Climate Research User Requirements and Product Specifications in Phase II SST CCI
User requirements for climate research in Phase II SST CCI were gathered through multiple methods including literature review, questionnaires, and workshops. These requirements were traced with unique identifiers and used to define product specifications for version 2.1 products in the Bridging Phase. Additionally, the CRG applied the data for trend analysis and automated protocols, benefiting CMUG WP4.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. Download presentation by click this link. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
E N D
Presentation Transcript
Breakout presentation Nick Rayner, CRG lead, Integration Meeting, Met Office, 29-31 Oct 2018
Consistency of User Requirements The Phase II SST CCI URD (http://www.esa-sst-cci.org/PUG/pdf/SST_CCI-URD- UKMO-201-Issue_2.1-signed.pdf) presents requirements from climate research users, which were gathered via six methods: 1) a literature review of relevant documents from bodies such as the Global Climate Observing System (GCOS); 2) review of 'lessons learned' information provided by other projects; 3) questionnaires, which asked about future needs for SST data, five years from now; 4) discussion sessions; 5) review of user requirements found in the ESA SST CCI Climate Assessment Report (CAR) from Phase 1, the CMUG requirements documents, and the MyOcean Weather, Seasonal Forecasting and Climate User Requirements Document; and 6) a user workshop on uncertainties. PM11 | Reading | 10/07/2018 | Slide 2
Traceability of requirements Throughout, requirements are each given a unique identifier. The format of the identifier is SST_CCI-UR-SSS-n, where UR stands for User Requirement, SSS is the method of gathering REF: through analysis of reference documents; QUF: from results of the first questionnaire; CAR: from the Climate Assessment Report of the Phase I products; LLP: from lessons learned from other projects; MYO: from the MyOcean User Requirements Document; UWU: through discussion at the User Workshop on Uncertainties; and QUS: from results of the second questionnaire, and n is the user requirement number. PM11 | Reading | 10/07/2018 | Slide 3
Product specifications The SST CCI Phase II PSD (http://www.esa-sst-cci.org/PUG/pdf/SST_CCI-PSD- UKMO-201-Issue-2-signed.pdf) provides descriptions and specifications of version 2.1 products that will be produced during the Bridging Phase. The specifications were originally developed from user requirements defined in the SST CCI User Requirements Documents and have subsequently been revised following experience gained within the project and from trail blazer and other users making use of the data. PM11 | Reading | 10/07/2018 | Slide 4
Use of products by CRG In Phase II, the CRG are using the data in the following ways: Analysis of trends and variability in comparison to other climate SST data sets. This has been packaged up into an automated protocol to apply to new product versions. This also includes a simple autocorrelation metric which may be useful to CMUG WP4.1: Evaluation of modelled system memory when considering differences which might arise when using different data sets to assess system memory. PM11 | Reading | 10/07/2018 | Slide 5
Use of products by CRG In Phase II, the CRG are using the data in the following ways: Dedicated modelling experiments (both low N96/1.25 x 1.875 degree - and high N512/25km - resolution) are currently being analysed. These simulations are comparable to those run under HighResMIP forced by other SST data sets and will be used to assess the impact of the SST CCI analysis in providing lower boundary forcing. This will explore: the impact of any large-scale differences in the SST CCI products cf other SST data sets (e.g. relative biases or differences arising from sea ice data included); and areas of the simulation that are likely to respond to the higher resolution information provided by the SST CCI analysis, e.g. precipitation and wind stress. PM11 | Reading | 10/07/2018 | Slide 6
SST CCI uncertainty estimation L2P and L3U files contain uncertainties broken down into components from errors that correlate on different spatial and temporal scales: uncertainty from effects that are not correlated from location to location (such as random noise in the satellite sensors). uncertainty from effects that are assumed (provisionally) to be correlated over distances of 100 km and 1 day (related to atmospheric conditions). uncertainty from effects that can be assumed to be correlated everywhere and over long time scales (such as over all calibration of the satellite sensor). an adjustment uncertainty is provided which is applicable if using SSTs that have been adjusted to a standard time (10.30/22.30 h) and depth (20 cm); (provisionally) assumed to be correlated over 100 km and 1 day. PM11 | Reading | 10/07/2018 | Slide 7
SST CCI uncertainty estimation For each individual SST, the total uncertainty can be obtained by summing each component in quadrature For applications combining multiple SSTs, such as performing a regional average of the data it is essential that the correlations in the uncertainties are taken into account when combining the uncertainty components. We provide an example of how to do this in the Product User Guide and we explored this in a practical setting in a dedicated User Workshop on Uncertainties (see http://www.esa-sst-cci.org/PUG/workshop.htm ). L4 files contain a single uncertainty field representing the total uncertainty, so it is not possible to consider different correlation length scales in this case See our Uncertainty Characterisation Report (http://www.esa-sst- cci.org/PUG/pdf/SST_CCI-UCR-UOR-201-Issue-2-signed.pdf) for a full discussion. PM11 | Reading | 10/07/2018 | Slide 8
SST CCI uncertainty estimation This advice on uncertainty propagation is key to: CMUG WP4.2: Evaluation of model results considering observational uncertainty when calculating averaged quantities; and CMUG WP4.4: Optimal spatial and temporal scales for model evaluation , since L3C and L4 products have different complexities of uncertainty information and would produce different answers to this question We have tools to propagate uncertainties into regional averages and to regrid, but they are currently sub-optimal and need further improvement this is complementary to these CMUG work packages and they would benefit from interaction as we evolve our thinking. PM11 | Reading | 10/07/2018 | Slide 9
Data production schedule SST CCI v2.0 products, 1981-2016 (L2P, L3U, L3C and L4) are currently accessible to trail blazer users. Once the Climate Assessment Report has been completed, they will be made publicly available Come and speak to me (or email nick.rayner@metoffice.gov.uk) if you want to be a trail blazer user of v2.0 products over the next few months v2.1 products (an update that includes e.g. anomaly fields) will be released after the CAR is complete (March 2019) PM11 | Reading | 10/07/2018 | Slide 10
Data production schedule Subject to successful proposal, v3 products would be available to trailblazers from Q1 of year 3, so these CMUG WPs should plan to use v2.1 products: WP3.8: Evaluation of the impact on skill of Hires-SIR on seasonal prediction (using SST to nudge and to evaluate reforecasts) here our PMW prototype products may be of particular interest WP3.9: Biophysical feedbacks in the global ocean (using SST to evaluate assimilation with and without OC) WP4.7: Skill assessment of the DCPP decadal predictions (evaluating decadal reforecasts) Phase 1 and (some) Phase 2 data are available under C3S, but not yet via the Climate Data Store: ATSR L3U v1.1, 1991-2010 - brokered, not available in CDS AVHRR L2P v1, 1991-2010 - brokered, not available in CDS PM11 | Reading | 10/07/2018 | Slide 11
Data production schedule Phase 1 and (some) Phase 2 data are available under C3S, but not yet via the Climate Data Store: Analysis L4 v1.1, 1991-2010 - brokered, available here: https://cds.climate.copernicus.eu/cdsapp#!/dataset/satellite-sst-esa-cci ATSR L2P+L3U v2, 1991-2012 - brokered, not available in CDS AVHRR (1995 / NOAA-14 onwards) L2P+L3U v2, 1995-2016/7 - brokered, not available in CDS Analysis (1995 onwards) L4 v2 is due by end of October, 1995-2016/7 ATSR and AVHRR L3C v2 will be available soon, 1995-2016/7 Early data AVHRR and Analysis v2 up to 1995 is due next year (after the CAR is complete), 1981-1994 Ongoing processing will also be available from C3S PM11 | Reading | 10/07/2018 | Slide 12
Other feedback from CMUG session The following CMUG work packages also plan to use SST CCI products: WP4.3: Evaluation of model results considering abstraction level of obs . Considerations here include: Use of L4, L3C, or L3U (decreasing levels of processing consider here also the L4 single uncertainty issue) Use of SSTskin or SST0.2m (latter involves modelling) Use of BTs would not improve evaluation of modelled SST WP4.5: Evaluation of model results considering internal variability Assessments of internal variability are data set dependent and detailed knowledge of the data set construction required to understand where they come from e.g. see daily variability assessment in http://www.esa-sst-cci.org/sites/default/files/Documents/public/SST_CCI- CAR-UKMO-001-Issue_1-signed-accepted.pdf) PM11 | Reading | 10/07/2018 | Slide 13
Other feedback from CMUG session WP4.6: Evaluation of model results considering a combination of sources of uncertainty The tasks that contribute to this are likely to yield results that are highly dependent on the way uncertainties have been communicated in products A framework will be difficult to implement in practice without consistency between data sets in uncertainty estimate provision PM11 | Reading | 10/07/2018 | Slide 14
Needs for ECMWF reanalysis data For CCI we need the ERA-5 re-analysis (best estimate or ensemble mean) and have initiated steps collaboratively with NCEO and C3S projects to have that available at CEMS. The fields required are atmospheric variables to support forward modelling and surface forcing variables to drive our skin-depth adjustments. The ensemble could give us error covariance information, which is of interest but not currently planned as a processing approach for practical reasons. PM11 | Reading | 10/07/2018 | Slide 15
Other considerations Complementarity of CRG and CMUG work How will you know if your CRG work is complementary with CMUG research? By attending integration and colocation meetings, discussing plans with colleagues and reading the workplan No duplication detected so far just complementarity By taking part in the climate science working group Meeting attendance over the next 12 months We will attend the next Colocation meeting Chris Merchant will attend the Living Planet Symposium We will attend the annual Group for High Resolution SST science team meeting PM11 | Reading | 10/07/2018 | Slide 16
