wiki.met.no

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
====== QC2 Acceptance Test & Other Possible Tests ======

Subject: Acceptance test of precipitation re-distribution algorithms in QC2 (if precip. missing for >= 1 days)

Participants: Per-Ove, Solfrid, Ole Einar, Matthias (author of this page)

Not present: Paul

===== Issues discussed: =====

  * Test with different kinds of precipitation data: Convective, terrain (complex/flat), proximity to Swedish border

  * Precipitation measurements have already 10% measurement inaccuracy; good result if we can achieve a certain (yet unknown) percentage accuracy in estimated values

  * Convective case especially challenging: Half of stations could have precipitation, other half not.

  * Automatic QC2 should give better result than HQC --> Could this be tested? (Hypothetical HQC test dataset could be used and compared with automatic QC2)

  * Tests of QC2 should be carried out using historic data. Results should be compared with measurements.

  * We have enough data for extensive evaluation and reliable statistical error estimations.

  * Should HQC still check results of automatic QC2 (probably not feasible)

  * Is QC2 giving better results than values of numerical models that are stored in KDVH. --> Test this


===== Acceptance tests with different datasets: =====

  - Precipitation type (convective/frontal)
  - Regions/topography/catchment areas
  - Station edensity/coverage