Darft Version 1.1 Minutes of the ESEAS-RI WP 3 Task 3 Team Meeting Place: KMS, Copenhagen, Denmark Date: 14 June 2004

1   Summary of T3.1 (Per Knudsen)
2   Presentation of work done in T3.2 (Luciana Fenoglio-Marc)
3   Preliminary results and challenges connected to T3.3 (Niels Kj\x{00E6}r)
4   Discussion - concluding T3.2 - recommendations for T3.3 (All) 

The homepage on Operational Satellite Altimetry were presented. It can be found at: http://manicoral.kms.dk/~eseas/.

Very briefly the main findings of T3.1 were summarized; Sea level height variations are regionally consistent, and for European seas the variability is steady in time (largely due to the influence of the North Atlantic Oscillation NAO), but this seems not to be the case for other areas.

The work done on altimetry in T3.2 and some preliminary results along with an outlook for challenges regarding T3.3 were presented, spurring a general debate on methodologies and recommendations for actions to be performed before finalizing T3.2 (see below). In brief, some of the comments and topics discussed were:

1. Low pass filtering data leads to less degrees of freedom, so a lower correlation coefficient for LPF data might not be worse, depending on the significance (see 'actions' below).
2. Generally only winter values of the NAO should be compared to sea level variations.
3. Lag correlations with the Southern Oscillation Index (SOI) might be considered. Also El Nino and La Nina years have different impacts on sea level variability, so it is interesting to compare years with positive SOI and years with negative.
4. Instead of calculating the spatial correlation of a tide gauge with altimetry, one might also compute the correlation of the T/P groundtrack point closest to the tide gauge with the rest of the T/P points, to learn about the spatial correlation of the altimetry.
5. The effects of low pass filtering the time series and possible aliasing (see 'actions' below).
6. Quality of some of the 'control stations' from T3.1, and thereby their usefulness in further studies, were raised as an issue, as station characteristics was not known to all participants. When comparing tide gauges with altimetry it is important to check that the selected tide gauge is actually representative for the region (see 'actions' below).

In order to conclude T3.2 the following actions were agreed upon, with a deadline of July 8th, which were met by all partners.

1. Significance levels should be calculated for all the correlation coefficients presented (between tide gauges, altimetry and climatic indices).

This has been carried out by P2 and P14, and the relevant figures and tables have been updated.


2. Verification of filtering effects; that low pass filtering the data with yearly averages performed every half year, does not create aliasing.

An example using the data from the tide gauges in Stockholm and Cuxhaven shows clearly that a filtering performed as described above, does not lead to aliasing.


3. Check 'control stations' (as defined in T3.1) for regional consistency.

Indeed, some of the 'control stations' were not in complete agreement with other nearby stations, and we will definitely consider this for our future work in T3.3.