toposoft Gesellschaft für Datenbanken und Applikationen mbH

ECO - Evaluation and Correction of ObservationsAutomatic examination and reconstruction of measurement data

Suppose, because of a data transmission error, at 23:30 the daily precipitation is entered as 13mm instead of 1300mm in the precipitation time series. This error is noticed the next morning at 7:45 and would be corrected (with ZRFix). In the meantime, however, the reservoir in the N/A model has been running full to the brim and is slowly spilling out over the next few weeks, generating a large increase in Q. This error, in turn, may not be apparent until a week later. But then the data were published long ago and the worst part is that it calls your authority into question.

To effectively combat these and other problems, we offer our service ECO. It constantly checks the data that Callisto retrieves for plausibility and corrects errors according to set rules. The time series to be checked is assigned a test profile. Profiles are a set of a series of connected rules. Each rule consists of a test method and, if necessary, a correction method.

When Callisto retrieves new data, it sends a test job to the ECO service. This opens a new thread for the execution of the test and is then immediately available for new test requests. Callisto can thus settle several hundred requests per minute. There is also a periodic check service (for example, once a day) and a manually-triggered one. The latter is useful for interactively determining the parameters of validation rules.

Test Algorithms