SST climate dataset paper accepted

Merchant C. J., O. Embury, N. A. Rayner, D. I. Berry, G. K. Corlett, K. Lean, K. L. Veal, E. C. Kent, D. T. Llewellyn-Jones, J. J. Remedios and R. Saunders, A twenty-year independent record of sea surface temperature for climate from Along Track Scanning Radiometers, accepted for J Geophys Res in July 2012.

Analysis is based on the ARC SST dataset which acts as the reference point for forward modelling and SST retrieval within SST CCI.

Regridding tool progress

Discussion with Ralf Quast, Stuart MacCallum and me on status of regridding tool, since Stu has applications requiring its use as soon as possible.

Ralf explain the use case currently implemented, which doesn't quite match. It seems worthwhile for Stuart to shuffle priorities and use the tool in intermediate stages of development that will be still fit for his immediate purposes and available in the next few weeks.

The conclusions were:

1. Ralf will advise Stu tomorrow (Fri 10 Aug) on what route for advancing the tool he plans to follow (whether further develop current code or switch to extension of new Beam rebinning module).
2. Assuming choice to base it on the beam module, we agreed a useful order for development would be:
(i) get it working for ARC input data using currently available aggregation -- and let Stu know so he can play with it
(ii) implement CCI method for averaging (using OSTIA climatology as reference) -- and send to Stu so he can use this seriously
(iii) implement the CCI method for uncertainty propagation -- this requires us to finalise the method and LUTs etc, and we need to have a telecom including Nick Rayner soon to do that
(iv) (if possible) allow an option for an alternative to the OSTIA climatology -- again, let Stu know, and he can then use the ARC-based climatology when required
3. Stu will send Ralf some ARC data files as examples
4. Stu will review the tool specs and may also comment on the approaches when the further discussions come
5. Output of regridded time series as CF-compliant NetCDF is fine, it doesn't have to be in full SST CCI L3 format as outputs from this tool are not CCI products
6. Confirm it does need to cope with input L2P for use on AVHRR GAC

Update on "demo product" strategy

We revised the demo product strategy from that originally proposed because of instrument failures that prevented pursuing the original plans (see here). Getting closer to this work, Alison McLaren pointed out that the notes don't discuss the depths intended for the analysis products for the demo periods.

So, this note simply clarifies this for the record. It is straightforward, since the choice arises from the purpose for each of the revised demo periods.

The "4i" option adds passive microwave sensors to the long-term ECV product, in effect, and will of interest to see their impact. Clearly, this has to include an analysis product of the same type as the long-term record -- i.e., sst-depth.

The "4ii" option covers Metop AVHRR 0.05deg, AATSR L2P and SEVIRI, all of which deliver true skin SSTs. The main purpose of this demo is simply to have experience of handling these datasets (which are larger volume than the long-term AVHRR GAC and ATSR 0.05deg data). The outputs will be compared with operational OSTIA, and therefore the analysis should be a foundation analysis, obtained by applying the same methods to skin SSTs as currently used in OSTIA operationally.

How to estimate SST in SST CCI?

The "Algorithm Selection" process has been discussed many times on this blog (here, here and here, for example). Now I can present the conclusions.

First, some general comments. It was great to get some external participation, and to be able to make very clean comparisons between different methods of estimating SST from space on common data, where we had controlled the procedure for comparison against validation data that had not been used in algorithm development by any party.

It turned out that there was tough competition between the algorithms in terms of the quantitative metrics. Just to recap: we generated statistics and maps of "bias" (taken as the mean SST-depth minus drifting buoy difference), "precision" (standard deviations of the same), and "SST sensitivity", and also various measures of stability (with respect to trend, seasons and day-night).

Generally, all the considered algorithms were good. Sometimes one would perform better on a particular metric for a given sensor, but on a different sensor, the ranking would be reversed.

For the ATSRs, the choice came down to using either the ARC-project coefficients, or a version of optimal estimation tuned to the ARC-project coefficients. Both were independent of in situ, therefore. The optimal estimation had a slight edge on precision, but otherwise there was no clear advantage that one consistently had over another across the full range of application.

For the AVHRRs, the choice came down to the same optimal estimation or Boris Petrenko's incremental regression approach. Incremental regression gave better precision, but poorer sensitivity, so there was an aspect of trade off there. For night-time cases, overall both were very similar on quantitative metrics. For day time cases, optimal estimation was a little better on bias and sensitivity, not as good, as mentioned, on precision.

So, in the end, the only very clear deciding factor was that incremental regression is an empirical approach, tuned to in situ measurements. Optimal estimation, being tuned to independent ARC SSTs, retains independence from in situ measurements -- this being an advantage for a significant minority of climate users of SST (according to our earlier survey).

Having opted for optimal estimation for the AVHRRs, it then seemed preferable to make the same choice for the ATSRs, to maximise the consistency of approach across the sensors. (The only exception will be ATSR-1, for which an version of optimal estimation is yet to be developed.)

Summary of Round Robin

From Gary Corlett: a summary of the Round robin participation process

The SST_CCI project is part of the ESA Climate Change Initiative, which aims to produce and validate sea surface temperature (SST) SST essential climate variable (ECV) data products.

In order to identify the best performing algorithm or combination of algorithms, the SST_CCI project held an open round-robin (RR) algorithm intercomparison and product validation exercise following the protocol defined in this document and using the selection criteria defined in the Product Validation Plan (PVP, RD.216, Section 4). By maximising the number of users participating in the Round Robin exercise, ESA expects to identify the best algorithms for a future operational system.

The chosen algorithm(s) will then be implemented in an end-to-end system to generate the first SST_CCI data records. It is expected that future algorithm selection exercises will be carried out for each subsequent reprocessing to ensure the best performing algorithm is always implemented.

Although participation in the ESA SST_CCI RR was open to all, participants did have to follow a protocol (PVP, RD.216, Appendix C), which defined what was expected of each participant, how the RR would be run, how results should be submitted and what happens next.

All participants (including those internal to the ESA SST_CCI project team) were provided with a set of multi-sensor match-up dataset (MMD) files containing all of the necessary information to run their retrievals against. Each MMD match-up contained multiple satellite image extracts (roughly 11 km by 11 km) matched to a temperature history from an in situ platform.

Only match-ups to drifting buoys were used for the RR and the drifting buoy match-ups were split into four categories: (1) training, (2) test, (3) selection and (4) validation. The training and test match-ups were provided at the start of the RR where the ‘training’ data could be used to tune a retrieval algorithm and the ‘test’ data were for the participant to evaluate their algorithms on an independent subset.

Towards the end of the RR period the participants were then supplied with the ‘selection’ match-ups, but this time the in situ measurements were withheld so the final choice of algorithms could be run on a blind sub-set. Participants then submitted SST estimates with appropriate uncertainties, which were then combined with the selection in situ data into the final RR data package for algorithm selection. The data package was then passed to the Science Team, who will carry out the final algorithm selection against the criteria defined in the PVP (RD.216).

In total, 10 research groups expressed interest in participating in the ESA SST_CCI RR, and contacts were made with Carol Anne Clayson (FSU now WHIO), Jim Cummings (US Navy), George Kruger (BoM), Haiyan Huang (Oxford), Bob Evans (Miami), Ajoy Kumar (Millersville), Igor Tomazic (ZIMO), Rene Preusker (FuB), Boris Petrenko (NOAA NESDIS) and Caroline Cox (RAL).

Of these potential participants only Petrenko and Cox submitted results to the final selection process (Cummings submission was too late for consideration). Unfortunately all other groups were unable to participate in the end due to other project commitments.

RD.216: SST_CCI Product Validation Plan, SST_CCI-PVP-UoL-001 (Oct 2011)

The current version of each reference document is available via the SST CCI web pages at http://www.esa-sst-cci.org/.

SST CCI at GHRSST

GHRSST is the "Group for High Resolution Sea Surface Temperature", and is an international collaboration based on task-sharing among producers of SST information. While focussed on operational distribution of SST information for weather forecasting etc, there is a climate angle to GHRSST via a working group on Climate Data Records, which I chair along with Jon Mittaz of NOAA.

The 13th GHRSST meeting happened in Tokyo last week. SST CCI work was prominent, since the project is supplying many experimental ideas and techniques, the more successful of which will in due course benefit the whole GHRSST community.

As a team, our SST CCI experience convinces us of the power of the innovative SST CCI multi-sensor match-up system. I explained how multi-sensor techniques are relevant to maintaining climate-quality independent SST between the ATSR and SLSTR, now that Envisat has been lost.  

The role of SST CCI in extending the ARC v1.1 L3C dataset to the end of the mission was explained by Owen Embury. Owen also explained about the future SST CCI L3C ATSR products -- how they have several improvements and will be at higher spatial resolution in response to user requests.

The scientific direction of GHRSST is often driven from GHRSST's Technical Advisory Groups. SST CCI featured as follows:

The SST CCI high latitude MMD and cloud/clear/ice classification was presented in the High Latitude TAG by me. The availability for wider use of the Hi Lat MMD was emphasised.

The SST CCI Algorithm Selection was presented in the Estimation and Retrieval WG, and again the central role of the MMS was discussed. The EARWG endorsed the multi-sensor approach, and sees a more comprehensive MMS as a GHRSST community aspiration.

The SST CCI Product Validation Plan was promoted to the attendees of the ST-VAL (Validation TAG) by Gary Corlett, since we are soliciting community feedback on it, as per the SST CCI user consultation plan.

The SST CCI PVP also featured in the CDR-TAG discussions, as did the Algorithm Selection metrics. SST CCI will undertake new approaches to evaluating climate datasets, and again, the multi-sensor matchup system is central to how we do so. The issue of whether GHRSST wants collectively to build a facility with full multi-sensor capability naturally arose in the CDR discussion, as a way of supporting the evaluation of climate data sets by GHRSST for the benefit of users. Also in the CDR-TAG session, Jonah Roberts-Jones presented L4 developments that include results obtained within SST CCI. 

Images of a climate data system -- updated

Algorithm Selection underway

Previous posts have mentioned the "Round Robin" process, by which different options for estimating sea surface temperature are objectively compared using common data. This has depended on having other SST investigators work with our data distribution and submit their SSTs. A big thank-you to Caroline Cox (Rutherford Appleton Lab) and Boris Petrenko (NOAA) for participating in a timely fashion!

We have overlapping results to compare for:
ATSR2 and AATSR (from Caroline Cox and SST CCI project-team)
AVHRR 18 to 19 and Metop-A (from Boris Petrenko and SST CCI project team)

For these sensors there are typically different channel combinations for day and night, and several algorithms, so the inter-comparison is a complex exercise.

Ocean temperature angel?

As part of the SST CCI, Steinar Eastwood has been scrutinizing a lot of high latitude images to create a data base of observations classified into clear-sky-ocean, clear-sky-ice, cloud, cloud-shadow-over-ice, etc. This will be used to improve the classification step for ice areas, so that, for example, an "SST" is not inadvertently obtained for a location with significant ice (which may be a very different temperature). In the course of this painstaking work, Steinar was startled to see, in the pattern of warm and cold ocean surface temperatures, a figure with wings and a halo, playing a trumpet ...

The image shows in false colour the Greenland coast (bottom), and area of low cloud (appearing red) and, in the top portion of the image, the variations in sea surface temperature encoded as different brightnesses.