ࡱ>  K;bjbj*.*. 3tHDHD1-q  8L|l?U4^LTTTTTTT,WuZrTTE"TE"E"E"TE"TE"E"BNrRЧUW!P*TU0?UQnZE"ZTrRE"rRSTTE"?UZ :   Limited Distribution IOC/IODE-XXIV/3.4.7 14 December 2016 Original: English INTERGOVERNMENTAL OCEANOGRAPHIC COMMISSION (of UNESCO) Twenty-fourth Session of the IOC Committee on International Oceanographic Data and Information Exchange (IODE-XXIV) Kuala Lumpur, Malaysia, 28-31 March 2017 PROJECT REPORT IODE International Quality-controlled Ocean Database (IODE-IQuOD) Submitted by: Catia Domingues and Matt Palmer catia.domingues@utas.edu.au; matthew.palmer@metoffice.gov.uk 5th December 2016 TITLE OF THE PROJECT: International Quality-controlled Ocean Database (IQuOD) PROJECT ESTABLISHED BY [IODE Recommendation/IOC Decision]:  HYPERLINK "http://www.iode.org/index.php?option=com_content&view=article&id=462:iode-steering-group-for-the-iode-iquod-project&catid=10&Itemid=89" \l "rec23.3" Recommendation IODE-XXIII.3 (2015) Project Leader(s): Co-Chair(s) of the Steering Group: Catia Domingues (Brazil/Australia), Matt Palmer (UK) Project Manager: (if applicable) N/A Project Technical Manager: (if applicable] N/A Members of the project Steering Group: http://www.iode.org/index.php?option=com_oe&task=viewGroupRecord&groupID=323 Rebecca Cowley, Ann Thresher, Susan Wijffels, Catia Domingues (Australia); Simon Good, Matt Palmer, (UK); Toru Suzuki (Japan); Viktor Gouretski (Germany); Tim Boyer, Gustavo Goni, Janet Sprintall, Alison Macdonald (USA); Charles Sun (USA/SG-GTSPP). Objective(s) of the Project: To produce, freely distribute and curate the highest quality, most complete and consistent global ocean subsurface temperature profile repository for Earth system, climate and ocean studies, with intelligent metadata and an uncertainty estimate for every observation. This aim will be achieved through coordination of resources and expertise into a single best practice international community effort. Expected outcome (result) of the Project: (1) Development/implementation of international best standard practices for automated and semi-automated quality control of historical (and modern) subsurface ocean temperature data. (2) Provision of intelligent metadata and uncertainty assignment to each individual observation. Both (1) and (2) involve community agreement on best practices; open-source software development/ documentation/deployment; personnel training (capacity building); and coordinated application of quality control procedures/audits (3) Template for future efforts: great community interest in improving the quality, completeness and consistency of the historical salinity observations and other ocean variables (e.g., oxygen and nutrients). (4) Important ocean data legacy (raw and added-value products) for a broad spectrum of users, which includes numerous downstream applications for Earth system, climate and ocean research and services of great societal benefit such as: Refinement of XBT(MBT) bias corrections; Key indicators for global/regional climate variability and change (e.g., ocean warming and sea-level rise); Development of performance metrics to evaluate climate models used for climate predictions and projections; Ocean model initialization; Synthesis efforts (coupled and ocean data assimilation); Multi-disciplinary (observational and modelling) climate research, including detection and attribution studies. Expected duration of the Project (if open-ended then indicate starting year only) (2013-open): Expected duration is for about 5 years from implementation phase. Actual duration, however, is dependent on funding sources and human resources allocated. URL of the project or of the page in the IODE web site:  HYPERLINK "http://www.iode.org/index.php?option=com_content&view=article&id=462:iode-steering-group-for-the-iode-iquod-project&catid=10&Itemid=89#rec23.3" http://www.iode.org/index.php?option=com_content&view=article&id=462:iode-steering-group-for-the-iode-iquod-project&catid=10&Itemid=89#rec23.3  HYPERLINK "http://www.iquod.org" www.iquod.org Activities implemented between April 2015 and March 2017 (eg meeting of SG, workshop, development of products/services,): Endorsement letters from data assimilation and climate modelling community, April 2015 Establishment of the SCOR-IQuOD working group 148, Dember 2015 http://www.scor-int.org/SCOR_WGs_WG148.htm Dissemination seminars: OOPC meeting, March 2016 CLIVAR Conference in Qingdao, September 2016 Workshops/SG-meetings: 1st IODE-IQuOD workshop, Hamburg-Germany, December 2015 2nd IODE-IQuOD workshop, Tokyo-Japan, October 2016 Open session at AGU Ocean Sciences 2016. Problems experienced during April 2015-March 2017 and measures taken: Funding or volunteers to support programming and other IQuOD activities. During the 2nd IODE-IQuOD workshop in October 2016, we have formally established a task team to actively seek and organize funding opportunities. Results achieved between April 2015 and March 2017: Progress has been made to re-code and further develop the Cowley et al (2013) intelligent metadata (iMeta) algorithm from the original MATLAB code into open-source Python. We are working towards having the IQuOD iMeta datastream implemented in World Ocean Database in early 2017. The Python code will be made available to the community through the online repository discussed immediately below. The iMeta algorithm will be documented in a peer-reviewed paper that is currently in preparation. IQuOD has established an open source software suite for benchmarking automatic quality control checks on GitHub see  HYPERLINK "https://github.com/IQuOD" https://github.com/IQuOD. This contains a framework in which quality control checks can be applied to a set of input profiles. The result of applying each individual check to each profile is output to file. The software is written in a way that it is extremely simple to add new quality control checks in the future. Through international cooperation there are now 49 quality control checks included (many of these are variations of each other, which allows the relative performance of each to be evaluated). The software suite accepts profile data in the commonly used ASCII format in which World Ocean Database (WOD) data are distributed. In order to benchmark the quality control checks we require input profiles of known quality. At the workshop in Hamburg 2015 there was brainstorming of which datasets could be used that meet this stringent criterion. Four of these have been converted to/output from WOD in the WOD ASCII format and will be used for benchmarking the quality control checks. These will imminently be run through the software suite to obtain the quality control check results. Once the quality control checks have been run, the next activity will be to identify which combination of checks are the best. The definition of best will differ according to the application, but in general we want a high rejection rate for profiles containing bad data and a low rejection rate for profiles that do not contain bad data. Software is already in development (which uses machine learning methods) to find these best combinations. The quality control check code is already freely available in the GitHub repository and people are welcome to use the code in their own applications. In the UK, IQuOD members have approached the University of York and put forward PhD proposals on machine learning methods for improvement in climate data sets including IQuOD subsurface ocean profiles. Machine learning approaches to provision on next generation iMeta algorithms will be explored at the UK Met Office with work to begin in early 2017. A first cut estimate of uncertainties has been prepared and further refinements are underway so that we can include them in WOD in the next few months (early 2017). This will build on the values reported in Atkinson et al (2014) and efforts from working group member V. Gouretski, see  HYPERLINK "http://iquod.github.io/uncertainty/uncertainty.html" http://iquod.github.io/uncertainty/uncertainty.html. We will work with the wider oceanographic community to revise and improve these estimates for future IQuOD data releases. Deliverables produced between April 2015 and March 2017: The first element of IQuOD to be delivered will be data streams for iMeta and uncertainty estimates to be served as part of the World Ocean Database using their standard data formats (NetCDF and ASCII). The aim is to deliver these by March 2017. The first full IQuOD database, with automated QC flags applied will be released in late 2017-early 2018. Interim deliverables (as explained in 11): - first cut algorithm for intelligent metadata - first cut estimate of instrumental uncertainties - coding of 49 community-based automatic quality control procedures on github - algorithm for exact duplicates check Should project continue: YES Work plan and Budget for the next period (April 2017 March 2019) (if project should continue): 14.1 Description of activities, expected results and deliverables to be produced during the period April 2017 March 2019 including timeline. 2017-2018 will focus on: Benchmarking of the various AutoQC procedures using a number of high quality regional reference data sets. Benchmarking analysis will identify the most effective combination of AutoQC checks and the work will be submitted to an open access scientific journal. The AutoQCed database, in combination with any advances in random error and intelligent metadata assignment (work in collaboration with the XBT Science Team), will constitute version 2 of the IQuOD database. 2017 workshop will serve to discuss the outcomes above, to share the results of the benchmarking tests and to provide an international forum for a consensus on best practices for AutoQC procedures for subsurface ocean temperature observations. 2018-2019 will focus on: Preparation and submission of scientific papers related to the AutoQC benchmarking exercise; on the publication of version 3 of the IQuOD database with updates and improvements from the previous two years, including related documentation (reference guides and software tools). We will also report on the feasibility of using machine learning (or other novel computational) methods for the expert quality control step, through publication of a discussion article. The 2018 workshop will be organized as a large international workshop for knowledge transfer and capacity building, to encourage rapid and wide adoption of best standards for quality control of historical subsurface ocean temperature profile data, inclusion of intelligent metadata and uncertainty. We will also seek additional funding sources to ensure maximum international participation, particularly from developing countries not yet involved in IQuOD. 14.2 Required financial resources (indicate in-kind human resource contributions We are looking for python programmers (volunteers) to help code extra automatic quality control procedures. In case, IODE could help finance this activity, this could also be an avenue. https://github.com/IQuOD/AutoQC We are also looking for support/exchange of ideas/discussions with IODE-related community with great experience on the development/implementation of capacity building and knowledge transfer activities. (Budget expressed in US Dollars) Activity (as per work plan)Timing (month, year)Total required fundingFunds already identified (and source)Funds requested1Completing AutoQC programmingEarly 2017 2K2K2SG-IQuOD-III and 5th IQuOD workshopLate 2017 (potentially together with IAPSO meeting, South Africa)25K15K (SCOR)10K3SG-IQuOD-IV and 6th IQuOD workshopLate 201825K15K (SCOR)10K Actions requested from the Committee (in addition to work plan/budget) or other comments. Support SG-IQuOD meetings by IOC Project Office for IODE Draft text for the annotated agenda and summary report Through coordination of resources and expertise into a single best practice international community effort, the IQuOD project aims to produce, freely distribute and curate the highest quality, most complete and consistent global ocean subsurface temperature profile repository for Earth system, climate and ocean studies, with (intelligent) metadata and an uncertainty estimate for every observation. Major activities during this inter-sessional period included: (i) development of a "first cut" intelligent metadata algorithm; (ii) development of "first cut" uncertainty estimates; (iii) implementation of exact duplicate checks; (iv) implementation of 49 community-based quality control procedures on github; (v) selection of high quality regional datasets for auto-quality control benchmarking. Any other comments: n/a  Intelligent metadata algorithms are used to infer XBT probe type (and potentially other details) based on known but indirect information, such as maximum recorded depth and country code etc. This is important since about 50% of the historical XBT measurements lack metadata (Abraham et al., 2013).     IOC/IODE-XXIV/3.4.7 Page  PAGE 2 IOC/IODE-XXIV/3.4.7 Page  PAGE 3 PAGE 2 PAGE 3 #)*+;O# $ % & 4 ôwfwwUH;h1h D5B*phh1h#5B*ph h1h9C5B*CJaJph h1h9/5B*CJaJph h1hh1h9CB*CJOJPJQJ^JaJfHphq M N   3|b4!K$@zYgd2gdN & FgdZ & Fgd ^gd & Fgd & Fgd1 & Fgdmrugd9C5 6 7 ! [ a b ~  bcnoxy3DM[\]^ƿƱї}uufޗ[h1hsB*phjh& z0J$B*Uphh 8B*phh1h5B*phh1h#5B*phh1hs5B*phh}hPJnHtHh}hPJnHo(tH h~Bhh h=:hh1h5B*phh1hB*phh1hGHJLMU^aeuɼ铌vh1h#B*phh1hsB*ph h{hhhhh=:hh6 h=:hhh1hhB*phh1htB*phh1hs5B*phh1h#5B*ph hhhy%(hhh hAhh1hB*ph. !-LMNPQstX׻׻xmxbWSOhPLhNh1hNB*phh1htB*phh1h#B*phh1hpB*phh1h#B*CJaJphh1hpB*CJaJphh1h & FgdZ & Fgdmru & FgdPLgdN & F gd2') ""@#F#%%%%S&`&&&'''R'S''''( ( (춲xmbh1hcJB*phh1hh>jhcJ0JUh;JhcJ0JjfhcJUjhcJUhcJhXhhqhjh>0JUh;Jh>0Jjh>Ujh>Uhvu*h2rh>h1h>B*ph& (k)))n*o*v*~************}+++++, -;-l-.../0 00E12ζΖ~s~~ff\fffffhyB*\phh1hkbB*\phh1hdB*phh1hmruB*phh1h(5B*phh1hd5B*phhyw4hmruB*phh B*phh1h#B*phh1hmru5B*phh1hCJPJaJnHtHh=:h CJaJh=:hyw4CJaJh=:hy%(CJaJh1hdB*CJaJphh1hHcOB*CJaJphh1h(B*CJaJphh1h9B*phh9h9B*phh1h{mtB*phh1hmruB*phh1hkbB*phh1h9&AB*ph44kdW$$Iflֈr7  n t0:!44 lap<ytyw444444444 $Ifgdg444444444444444555555P5Q5R5S5T5U5_5`5b5c5d5e5f5g5p5wiӤ[iӤh_CJPJaJnHtHh1vCJPJaJnHtH h=:hwCJPJaJnHtHhy%(CJPJaJnHtHhwCJPJaJnHtH h=:hy%(CJPJaJnHtHh=:hy%(CJaJh CJaJh=:hyw4CJaJh >CJPJaJnHtH h=:h CJPJaJnHtHh CJPJaJnHtH#44kdX$$Iflֈr7  n t0:!44 lap<ytyw44455Q5U5`5d5 $Ifgd > $Ifgd1v $Ifgdw $Ifgd1d5e5kdY$$Iflֈr7  n t0:!44 lap<ytyw4e5g555555 $Ifgd1v $Ifgdw $Ifgd1p5s5w5x5{555555555555555555566;6<6ŷន{laP hy%(hy%(CJPJaJnHtHh1hy%(B*phh1h#B*CJaJphh1h#B*phh1hA5B*phh=:hy%(CJaJh=:hyw4CJaJh_CJPJaJnHtHh1vCJPJaJnHtHh >CJPJaJnHtHhy%(CJPJaJnHtH h=:hy%(CJPJaJnHtHhwCJPJaJnHtH55kdZ$$Iflֈr7  n t0:!44 lap<ytyw455<6s68999:::::::::: ; ;.;/;$a$gdyw4gd(x"gd(x & Fgd(x^gd_| & FgdA & Fgdy%(gdy%(<6s6G8d888889999999999999: ::::+:m:q:::::::::±{s{s{s{s{s{s{s{hd\d\jhgMUhgMhlhyw4mH sH hyw4CJaJhlhyw4CJaJhyw4jhyw40J$Uh1h(xB*CJaJphh3 K5B*CJaJph h1h(x5B*CJaJphh1h2|B*CJaJphhi!hi!B*CJaJphhi!h2|B*CJaJph h1h_|5B*CJaJph$:::::::::::;;;; ;;;;#;$;*;+;,;-;/;0;6;7;8;9;:;<;=;ʷtpʷtpf`fUf`pfh3 K0JmHnHu hyw40Jjhyw40JUhyw4*h3 KB*CJOJQJaJmHnHphu(jhyw4B*CJOJQJUaJph.jhyw4hyw4B*CJOJQJUaJph%hyw4hyw4B*CJOJQJaJphhyw4B*CJaJphhyw4hyw4B*CJaJphh1hyw4B*CJaJphjhgMUhgM!/;:;;;<;G;H;I;J;K;gd(x &`#$gd-_f=;C;D;E;F;G;I;J;K;h1h(xB*CJaJphhgMhyw4h3 K0JmHnHujhyw40JU hyw40J1 0:pyw4|. A!"#o$% DyK  yK &http://www.iode.org/index.php?option=com_content&view=article&id=462:iode-steering-group-for-the-iode-iquod-project&catid=10&Itemid=89yX;H,]ą'crec23.3DyK yK Jhttps://github.com/IQuODyX;H,]ą'cDyK yK http://iquod.github.io/uncertainty/uncertainty.htmlyX;H,]ą'c$$If!vh#v#v#v#vn#v#v:V l t0:!5555n55ap<ytyw4$$If!vh#v#v#v#vn#v#v:V l t0:!5555n55ap<ytyw4$$If!vh#v#v#v#vn#v#v:V l t0:!5555n55ap<ytyw4$$If!vh#v#v#v#vn#v#v:V l t0:!5555n55ap<ytyw4%666666666vvvvvvvvv666666>6666666666666666666666666666666666666666666666666hH6666666666666666666666666666666666666666666666666666666666666666666866666662 0@P`p2 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p(8HX`~8XV~ OJPJQJ_HmH nH sH tH L`L 7Normal CJPJ^J_HaJmH sH tH p@p  " Heading 1$ & F<@&+5CJ$KH OJPJQJ^JaJmH sH tH@ pdHeading 2,DM Heading 2$ &@&a$m$$B*CJKHPJ]mH ph6_sH tHp@p  " Heading 3$ & F*$@&m$%56CJPJ\^JaJmH sH tHr@r  " Heading 4$$ & F@&/B*CJOJPJQJ\]^JaJmH phOsH DA D Default Paragraph FontRi@R 0 Table Normal4 l4a (k ( 0No List D@D "W Balloon TextCJOJQJaJToT  "Heading 1 Char5CJ$KH OJQJaJmH sH o pd Heading 2 Char,DM Heading 2 Char+5B*CJOJPJQJ]aJmH ph6_sH T@"T mruColorful List - Accent 1  & FRo1R pdHeading 3 Char56OJPJQJ\mH sH doAd ^Heading 4 Char0CJOJPJQJ\]^JaJmH phOsH tH nSn d Table Grid7:V0D@bD (x0Header  !^JmHsHtH>oq> (x0 Header CharCJPJ^JaJD @D (x0Footer  !^JmHsHtH>o> (x0 Footer CharCJPJ^JaJ.)`. (x0 Page NumberB/B 9Capple-converted-space4U`4 9C0 Hyperlink >*phB'`B .G0Comment ReferenceCJaJ<@< .G0 Comment TextCJaJNoN .G0Comment Text CharPJ^JmH sH tH @j@@ !.G0Comment Subject 5\ZoZ  .G0Comment Subject Char5PJ\^JmH sH tH 6@"6 #& z0 Footnote Text"Lo1L "& z0Footnote Text CharCJPJ^JaJ@&`A@ & z0Footnote ReferenceH*PK!pO[Content_Types].xmlj0Eжr(΢]yl#!MB;.n̨̽\A1&ҫ QWKvUbOX#&1`RT9<l#$>r `С-;c=1g~'}xPiB$IO1Êk9IcLHY<;*v7'aE\h>=^,*8q;^*4?Wq{nԉogAߤ>8f2*<")QHxK |]Zz)ӁMSm@\&>!7;wP3[EBU`1OC5VD Xa?p S4[NS28;Y[꫙,T1|n;+/ʕj\\,E:! t4.T̡ e1 }; [z^pl@ok0e g@GGHPXNT,مde|*YdT\Y䀰+(T7$ow2缂#G֛ʥ?q NK-/M,WgxFV/FQⷶO&ecx\QLW@H!+{[|{!KAi `cm2iU|Y+ ި [[vxrNE3pmR =Y04,!&0+WC܃@oOS2'Sٮ05$ɤ]pm3Ft GɄ-!y"ӉV . `עv,O.%вKasSƭvMz`3{9+e@eՔLy7W_XtlPK! ѐ'theme/theme/_rels/themeManager.xml.relsM 0wooӺ&݈Э5 6?$Q ,.aic21h:qm@RN;d`o7gK(M&$R(.1r'JЊT8V"AȻHu}|$b{P8g/]QAsم(#L[PK-!pO[Content_Types].xmlPK-!֧6 -_rels/.relsPK-!kytheme/theme/themeManager.xmlPK-!!Z!theme/theme/theme1.xmlPK-! ѐ'( theme/theme/_rels/themeManager.xml.relsPK]# xK3,/K3t 1Vcppps4 5 ^X' (24p5<6:=;K; !#$%&()+.3679M O04444d5e555/;K;"'*,-/012458(D LPsRK3XXXXX%,.JQSV]_cjls!!!!8@0( dniw*dniw) B S  ?puBGsxCH{OTuz di af % O T !!A"G"##J$P$$$^%c%"&(&&&''**,,,,--j-o-{--....D0E0319111122222222222222 33333#3I3L3{}T[/ 3 &&-%-1112222222222 3#3I3L3:::::::::::8 ̩3Ɨf1ZhT?9<ZN^@|=*fSCȞ oDZh Iz_%KGsZx/Zd*zBnE`o2b{y?>bw~ ^`OJQJo( 8^8`OJQJo(^`OJQJ^Jo(o  p^ `OJQJo(  @ ^ `OJQJo( x^x`OJQJo(H^H`OJQJ^Jo(o ^`OJQJo( ^`OJQJo(P^`Po(.@@^@`o(.0^`0o(..``^``o(... ^`o( .... ^`o( ..... ^`o( ...... `^``o(....... 00^0`o(........h8^8`OJQJo(hHh^`OJQJo(hHoh ^ `OJQJo(hHh ^ `OJQJo(hHhx^x`OJQJo(hHohH^H`OJQJo(hHh^`OJQJo(hHh^`OJQJo(hHoh^`OJQJo(hHh ^`hH.h ^`hH.h pL^p`LhH.h @ ^@ `hH.h ^`hH.h L^`LhH.h ^`hH.h ^`hH.h PL^P`LhH.h^`OJQJo(hHh^`OJQJ^Jo(hHohp^p`OJQJo(hHh@ ^@ `OJQJo(hHh^`OJQJ^Jo(hHoh^`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohP^P`OJQJo(hHh^`o(. ^`hH. pL^p`LhH. @ ^@ `hH. ^`hH. L^`LhH. ^`hH. ^`hH. PL^P`LhH.P^`Po(.@^@`o(.0^`0o(..`^``o(... ^`o( .... ^`o( ..... ^`o( ...... `^``o(....... 0^0`o(........t\^t`\OJQJo(hHl\^`\OJQJo(hH\^`\OJQJo(hH` \^` `\OJQJo(hHl \^ `\OJQJo(hH \^ `\OJQJo(hHL\^L`\OJQJo(hHl\^`\OJQJo(hH\^`\OJQJo(hHP^`P^Jo(.@^@`^Jo(.0^`0^Jo(..`^``^Jo(... ^`^Jo( .... ^`^Jo( ..... ^`^Jo( ...... `^``^Jo(....... 0^0`^Jo(........h ^`hH.h ^`hH.h pL^p`LhH.h @ ^@ `hH.h ^`hH.h L^`LhH.h ^`hH.h ^`hH.h PL^P`LhH.8^8`o() \^`\hH() \^`\hH ` \^` `\hH.  \^ `\hH()  \^ `\hH L\^L`\hH. \^`\hH() \^`\hHP^`Po(.@^@`o(.0^`0o(.`^``o(... ^`o( .... ^`o( ..... ^`o( ...... `^``o(....... 0^0`o(........P^`P@^@`.0^`0..`^``... ^` .... ^` ..... ^` ...... `^``....... 0^0`........h^`OJQJo(hHh^`OJQJ^Jo(hHohp^p`OJQJo(hHh@ ^@ `OJQJo(hHh^`OJQJ^Jo(hHoh^`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohP^P`OJQJo(hHh ^`o(hH() ^`hH. pL^p`LhH. @ ^@ `hH. ^`hH. L^`LhH. ^`hH. ^`hH. PL^P`LhH.bP^b`Po(.^`o(..0^`0o(...^`o(... ^`o( .... 2^2`o( ..... ^`o( ...... R`^R``o(....... ^`o(........h^`OJQJo(hHh^`OJQJ^Jo(hHohp^p`OJQJo(hHh@ ^@ `OJQJo(hHh^`OJQJ^Jo(hHoh^`OJQJo(hHh^`OJQJo(hHh^`OJQJ^Jo(hHohP^P`OJQJo(hHP^`Po(.@^@`o(.0^`0o(..`^``o(... ^`o( .... ^`o( ..... ^`o( ...... `^``o(....... 0^0`o(........!bw~bw~%K%K%K%K%K%K%K%KSC%K%KsZNNN8 8 `o`o`o9<1 oDBnm~3=* Ix/Zd{y`                                           *                  !N                 Kd  P4P Z}\>!notw]I(_|9/pwUGy*Ti!$,$|'y%(vu*/+>/yw478L: >9&A D.GIcJ3 Kt(kb_29N8fgMTi4b(x1v}^3W#3Mt XlAsaz79Ch " 811@1111z9!K3@Unknown G*Ax Times New Roman5Symbol3 *Cx ArialC  PLucida GrandeG5  jMS Mincho-3 fg7@CambriaK=   jMS Gothic-3 0000? *Cx Courier New;WingdingsA$BCambria Math!qhrLgrLgiC*ZB+\2012 3QH$P'72! xx Peter PissierssensPeter PissierssensT              Oh+'0t   0 < HT\dl'Peter Pissierssens Normal.dotmPeter Pissierssens2Microsoft Macintosh Word@F#@^U@^UiC* ՜.+,D՜.+,4 hp  ' UNESCO/IOCZ1  TitleP 8@ _PID_HLINKS'Aq 4http://iquod.github.io/uncertainty/uncertainty.htmlV https://github.com/IQuOD[http://www.iquod.orgNhttp://www.iode.org/index.php?option=com_content&view=article&id=462:iode-steering-group-for-the-iode-iquod-project&catid=10&Itemid=89rec23.3Nhttp://www.iode.org/index.php?option=com_content&view=article&id=462:iode-steering-group-for-the-iode-iquod-project&catid=10&Itemid=89rec23.3  !"#$%&'()*+,-./0123456789:<=>?@ABDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnoprstuvwxz{|}~Root Entry FPUData ;1TableC;[WordDocument 3tSummaryInformation(qDocumentSummaryInformation8yMsoDataStoreUĸUC1J2CSGBQBOADTQ==2ԞUUItem  PropertiesUCompObj `  F Microsoft Word 97-2004 DocumentNB6WWord.Document.8