Qc2 Workshop 20081121

Workshop Description

There are three broad areas in which Qc2 (and Qc2 work) can be divided:

1. System and Software design,
2. Theory, development and testing of Qc2 algorithms,
3. Operational aspects of defining, running and interacting with Qc2 controls and checks, (includes the specification of new flag values and how we handle these changes!)

The workshop will focus mainly on point NUMBER 3 of the above aspects.

Workshop Minutes

The workshop took place at met.no [09:30 - 11:30] with SMHI attending by telephone link.

Meeting Slides

The main results and actions arising from the workshop:

• Have to discuss and decide a common (SMHI - met.no) software tool for QC2. C++ or Python?

• A QC2 database (or netCDF-files), separated from Kvalobs, to store several versions of QC2-runs with model output.

• Define additional use of this model output database in several aspects for end user or HQC operators.

• A need of extending Kvalobs data table with a column “edittime” to log the last change in original or corrected. (Change in original means that a new version of the original value enters Kvalobs for some reason.)

• Interpolation methods to be further developed. Needs to quantify the confidence of model values, and make use of this confidence as criteria for using the model value or not. To avoid noise.

• Define control flags (fw-flag etc). Use a processflow for flag specification and implementation, to secure interactions with QC1.

• Defined categories of QC2-checks in Kvalobs system documentation are QC2-d (daily) and QC2-m (monthly). It is important to describe the events calling for QC2-actions. This may be data retrieval events, data quality events, user require data events and so on.

• Qc2 development can be divided into three main activities (see the details in the workshop description above). Different resources could be assigned to each area.

• Proposal: Next workshop during last week in January 2009 - extended to SMHI/met.no cooperation on Kvalobs.