For data upload during the experiment server name, user and password will be provided within each notification mail message (see: notification mail).

Data can be uploaded to the server as follows:

We can offer only very limited support for automated data upload from your organization towards the secure web interface and recommend, especially if you are not familiar with perl, to contact your own System Administrator on this issue, providing him with our documentation. Besides perl, Windows-based clients do exist that can achieve the same result.

Procedures: Notification mail

The e-mail message provides information on

• Date and time of the request
• Responsible officer, IDC point of contact
• Information on the server (location, user name, password)
• Cross-reference to the request document put on the web (allows for authentication)
• All required information on the computations
• Number of stations for which source-receptor matrix computations shall be provided
• Coordinates of the receptor locations
• Measurement start/stop
• Date/time (backward from measurement stop) until which computations are requested
• Expected response time of the participant (usually 24 hours for each request)
• Response form (to be modified only in case of non-operational participation or non-participation)

As soon as the computations are finished, each participant is requested to upload the requested source-receptor matrices in the requested form (see formats and naming conventions) to the server specified in the e-mail notification message. There is no further message necessary from the participant, since the server incoming area is continuously monitored by software that notifies IDC atmospheric transport operations as soon as new data are arriving.

In case of problems that prevent the participant from the submission of results within the time limits, he is asked to notify the IDC point of contact via e-mail (or, if necessary, via other means of communication).

What do we need - source receptor relationship

The IDC requests from the participants information on the so-called "source-receptor relationship" for a number of stations with specific measurement times. This information is needed, for example, to get information on possible sources. For one single measurement at one station, the source-receptor relationship is nothing else than the sensitivity of the measurement to a source at a certain position and time. We need this sensitivity for each point on the globe (1 degree resolution) and for each time increment starting from the end of measurements (collection stop) and going backward to a certain time (e.g. 144 hours; calculation end date/time is specified in the notification mail). The time increment will be 3 hours.

The following shall be (implicitly) assumed:

Source-grid: 1 degree globally

Source duration (for backward modellers: adjoint concentration output sampling interval): 3 hours

Source strength: 1.3 10¹⁵ Bq

Backward simulations (recommended):

In a backward (adjoint) simulation using an Atmospheric transport model, the required source-receptor sensitivity can be easily obtained as follows

• Simulate an adjoint release of 1.3 10¹⁵ Bq of a tracer (no deposition, no decay) at a constant rate at the point of the station location backward from collection stop to collection start (usually 24 hours duration)
• Calculate the respective adjoint tracer concentrations [Bq/m3] at a global 1x1 degree grid, output frequency 3 hours, time average of output 3 hours
• Simulate backwards in time to the requested ending date/time (usually 6-10 days from sample collection stop)

Forward simulations are totally valid, but usually computationally inefficient to address our problem (as long as the number of measurement stations is small). Anyway, the following approach would be equivalent to our backward simulation:

• You have to release material forward in time from a (pseudo-) source grid and sample the material from all pseudo-sources at the receptor location during the whole measurement (sampling) period (from collection start to stop; this is usually a 24-hours concentration average)
• The release rate at each grid point should again be 1.3 10¹⁵ Bq
• The (forward) release duration should be, as implicitely assumed in the backward computations, 3 hours
• The pseudo source grid calculations have to be repeated every 3 hours and for each grid point the receptor may be sensitive to (changes with time of the assumed release; the longer ago the release was, the larger this source grid area has to be assumed). The first computations have to start 3 hours before collection stop time with duration 3 hours. Then from 6-3 hours before collection stop, and so on, until you reach the requested end time (for example 144-141 hours backward).
• What we need from you as output (source-receptor relationship) is the resulting (forward!) concentration at the receptor location (please note that this should be an average from collection start to stop) from the release at the respective grid point and time [in Bq/m3]. This is exactly the sensitivity of the sample to a release at the respective point during the respective time (and thus equivalent to the adjoint tracer concentration in the backward runs at this point and this time).
• Please assure that your capture all points the receptor is sensitive to during the whole time period under consideration

File formats and names

Example file (click to download)

Longitude of station / latitude of station/ measurement (collection) start time/ Measurement stop time [UTC]/ Total mass [Bq] released in backward model / Maximum number of hours backward/ Output frequency [hours] / averaging time [hours] / resolution of output grid in longitude / latitude [degrees] / station identifier [as specified in notification mail message]

[there is no specific number format required]

Latitude of grid point / longitude of grid point / backward time step number / value of source-receptor sensitivity [Bq/m3]

In our example, timestep number 1 is 3-0 hours before sample collection stop; timestep number 2 is 6-3 hours before sample collection stop; and so on

Please leave away all zero-value grid points

You can write out the data in any order you wish to

(data loop can go across longitude, latitude or time, you can start with collection stop and go backward in time (forward in time step number) or with the start of model simulation and go forward in time (backward in time step number).

Please unix-compress all files before upload

You should name your file as follows:

During the full-scale experiment, we will request, in total, on the order of 20-30 computations. You can expect 5-10 stations (measurements) per request. To facilitate data upload, we open the opportunity to submit standard zip (archive) files or unix tar files to the server. The following procedures should be followed:

• please name these files XXXX.nn.zip [or .tar, if a tar file], with XXXX being your 4-digit centre identifier, and nn being a consecutive number (01 for the first file you submit, and so on)
• If you use zip, please make sure that we can extract files from the archive with the standard (unix) unzip command (no special WIN-ZIP !!!)
• The content of the zip/archive files should strictly follow the defined file naming conventions

THIS IS ONLY AN OPTION, FOR YOUR CONVENIENCE. IF YOU WANT, YOU CAN OF COURSE GO ON SUBMITTING SINGLE FILES !!!