GRWG/GDWG Web Meeting 2014-07-23

GSICS Web Meeting on Inter-Calibration Products to Primary GSICS References

Agenda

1. Tim Hewison (EUMETSAT) – Background Motivation & Overall Proposal
2. Tim Hewison (EUMETSAT) – Correcting the Corrections - Application of Delta Corrections to current GSICS Products
3. Tim Hewison (EUMETSAT) – Blending inter-calibration with multiple references
4. Viju John (EUMETSAT) – Implications for Archive Re-Calibration products
5. Masaya Takahashi (JMA) – Proposal of new GSICS product convention for a primary reference
6. Discussion (All) - The way forward - Asking the Users

Attendees

CMA: Scott Hu

EUMETSAT: Tim Hewison (Chair), Viju John, Rob Roebeling

JMA: Masaya Takahashi, Arata Okuyama

NOAA: Fred Wu, Fangfang Yu, Manik Bali

Summary

Tim Hewison presented the following proposals:

1. To define one Primary GSICS Reference for each spectral band
2. To define Delta Corrections based on the mean double difference of the coefficients of GSICS Re-Analysis Corrections using different reference instruments,
   averaged over the reference instruments' overlap period
3. To blend the various GSICS Corrections using different reference instruments together to generate a Prime GSICS Correction, traceable to the primary reference.

In the future the concept could be extended to obtain SI traceability if such a reference instrument becomes available - either using the double difference of GSICS Corrections, or by applying the SRFs to direct differences of hyperspectral instruments obtained by SNOs.

It was pointed out that after the overlap period, the traceability chain is strictly broken. This means we need to characterise the uncertainty growth - particularly in the case where there is a significant trend in the references' double differences.

The necessary stationarity tests should be applied to time series of double differences of GSICS Corrections using AIRS and IASI. [Masaya Takahashi (JMA) plans to investigate this.] They could also be applied to time series of IASI-A/B double differences after applying synthetic jumps.

We also discussed applying the concept to blend results from different inter-calibration methods (e.g. for GEO-LEO VIS products). It was noted that in general the uncertainty of the blended product could be larger than that of the best method, if the methods are applicable to different scenes, as this could cause them to underestimate the uncertainty when applying their results to more general scenes.

Rob Roebeling delivered Viju John's presentation on the impact for Archive Re-Calibration products. This analysis will need to account for diurnal variations in the geostationary imagers' calibration. It was greatly appreciated that NOAA (Manik Bali) plans to investigate this for both GOES and Meteosat imagers.

Masaya Takahashi (JMA) proposed changes to the content of the netCDF files and GSICS servers to accommodate the changes proposed above. Masaya's recommended solutions were approved in principle, but the details will be discussed further by email.

Outcome

Action: Tim Hewison to circulate proposed scheme for scoring reference instrument.

Action: Tim Hewison and Manik Bali to prepare presentation to explain proposed new GSICS Prime Corrections for Users Workshop.

Action: Tim Hewison to coordinate inputs for a white paper on comparisons of IR reference instruments.
This could include analysis of double differences from different geostationary imagers and reference instruments, including multiple FY-2 instruments by Xu Na (CMA), and from NWP+RTM models (NOAA MICROS team). It should also include a validation of the delta corrections using direct SNO comparisons from different LEO reference instruments (e.g. Dave Tobin).