will help us clarify problems faster without losing track. It should strengthen your attention and point you to check the files carefully before they are submitted. If you think we've made a mistake or have questions about any issues your data connected with your data, please contact our colleague Matthias Büchner via or email We always accept example files sent by email or placed on the vre1 server or a website.

In the end it will hopefully be much easier to include all your data into the analyses if the files are correctly formatted. Please help us in obtaining reliable outputs and visit this page on a regular basis.

You can access the spreadsheet via the following link:

on our website -> Output data or via the News & Events entry on the start page.

NOTE: Unfortunately, write permissions on the vre1 server were accidentally changed by DKRZ such that you may not be able to upload any data at the moment. DKRZ are trying to fix this problem, but it may persist over the weekend. We sincerely apologize and ask for your patience.

2. Please treat current output data as preliminary and use with caution for analysis

When downloading output data from the DKRZ server to use for your analysis, please keep in mind that some of the output data currently available on the server will be frequently replaced by the model groups (see above). Therefore, please treat this data as preliminary for now and use it with caution. 

It would be expedient if you could check in the interim, if new output data was submitted. You can do this easily by using the link on our website -> Output data -> output data statistics. These files are updated daily, and include a detailed list of all available output files, and a list of newly added files.

3. ISI-MIP Analysis Workshop

For those groups who are not allocated a presentation slot in one of the cross-sectoral sessions the workshop, it would be useful to bring a few slides along that summarise the main results from your model so far. This should not be a presentation; just any plots you think will be beneficial to have on hand for the sectoral discussion rounds.

4.  Protocol changes

Please also note that two minor corrections were made in the simulation protocol, regarding the file ending of NetCDF4 format (nc4 instead of nc, sections 5.2.1 and 5.2.3) and a change in variable name for the health sector (section 6.4) - please see -> Simulation protocol.