ࡱ> Y 7YbjbjWW "t==BRT]  $$$$8\ h4$::2.BV 9::::::$l;`=%:bbb%:*  * * * b9$$b9* x* S89,v6H$$ &9,NOAA 2007 Progress Report Virginia Institute of Marine Science 2007 September 17 David Forrest drf5n@maplepark.com 804-684-7900 (Much of this document is the SCOOP Consolidated Statement of Work, The Objectives and Approach Sections from the template should be covered by the SOW text. The VIMS 2007 deliverables in response to the SOW deliverables is in the Results subsections, with summaries in the Problems and Impact subsections.-- drf) In addition to its roles as co-team lead for the Quality Assurance/Quality Control and .Retrospective Data Generation tasks VIMS has provided extensive support for several of the SCOOP Consolidated Statement Of Work goals and tasks. In the Quality Assurance/Quality Control task, significant advances have been accomplished in automating the discovery of workflow errors and product defects. In the Retrospective Data Generation task, significant advances in design, implementation, and management of retrospective storm product data sets have been made, however, focus on delivering updated real-time data products for the current season has delayed some progress. VIMS has provided support across a wide range of infrastructure developing tasks, as well as SCOOP administrative support. Task Lead Items Quality Assurance/Quality Control - BIO, VIMS, UF, TAMU, UAH (Team Leads: VIMS & UF) The SCOOP infrastructure will put policies in place for ongoing oversight of product accuracy by coastal scientists. Policies will be implemented so that products on the portals are reliable and, if not, assure that we can intervene as appropriate. These policies will be automated to the extent practicable, but we anticipate some level of ongoing human oversight by coastal science experts. This task is expected to involve substantial input from the coastal scientists. Deliverables/Milestones Develop a policy for human oversight by the Coastal Science partners. At minimum, real-time and retrospective products should include a documented oversight/review process. Ideally, an oversight mechanism or procedure(s) would be included in the Catalog. (All Partners) Derive a set of automatic statistical tests that check forecasts against real-time observations and issue QA/QC flags. This too should be documented in the Catalog. (BIO, VIMS, UF) Ensure accuracy of visualizations or text messages (BIO, VIMS, UF) Provide model-data visualization for the purpose of displaying verification anomalies. This effort will result in a graphical (and statistical) model verification tool for SCOOP efforts. (TAMU, UAH, UF, BIO, VIMS) Add QA/QC capability to OOSTethys, and incorporate this capability into the SCOOP infrastructure to the extent practicable. At minimum, document the relationship between the two. (TAMU, UAH) Results The development of reliable human or automated oversight and review of model results requires fine-grain data extraction from the SCOOP infrastructure. VIMS has worked with SCOOP archives to develop fine grain, interoperable data extraction tools to support QC/QA operations. A large part of this effort is the development of Open Geospatial Consortium(OGC) Web Feature Service (WFS) extensions adapted to multivariate time series and forecast data. delivery. Collaboration with TAMU as the primary SCOOP data archiver has been essential for progress in this area. VIMS has developed low-level monitoring reports and provided RSS (Rich Site Summary) feeds of exceptions to enable automated exception monitoring of dataflow of real-time storm events, retrospective scenario events, and test events. VIMS developed prototype RSS-based event detection and verification monitoring systems in anticipation of point extraction from SCOOP. These RSS event detection and notification services are operating in experimental mode using the NOAA/NOS/COOPS tide observation data provided through web services. VIMS developed some model-data comparison visualization services for non-SCOOP data and is working to adapt them to SCOOP data. Problems Assessment of the accuracy of SCOOP product visualizations and model-data comparison visualizations has been dependent on the development of mechanisms for extracting fine-grain data from the system and matching it with the visualization products delivered by the SCOOP infrastructure. The development towards this goal is ongoing. Impact The recent progress in adding point and time-series extraction services into our data archives will enable automated QA/QC procedures to ensure data quality and reliability. A significant side benefit is that this approach will enable nested modeling approaches through the use of service-based data extraction. I.e, by putting model data into SCOOP and making it accessible through OGC WFS, other scientists can access and use the data independently of the data format. The development of prototype filtered RSS feeds provides for tailored high-quality notification. In addition to real-time monitoring of the quality of SCOOP models, this mechanism can enable notification of tailored event conditions from direct model output. Retrospective Data Generation - UF, TAMU, VIMS, UAH, LSU, UNC, GoMOOS, BIO (Lead: UF & VIMS) AllPriority Productsthat leverage the retrospective scenario, require that thedata exist in the SCOOP Archive and areaccessible through the Catalog via the prototype end-user portal, OpenIOOS.org.Much of this data does not currently exist in the Archive and must be generated bysourcing various data for those retrospective time periods or running models to generate the retrospective data. Deliverables/Milestones Identify specific applications and users outside the program to test the infrastructure. Program partners have already made some of these contacts. (All Partners) Source or generate retrospective model-run data and observations for major storms as well as non-storm related real-time observations from 2003 to present. (TAMU, UF, VIMS, UAH, LSU, UNC, GoMOOS, BIO) Create a Retrospective Portal job submission & workflow interface based on available resources (All Partners) Results SCOOP has delivered historical model data to users outside the SCOOP program through the SCOOP catalog interface.The development of reliable human or automated oversight and review of model results requires fine-grain data extraction from the SCOOP infrastructure. VIMS has worked with SCOOP archives to develop fine grain, interoperable data extraction tools to support QC/QA operations. A large part of this effort is the development of Open Geospatial Consortium(OGC) Web Feature Service (WFS) extensions adapted to multivariate time series and forecast data. delivery. Collaboration with TAMU as the primary SCOOP data archiver has been essential for progress in this area. UF has delivered ensemble wind fields for several historical storms within the SCOOP infrastructure. Pre-season experiments have shown that SCOOP modelers have been able to make use of these wind fields to produce retrospective model runs of surge and waves. Data availability for historical events have limited our ability to duplicate the current configuration of SCOOP products and services using historical data. SCOOP has developed a low-level mechanism for retrospective job submission, SCOOP (UF & VIMS)has designed and developed of an empirically-derived ensemble of hurricane track forecasts and associated windfields for use in forecast surge and wave models. Problems Historical data availability issues have limited the ability to duplicate the current workflow in a retrospective format. For example, initial condition hotstart files for surge models based on historical NAM wind forecasts are not available. Model development to adapt the models to use alternate data sources seems like the only option. Development of a portal-based workflow interface to provide retrospective runs has been dependent on resolving these modeling issues. Impact Retrospective runs enable the estimation and assessment of error on our current products. Developing the infrastructure to build retrospective runs has produced benefits in the design of the SCOOP architecture and monitoring infrastructure. By developing tools to determine if retrospective runs are complete, we can more easily determine if the event-driven system is operating reliably. Non-lead items OOSTethys and OGC Interoperability Experiment- TAMU, UAH, GoMOOS, VIMS (Team Lead: TAMU) The OOSTethys project develops open-source software and reference implementations that leverage interoperability standards developed by the OGC. This capability supports collections of observations for use in the 24/7 and retrospective use case scenarios. The OOSTethys project is a crosscutting initiative tying efforts of SCOOP's observation reporting and web-services implementations with the Marine Metadata Initiative's (MMI) efforts in the semantic web space. This effort also ties into the creation of a SCOOP-related Open Geospatial Consortium (OGC) Interoperability Experiment (IE), which will evaluate SCOOP's use of OGC open specifications including the Sensor Web Enablement and Web Feature Service for data dissemination. OOSTethys will require ongoing representation of SCOOP in OGC working groups and meetings, as well as data within the Archive system to provide guidance, substance and data for demonstrations carried out in the OOSTethys work, and the Interoperability Experiment. OOSTethys will result in downloadable software tools and a set of cookbooks describing how to implement these specifications in a stepwise fashion for programmers with a rudimentary understanding of web services. A similar cookbook will result from the IE, along with recommendations to OGC for changes to accepted specifications based on IE findings. The OOSTethys work will be presented at the next OOSTech meeting (date TBD). Deliverables/Milestones OGC Interoperability Experiment (TAMU, UAH, GoMOOS, VIMS) Foster integration with SCOOP activities (UAH, TAMU, GoMOOS, VIMS) Refocus the data and metadata standards efforts to promote the work envisioned in the MMI in support of SCOOP goals and objectives (TAMU) VIMS Delivered Milestones As an institutional Ocean Observation System (OOS) data provider, VIMS has implemented OOSTethys SOAP servers for sharing VIMS OOS data, and is upgrading to participate in the OGC-IE/OOSTethys using the newer servers. VIMS has developed controlled vocabularies for observational use with the OGC/OOSTethys Architecture Components Apart from the final item in this section, component-specific tasks describe system enhancements. The final item addresses routine maintenance and crosscutting needs. Visualization and OpenIOOS Development - TAMU, GoMOOS, UAH, LSU, VIMS, UF, UNC (Team Lead: GoMOOS) The SCOOP program will continue its emphasis on Web-based mapping standards and Geographic Information Systems (GIS) that facilitate display of high resolution digital elevation models (DEMs), land use/land cover databases, and other base layers that are needed to support the Priority Products. Currently various visualization tools and interfaces appear on OpenIOOS and directly on the SCOOP Archive. The intention is to migrate the Archive interface to the SCOOP portal (scoop.sura.org), and to maintain OpenIOOS.org as a prototype web site for a Regional Association, as well as a front-end for visualizing applications of OOSTethys. Both UIs will be enhanced and consolidated to provide a consistent, highly usable end-user portal for visualizing the SCOOP model and observation data for all three user scenarios. In particular, all of the Priority Products require visualization enhancements to OpenIOOS including visualization of inundation and flooding of land as well as model-model and model-data comparison for retrospective data. In addition, data that is stored in the Archive needs to be discoverable through the end-user portal of OpenIOOS using the Catalog to search, browse, and view archived data. Deliverables/Milestones Develop requirements & storyboards for all end-user priority products (VIMS, UF, BIO, UNC GoMOOS, TAMU) Integrate searching, browsing, and downloading of retrospective model-run results and observations into OpenIOOS (TAMU, LSU, UAH, GoMOOS) Develop visualizations required for Priority Product #1 showing high-resolution inundation using web-based GIS and showing overlaid streets. (TAMU, GoMOOS, VIMS, UF, UNC) Integrate visualization of forecast ensemble-products into OpenIOOS (TAMU, GoMOOS) Enhance OpenIOOS model-data comparison to support retrospective visualization for the Priority Products needing retrospective model-data comparisons (GoMOOS) Create model-model comparison capability on OpenIOOS (GoMOOS, TAMU, VIMS, BIO, UF) Develop a PDA Interface in a simple text format in support of the24/7 user-scenario (TAMU, UAH, GoMOOS) VIMS Delivered Milestones VIMS 24/7 surge forecasting product for the Chesapeake Bay is now available through the OpenIOOS interface. VIMS developed prototype model-model and model-observation graphics and reports for designing the OpenIOOS model-data comparisons. VIMS developed a high-resolution inundation model for the Upper Potomac river and DC area and delivered the results to SCOOP.  Figure XXX: Upper Potomac Inundation model visualized using the SCOOP infrastructure. Note that this graphic shows model results on land in the Upper Potomac river up to DC. The land cells approach 20m in size. VIMS developed prototype model-model and model-observation graphics and reports for designing the OpenIOOS model-data comparisons. VIMS is also developing inundation models integrating SCOOP winds and regional NWS forecast windfields, to produce other inundation models for SCOOP. VIMS is working through the open source project Oceanmodelfiles to help extend the COARDS/CF NetCDF convention to cover unstructured grid formats to enable more interoperable point extraction, visualization, and sharing of unstructured grid model results within and outside of SCOOP. VIMS developed Googleearth overlays using SCOOP Web Mapping Service (WMS) to provide for rapid prototyping and delivery of SCOOP model products VIMS demonstrated high-resolution inundation model results delivery using both GIS an Googleearth. XXX drf: Figure: Alexandria inundation . Data Transport, Translation, and Point Extraction - UAH, TAMU, GoMOOS (Team Lead: UAH) Data transport protocols allow SCOOP infrastructure components to retrieve and/or input data into the distributed archive using SCOOP partner storage and network resources.To this end, data from numerical models need to be retrofitted with standardized interfaces for input and output data streams, or wrapped with data-translation capabilities that serve the same purpose.The architecture can pass data between models, that is, the output of one model may become the input to another. To make this work with minimal impact on the model code, data-translation and filter services operate on data as they are transported between the modeling applications. These service interfaces include format conversion, region and parameter subsetting, point extraction and re-gridding. The service interfaces (as opposed to the components themselves) are often customized for specific products or outcomes. For example, data extracted from model runs are compared to observational data in real time. This enables a model-data comparison capability to facilitate ongoing quality assurance for model predictions. These location-specific point extractions are also used to generate time-series plots comparing model output with corresponding observation data. This capability exists on OpenIOOS.org for the 24/7/365 scenario. The real-time capabilities leverage Local Data Manager (LDM), which is a key infrastructure technology for data transport in SCOOP. Deliverables/Milestones Develop aconsensus strategy and plan for consolidating LDM operations and configurations, withthe goal of improving LDM performance and reliability (TAMU). Assure that data-transport capabilities and system functions (e.g., triggering) do not become overly technology dependent. Investigate the relation/addition of existing technologies for data transport including wider use of Thematic Realtime Environmental Distributed Data Services (THREDDS), and the NOAA National Operational Model Archive and Distribution System (NOMADS) system, as well as OGC specifications such as WFS, SensorML, and Common Observation (TAMU) Extend data extraction capability to include retrospective data for new models and models currently integrated into the SCOOP system (UAH, GoMOOS, TAMU) Generalize translation and extraction services to provide a standard mechanism for new data producers or modelers to use them (UAH) VIMS Delivered Milestones Since point extraction is an essential part of .QA/QC, VIMS developed a use-case for general archive-based point extraction and worked closely with TAMU and UAH to design an OGC WFS VIMS is working through the open source project Oceanmodelfiles to help extend the COARDS/CF NetCDF convention to cover unstructured grid formats to enable more interoperable point extraction, visualization, and sharing of unstructured grid model results within and outside of SCOOP. Cross-Cutting Components, Operations, and System Maintenance - All SCOOP Partners Several components of the SCOOP SOA support multiple layers of the architecture. These include infrastructure resources, security, web services, and data management. Infrastructure resources include the collection of physical resources for computation, networking, and storage. Any of the user scenarios or Priority Products can introduce enormous demands on infrastructure resources.While the SCOOP architecture is currently deployed over several SURA institutions, the intention is to leverage established grid activities for additional resources, such as SURAgrid, the Open Science Grid and TeraGrid. Security requirements increase substantially with shared physical resources for storage or computing, especially when the resources are distributed across multiple institutions.To facilitate resources sharing, most SCOOP services rely on web services that comply with various standards bodies, including the World Wide Web Consortium (W3C), the Open Geospatial Consortium (OGC), and the Grid computing community. These services enable machine-to-machine interactions. Web services become especially powerful when matched with earth-science content relevant to coastal applications, such as those developed by the Open Geospatial Consortium (OGC).With OOSTethys and the OGC Interoperability Experiment, we expect web services to become the norm. Over time, we expect to see the number of products offered via web services grow significantly and encompassing virtually all accessible data and translation products. Data management encompasses a range of services needed to coordinate observations and model input and output. These data will reside in a variety of places, including the SCOOP archives, the OpenIOOS database cache, and the myriad data providers external to the SCOOP program. Data management relies heavily on the SCOOP Catalog and associated directory services. The data-management component coordinates with resource management to assure that data make it to their intended destinations as efficiently as possible. Even for a prototype, the coordinated system components require regular maintenance to accommodate the proposed upgradesand tokeep the system operating smoothly.This maintenance includes things like updating the SCOOP website, documentation on the wiki and communications (con calls, meetings) that keep the implementation teams on track. The SCOOP public website at  HYPERLINK "http://scoop.sura.org/" http://scoop.sura.org hosts the SCOOP Portals that currently exist and are being developed. Deliverables/Milestones Adopt community standards within SCOOP architecture (All Partners) Implement compatible authentication processes across multiple institutions (All Partners) Develop security policy for archives, catalog and other SCOOP infrastructure activities (All Partners) Document resource requirements, (e.g., disk, throughput, compute resources, DBAs, reliability, etc.) (All Partners) Implement centralized source code and document repository (All Partners) Increase storage capacity of archives (TAMU & LSU) Maintain SCOOP public website (scoop.sura.org) (UAH) VIMS Delivered Milestones VIMS is working with the OceanModelFiles project to develop a COARDS/CF NetCDF convention for sharing unstructured grid model data VIMS is working with OOSTethys/ OGC-IE to standardize its delivery of observational data. VIMS implemented a source code and document repository for management of production code and shared documents. VIMS also shares SCOOP configuration files with the TAMU repository. VIMS maintains the SCOOP private website (scoopwiki.sura.org). VIMS coordinates the SCOOP technical communications.  THREDDS is a Unidata technology and middleware project to bridge the gap between data providers and data users. The goal is to simplify the discovery and use of scientific data and to allow scientific publications and educational materials to reference scientific data. NOMADS is a Web-services based pilot project that leverages community and in-house software development efforts across multiple institutions to provide and promote distributed format neutral access and model inter-comparisons to Numerical Weather Prediction (NWP) model input and output, climate models, reanalysis, reference quality in-situ datasets and now under development satellite datasets.     2007-09-16 --VIMS/drf //: NOAA_progress_2007VIMS.doc Page  PAGE 7 77CCQQRQvQwQQQBVCVDVXXXXXXXXX)Y*Y0Y1Y2Y3Y7Y0JmH0J j0JUCJ j0JU jU ;   . < 6  " & F & F & F$ ;   . < 6  "9Myuqke]U  X                :  C                 !9Ml \ !!""U#$$$****+++ & F & F & F & F l \ !!""U#$$$****+++,,, ----.222_33û}zwqlifa[SV     >?    x                o     +,,, ----.222_334455A6[66K7778$99 & F & F34455A6[66K7778$99:f;;;)<L<AB(CDEEEmFº}xph`XSK=  W  u       O          j    Z  t  /         9:f;;;)<L<AB(CDEEEmFGGGGOOQQQ5RRRjS & F  & F mFGGGGOOQQQ5RRRjSSST6TTUUVc}FyF3~zޠ 1~l{lCOBg)}GuS9=<3HÕV%Qz[g5ӮG<4>0#a[ts>ы3>w~Ǩ`64;7-jb-M?#"I$iߏMQꎞ8s.;/jI4Z/l?Rz 4 %d)k|pcC_ fSC 4ov1xH/ug=wWgc}ԝs\78ݽ3~_D$ dht*42{9ѪiԌ0~q G&r2$Nm#M#:6y5M6O-ikfK/z< Ai۰PT,w9a۳ [Xb{VJ)1D[8>ntC@=9SMvJ /V3(|ϵh18ݖ2tXqs3nhg6̘g'Ozv9 /kRyn/mk]ZP*cOU ̙*,sKe1zަZyBUXu?GYߡ FK"̷a>VOr/}9YTy٨wӴI>㇆$KpѦJ8* \zyrL}<5.Ԃ.7*1JA6&|8>WVR9 }G2&g+h= )'6\BeT/lAOhbLޞn@|0HJ8o@/VKBYe(+, َeM2Ɋ^Se:]e'uk\h@Hi/CN-uK Ԉnd%b@CIzkLg c4RY2CU~~Q. m_u :r/C4:rƧ:]|46sP,^[.O >1ޘ L, V.zE(Tn̥V?L'j ;6muӂ'JQd+5IEAyZ3~wUcKI+\g˥ C^csp:x֟-eB<1 ׼_c<,D OE#cW U~Q|qmRwk!$!˦J 4 ѩ e?$HY+P!? F;^XX)ejJVzW[[Ȫ)n&xҐ+31yZfØPy#Dt:ǞkW.PTh ]Hݙn>v.6@|( Yz̅uimbjFir4m0: %r;Ad"|X[tpmF㷐: {j,js't{1ۨ 4\FcH7ox]QaHP.L7 >&kudr#b4>Ad˼v1ipծ+9I^?DQؙ!`IʼnfXl/M.OxG  R[!pW죋tBdBnDJ2=8.iQ˙ԝzRC^9-"p/*dKFp/O\8S: ETl2wt" &@/e{ߍ6\t䢿;s^՞׽G鎞#4\[:!Ug'd; (6<J{ٟtH3S.\f<oȑjx lu=_֨l F#`>`mZi`zmdJ=$BNQ3Z$pH\P8 'vlX|ş2v8.MNrtlCSRXx8w[NCض9# #1o,QWg'MSUlhYkϕ4+j.k﯀}`Z^s3nTcH3>RcaYRk\Lz(~QkHĽFEH Щ:r w>yNZ0(b)NʅZ;`DΦN/Ȇ?*N<͈۸<ﳈI"+;vLLDq1XS[[4cvQN@"${ 5 Wrlj¦>MC$ D]h]{Aw뿂-dT`紨_kT$յ-L$D*sSBW"pDA:+Xr !֌ tN2X)/\TV0J'.F[r:1QQꮬ2#1t7afrŨ[4# r݇. +Cvx~Fļ BH '7t)HY@ QXĿAW^426PV}_&'#&,ksf ah0+~Yl)9Z?ˏc1K#Yb5s*sxy7|axX⦄A#TC-AͶ5ާnE!x;AY.b2>|F&DKY4m 5|(4rW N?vJ\-qqƙƑ{n[N3M%i pD:`F~z A]R2EJAGʼnr#$鎊K_Ojc ;*OxG@};+]-)/Q<گ&<u.ۦܪ6icgZ76j'%G/[}t鋶] H r F}d\^jVOxWRt$dIs?>oIRd+$NCr!1r;Y:~;Is㖑t0Y?^?z7~6Nbኩ1*gR[C96WК%cdŨ]x)68 햌˶HALQ mtEvCƑL? J[D@eU'EJ导)} =$Qڽ*4H$pct!ctX@<*ԉAYExׄcI81S#%bRxQEC\keKG"~su̡OkW\?ytvJyn0Go8c7 +# {g r'^://W=rC7AbHZOmHߟATLf}$hzXn&XZNz yMMl.йK?"9h^6TI`$5oJ [}FXE;$:NmNȖ<Kъu7Gk)wj?xo<JDmy{co>oDϹuߝ^`lx&2, ZJD_OxW/c":R :¹a;;ߎ+%rye~XXخvxϺNnK.>bwkǙrTF+uTpBKθ0: ?p䷆XMdcrA!⸗AB )?r%w"^ Rp 後|/?.J]gjΛrٵ:tcWi/GuG]:SÓOCrA@At;E1$(9Xt*N I.Q?zK?= ~kwȁA]z+AU @J~$8V{ynWP_gA_z{TǯS:"I#7 ڔ'6nXR&{431.L{LJ99+7~% :*:Rt58d+6z_/8?cМ&J(s&*(E“GBeu3AG"rɷ:ԁD+?ӈRX~yCBų"y18R76{S>3ihfˆ HC(]S+Ax%*q 76:a`,m쎇&JS<ECrG!&NWXeAOTPDxr}U]ح2¨|/1/V*:RIwl+k \mþn[@ѐܫODFمd}QC:wٺ3:gܟq@Q'x!A-vFU8<Ʈ8DC#iь-"$ސ.n6zՈnc3¿EPϣ ՉjWNaN$Dr;hhZ\'i <'] c"4vmSI#I*N Ew TjC F%0]l8b@n>&fF%@uL. ɆK?Km-i%`1>BsH> G4io =xzbIPJ9]jD[˨!ճ@aZ'q˕ɘjTFWxH³le-X9c]?$F'YǺqK}#~J[)`N: 'VkJGYS>bb[wƊ̴[Jxj\6m{nӇ1|QibDdȢE|2M0\)BO%/_!~qJjHpR-~a+HMpD!t(ii{~L\QxQ &4*1N8㣔ґ\bPRei!$ܰQNM[E)dШJ/%I (h!$Iߓe]%'ޖԝӾIIw 8Rucd͸#iT""Y9OӨN.c+}T{A1'-Cw8ͨMZGYNxyq{lxrޗĵ/B\x.S2;I|b`Thh+5r.CGEfȪԩv{rr?8W0VK) zDǘ}-+eKnS^!g*m (K-ȹ1~w owjTQAEà؂pVd ۈFۮh&$J(r+ٰAE3>b炩gSs̆?eIWCٴbE2%=(U۟~gW"; A:Xv J9ڄD͠,im jJ)"bQ.~9X?2]߶{ whv  .}`3*d:OuDˎ Bـ$^¨x+<۾~Ҵ o+'(ehs%LTbK@"z:ƔF ,EapՒ㯃`x-!o^^7^ss+|~mj[O5N/q~M:csoa2YȕŢߖWb*SUJUM ,/^Č26}AsyMđ8H5 ލZLO֐K XG|Ə3:hW{&M hsSaGd/iDuk{*GGIF_k10adt{'Xh N9b)_OJ)ԁalַgJ2Ć^tBzm^5o٬gpQ,Ox')A#>E'˸ϖCH>R'nP6-+o0oٙ$L2)hG'ƻoWaZy3X8[͡T?Q!#y˖Ze!-7v<8^#JLMHi@ȸ ,4`l_J#d"|ş ZIBDa1[ H Dv3Z*@ K+cJ;譟-$ (Lj!x!Z ,udWILκGI0>ݣ,(egU3PJM ݯ,rRJ@h(zT"QՓ0Rj~f7o8G!Bp=#jS@?E-R+s 40UYG|Z+`65ֲl+[${*i4_8Ea*tM1GDuV$B:G+f Po#)][EQũC!cc3arށRlbBG6t&S݅O8JMGUh$ 1=ślxd}/5 Chݸc_'Ɋ cV`f-s `[_ LhYbw:?D>5bJP8tBR`X 85C䑂B?ɂyr>A{'Ũ=ʖ\6Lp}~"ŀ_SF/ JQE˜+Ju] [BNZVxD̓: D[ߢ*R9p8$@~6JL0#mƍj$VRw̜#~j_nܖz˓N5k'!B[a,L9N{`p XN|<6g>wSX8Uf&Q&Tx|dVL82[* Qџ3Qb5 vNT9ss3>ҋӭL/$:=`e㝳7W#n-CXƦK!-F1o(4C VU\nCFa!8Qȉ 3Pa?.DꌾW再+],i[al3%AC)C>]I҃*[!KS"pwRQ/9MfT9 ld0 `y˪'{1Q|P<ѿםV g$Ismrę~Ń .)Rʢ'NbK/Sϊ*@¶qK֎ #![Rxg1⯦쉙f["/,'贵i-OĸRwJ:8G+lɎ3O[J#x ( V|>)`R %S¼rBtȨw~0HgE%wXT6 ¤tebKF/#ْOU e`6aٜM'|C'm4bJ[jYC"T O0ҮvP?>x3>71_ +`BQJYYD!,B&L;c=!OfQV ܶa(/OP(@6e˜HA6#“F/62y8C\w8O}}ؔצL6e+)yeYD>%Xd϶PՆyZ܎`YOJtD]xz$`aQ_$$$U@ME,)W~,'HVNx'}P6̳bIYVca r+g`IRe`,OŒhیxoI*-*c~T)NC,|i T f2S,*LEx|x_ʹu@mw!ZV2DBp\|*v*ڔʊ.w /* 3DDblW2Q9x㫵M3ksɱT`vT lqÔ'B0a6Ƒ%S%!؂n{p[M [L"ʱrTYu&x9Ŗj'UJ4LojU_HLd_$q0?] 9Ő'*I@p_;8d냦>RU#dm}ƵH`TDoYΩ/F+N6-L ̝n&y-‰L>8j>uS޸&i*QPQ:dsD?ZWeCfB[dDUUFLcCLOOvoqކ- }ʸ+tޘd 6Ӿ_fz=05DT[܆烝GXBPwG˸mouBd[eoP![%32Oa1rvSYY/blL KB7ReN,Oj)VY*X5Mn~EJk<''.}, EYk " v{ gQʖfc[T'+pc/ur6OP6 CxϯYuۡ%;ЊJfd^EsSU٦`+p0 *?x3gccG[z8U.5o۵ S%{:b,% <m0 !GBr@e#~c*W$'ys.cGUѝ}Pd[ 3Oؾ5s$,ZFO S5{^jW-6$v-2Qm6wXi Dh(QU{{fYEswɆTHՌ97y5%MYPUfMyA5~C ȸG=U<~lO`*s N7/&_䴑*1Α'pFg @lQpןv!AL!OX9p T/OP%H T92/ft1"LZ:l0xrFa:ʓrx{#b/=VҹO7>T؏161]o)["Ρq3T;i݊'pMeh u%ib~2SReA0Cl_?6!ls"KJl1u1U0u gwǜ/TYFzKRUe^.|0A 5kyBREf.ռ xWxO"TcFI=fޡdf?9m.XrY.Cc9u@qG[e[Hh'O|!jEUyTهZjڰ _N;[l2Fٮ[&_"={qKPeؼ zB'T oRXI赴Rvzm{.FYSc IDAT+Kٿk Qe_α*M.ٳ@_y8 1( %^f}*/Gu_JB!KYɥ¬͈*%XkUŸmDAERw_ Sӿl:?>2;,&lH]kF6[$"^v6r^YޓBv^8'ulb<2)kd3~3 Y)uw^'*^@Oذ^o BX;yVSg[G3CIKI !_*ڽŖIBh-C.o=DUW Y$>ӸBfWw'}#˸n:@b0QEk#U-&M#gA#M/;|seJD*P `Nh(=1co a@R4.zM-J%#`. ZAJOq,_3dp3 2۵kj7#[XO!U{\ է`ذOk}U2 (AJZ]޾2FRC#`.Ïy򄙒lƨm(0堊>םiOP:|<%dzC,08U3OdM,#)`do,qR$C \]˲xlƉL:ઔ]+A]$-aʈ:9Je0mW'cXN񷲗Q|=},O>9C~7L/!9 PY,na>b ם(R8,F1xgkmpȚ-- Tddg1&.8cn\Nԧ?eE!6 bgtǀTu*p!W8nbҠ£F_<%u*'063 ~"Ty2M.RE',@z f& XN3#Ps*FzKNh-U$OdNkc!ÖE,$"o#Olo7vTm ֯HW\Qx|79wcaG'6tWo8?N+v~.vO?(e\iJE/ :yNv==ˤ}Z} 0 =э'2;'샮Cr P쨎NeymѢ 扳Voz*)yV%c+B̉L\=+oL!M@lsI{L};"4] ~{i*1NJzajh4q=e$o pԥSiq#JfS7~_*y4FgT>#*N ?wvr}SRXI54\x:3:68CЊ@_J D:ɊFC F)NcF1N R19S?ԝ 8sJy@PIw\(D#mETB_ )U q,k3%7{h@[v8dRw%,s#n*l@K&]^ϒ!@nk0 #gm" PwP*l|îƽtFAF1UF1ۥF YH}b$dbh5TyfړH9CQZ,JRƆ3%Fb -zEw2+]aF6 q'u6d8O.þ=QdnMk0%xat 3 A#θ-qzƶx8쏐wHQY=Wzbk=QQux *ʼcGB[zGo$|iyAJyV{1h96sl*Qs,L UD{Ĥ@`9ӟ-;Z7D'wYR3aFݷAI4mµ5Yܾ5ϑ(p&khϚJWzJݥ:(u 37nOCa^ OW;LC[К4`2sve CvE}c#"CJL=>a]Fn">"}ikT mܠ>6T1ïs~)i$yU\7d񥔢.[ L@ t41c12 ^ 93'ˀHTGCl$VYR LԘ7a0 )`~^/ ;N0O!.[sje(m T/_~HlXQ?RAvUGXO/l7-RM *r'd3TVlAĠ #DhIVg-O6={ u6Bf1X;Q۪l$0=<~8Q?`OA#~H\amZd%f6)VgD]ɥ~y"Q4wB 9sG zU޾8Z`քIY6oCzeb̐>c+NB5U#hD*Rq j'rEG1bG/κSOSB;"U_d"ґ:zZk`lWON)rH~/-" 6DL@)SL]QCG:DlSRdǑګ[egq$ZHJ$'UTBmŽ-jdqKĺ 'ZO7?3;g$PNT0hIM! ,׫Nr2>^ܫO2)O흌̗')ԩrX E[:3-4l)?{pEyF#bji 3@T O1KɘqQ1Z$U$OX,'3gRwx*_^}Yĉ'ǮugfCټ1"˜ʡYoՉX}^*X)Q#z6(pQjݍ;oφ*)ZEef}2BP_H *9ALDv +uW2U/̖bLFmdu? b٠1q`1GٍrX&E,Hx(s"7rI GjLwHcabuʍ-(,U(HX1jU3mJ"ǣJm!>jqHuưR eQT  3?:xOl16bB:0oHLճxTaD[+p*p eXĉpff *i8Fe[^ C#^8 _ˎi7U @DUɢl!K]җCZ3eOF4+%#=eov$RyQJRMl|̋*"PS`D=Ӡ|m (6=UHܫO/¬ XW hk`|zkdKN$P0UYm`znU{>ڻ/lݲ޼Hio~Qj ҕITob?x7H)scF%9KJ:Pjho8-tu<)Io>bA<5o߲<[>h}9jkChF>QȚ={ϥ]E)U$Ɣ0˖D *턿xyvJ)8!@u^+#Ȯ|]^|4c4tg^GA[ +v9# /R,eZͭqK"yG}P6wo)3>MlC<:5'I]LȓԴ/U>#k'{p [J]?+?$7ܒqumYV ٲ@"[)t ʥPT9=|(>WAJ+aۭ<ԬwC;K_]tݺR{Fs gdb[,^RG<*wws#@QLo-ϸ X)*Pu*Xۖzyum`F KdYpmsƚ*Tl3rR' QG("a83iRa;־dr~>Q'N1y*&5JbkKbSR%S%!DnBmSWTieR$RؤǬf8U2O:`pa"Afzгks` 19OfuQ`'rU}7(UƠ<,UaT6Sel?M\PR 7R2U*JST VF (/ 1S*!JFd ޸+c1Xf(7L}6IeLkWB*C3ho$\xdi_dlc"*x@HHAddEnG&vӂe2U*짾fd*@M5d r|5!.բ|ƋX"DL8`;D~mle͕qt=`QDR Ko3 Y< 6ױM˒JL(uW%s]Ei/Q+3addD`}c^o?R/o b haJ [ E(qKMkڜ Ub#A7 c\? Ѡ'6I2mAqN)mEyQԚg%UlܠӁZU͋ 7TF Js HEQB8RW LbZw8GsZ:ؼd"HMAgY` |{uY2lɕIafR l1L4ot08e '"^>&,URA)]Fu9`2ILefIVJQPTQ lq^Rn sd-c~z"tˠ{}^ry)8[뼽Jy IDAT ̓y`UxR2S F{ 5ce+vӤ6uM W*QTya/KK:dZ,0?twCƈ*,idErPOUA^X,K5TpSTuF9wɆ"_f'A&/E,8ඨBK?ϱCE܆!X,Y$Ÿs)7w ʍn7c",H,㏝'tpyYxc#_|S=ޟoyj~#~}ReyQttÛ$0~-=xȾ'PfmK5Zڏ?o|Ϭc, J! wkܸƶ쫥ѳĂmz̎*u. aĜ<`; jh"Ӿ0;[ -:R95-PܩbwS8AT#-c4TI*~"$ݛ -IaRx=8Bm}TD#`VeReTiu[5܏Dȝ.(x`Tw]cXyWrƓD*YJ98[pR2?[%'A<1Rc{OV5FˤcmMe L̋DT[; ,BTv~KK ߧ3Syx`c"ʠԤ~[PTF.UUVV)'[L 78ɓճe#$xN` S%npMYj_i}tX61l/,e~ޥrgbXƿ>73G |e56,} J¢8n ~dtDV_axR]s#YԷ6tVNT:qˋT('vzRKfd/cE&;7}]!7yx>2퍍 !ju-%r!V3i^E3/Ձ=*/Mdn;㷎+} E4z=ɍuXb1E222,θQn Š~.yBC 32,z6f% Pd.2sYv^mYqg4-g.nɚYd.˕^HOH> EDo힑Fc7eSI/n;Yd1% `@rW?Ytgn}TȾTZkzMmwӼɡ!J)`쒦6)(,T2=5,Zxu_~yqugI~-Ȩy]8 HJx3 X1#3{ZݞS{#́ f}>qB1wIr?zWO냺//Ve^*-eCrt#g Gx;[ X)y^p|J|6l6=(,w?:$#bPc|Pث'ğIl8\yȁ4g:;3ڽ,T2Z#9iї+,KPAT?='aFxJ6OP"bI~JFlRb_LqY'8-[E_ Y5BhR,NRh  -TB}0ḥˑ=c*)v,T2?%H•]ф ~``89ϊ,8&D*7#_K )6 bzGGNpA[N Q@a-1S~C,ZOV[0JBv!vE?>/hqR`tcd+jHb0ךO-+/۔%,T2|0tF!T!=t5\ 8ji8Jᖩиc~ Yd$(..)MZ=yI Zr >FsBI\ gN'qqAorŐ^j6H') $>{Bepv#|]v:[5/ 'D": eX,87Dn؉kNK X˴$Z2Gsgt\w8ks+v7?B6bԈ Z>bD%-eɯb,x6klmB Ex&7G7V'\pe14mn{kh-KMb[rjvK4ŝ?A+4~,b=NOʨ-%1e29^(Jݕ݆)D + M駲 `Yz]F^:tVIiZ*3#BgOTtID89YYȟ%K`Tޒ腴Wg! ]4,QBP>q-FMgnQz-Hd 0Nll4 >Na݂g2(PIN{Qδ.:(=1"xDK\BdYHvSfL}/,Tfn;/UFۥҽ BHy&֥ͷYj+Ao !)u@)1Ji3 lyѹV*HdLB%AX\/TA "&k'7Y* jrey^4y!aW*`4){2J̀,:,R{ュt#S`¾F|/}#-ЏxrP'X=cR(#~ 4Pr#i\s;k6n9Čz}Ϻy™flN_W b/Db8˘('dJqtaW gaԝʲ dbKf2BsJqYK #~í-6[MvFyc nqe^Ҡ3Eh 2.K9\=6WtY(C)΋F2v2#aU[(Enbf_ihm"UcvIzwR~5N ƚ^w pBy9fBN%9' i ,n9v1~cMhֻ: Q"[D%-шd a1IM}f!Q, cL7hBℰ;zwe@ έeKL |)u;a٘r;\4e뚆ٲӬNrր<#B0fFqyqi u0ߣ~^4N 2H8eDB~lJ_ll"fg`1}V8[ +-|#ķWO#un'(uWur匂//(Ȏy=b4Θ#:4!Ӈr,0r=]CBČG&5mղ[F/Ώ0i*m%J]s4.É DؙpT3G޺VҥnQb%%^<'Tf7U$R[|C:[;̮b cyPƵX//%&Ge6XGi 7` $ ,R䴮%!b/czwP'2Yl1#Ro$dfp]cy+v|{*Pa˃,CطQJ)YĂL1fBKD[Ɛ,9cH ,ZH`&Iu}do-Ul WF8B)oզ8zi aqBPIMLA! t-WWefyq zs]/fM;I2<, gL{T.u,v|B&[xoR_ a'K?U2Ys*ٕ1!'TuA (p9? \)?dz/yC2ۼFxH9-ӒB  2 OM :hl;8KHF̨ܾm3|YcKF.aB^rA[5ד5ED"ƒ($;i=Ȍ~[¼CP * H+N2d-t0B/-!N,O Q0{.<%- ąaР~ycm%;CEKƠXPA\"C爁$J]21 VB`RgBKXPtK1J8(-99EfX߂PAN>"GZEC_(t}~0~Qy64VgYkytȷ1G,R!ߗ]9_KEv@!%u!z{S սsoXi1h<ڴ4b_mB+f-$J1鋿QFW)S.DyQ󒝑4,Dp=U;d4+F7RwPo f;z& -{8A7$'ңjlfQ*N갫O) 8S*u|HD>RX0-TQqniP<=Cl +J"8iE`4^GH]dצ8**22Y wFqKmg\D'pLG*' "rZp&;r-A,\s{z:I\xg+ݚ=E2L)[z5DGEoC{ʈ9Ҡhp}A}Hpʽz6ߨD$D 9:HS;&%xl &N-vɜW״|LWiVlYRlT-O%ԏEiK43dMI-%Oc9SX){Z#X1T8riCVoʳ[D`*P/ Uc?(z aJqc%wiG]Y >PWmGxkx)oB귦nAbd`:˂1fD}L'Kl8<*T'WR,_n§FaǤF ÔS%];jVHc"dY'R;IҲ^K9_{EirJqX*fc.4 N%b풴Qk$Q+SoK/"x Dz#sB4}\A7}?maJ"FSH[x  Zz6&EY=]v#d)~5> 47_%r "RFѓWp, ڿS4&uٳ''_ Er%Qz ~P x2T1 a)$ͅuMR?094G2Yo z=6V!^HX/]jh"$`(<0}6_XX!#A W˽x>:w=oI2(ПbLHTx7fLYѮO~uF\6O]K6A郕l~)ZE?ۺc DZb6KH%ֈ&+x0Fyn蟎%S5_*"+&z#E!@*1 kHůpdZ _ )/>IQ#Z)-ORI%TZj;X/E]VKnO.\-"Sc=','N:Z~FHxJTxKGtHy FOk^anP~滯S扟~hTrlLDc]=HS2R>T_@ +ϔHF6>3ZE> +3Z0=.e@_\:Lu /<%];BI{Ev|=y BEZ3e+'W鲓8GR ɩX0Eݮ4I"&1993=&5+^Gd5P^flqPZ""D=f 5c~vߕX,+Y-zW7 ,)2|OsBke^9jIK%/[@+]GdCR  :04CeNNhAͼۺl,6P+`(2'2\ ɯ81Yp;a݉$|[(~ 8GI X8F.D?맘W<3: ;|QBC`LEdLo9L]tK fi$B"劙R][b`ZSmŜ;. a eBeK)20k?14ӰQ :9Xm]T0 }m\Wڇ3 A- _d\yGW,9͇]s&/} ArjPDSg>̫RX2[d"_< 1diCZ0˱9T`CeRPTˢF{U PL %F74R /'R hS6/=+¢ӼZ"ΒHED{u];5JW\*/'RI1;Yu/ Pd= SaV,S[80qDiR@". u|caJ|3? ST4 .{R)C,H"%g_;S$]o3~Q/,=(HJz!=^sԽHLyugnos%NôW,AyxEcZ {ItnlI~cb vsI|>VzE$[U&V&-ubӘ7+9qIbwG-oғEG`IDATj7JإNN'꼢Y1K&ý- J%$J垾>) IV(#4[񽨥?gǤ;HK'Z(1efE̦†sjϺm>^'BtV:1\fNXFc`́Tz69#Ԃ" jOU!=ۯ֍GIopsN&l ҃臭o.Y(JR$u㞾n} 'Rr+jrf?,gΦ'Axv*fT}W";Q4 @*=t)p%^!W,F29g|~3=~z<% lP(bc(D3J''EFzBe8E:X%!MeFXxu4} ]01AU^aQB#"+滃;SD{( @ 3Hb" 5c*2x.o/k/!G'|F; o'Q>=}H(?4 A2e2J ^avab+'vrE-H$ |HE^$0w'2P΍e&(_ vw[>XjkLLzE~l^L?}} n"|>%AC:i%̧in :zVL5FA2LHedn22j13KβmhdN%DwQ.3ѝ"=(LDFbX-5o[tK%+cS/Ԣ2LKhmH}=R`J`H zĻAMyէV+v`p[%X+zs o m̔%&&XNf`(rr@*e`^lEb!^ν,IuuZiL+z옲H_Yl2 tR?' R ͞T_L[SrmsP>SW\5yPR)jW4>v+E|ܐq,s* i+i*,։8]N0~rB=?̛I~m}Ew'e.^EIEg \to47_He)b=M ۖXDO㫡wr\u+#s$@* 3S$c.u5 yj^mZe]Ȫ2y:O˂q@*}6L?GoϑAYoS0^1  jfvT6`P3f!<^&֤TdL}z%rF$0 lèeEWs᯸Z~b(TM*b/}31ͤf6Q\ʤ v֙ϐ!@مW&~ۋW ?h˙՗ugTWC=#+ ^BgJǂ4F>Rѭ!MZy%0'^1_T4^4 <; X= $L't}mihx6/t-U=ܟ3TawF!H%5#Z{K8}NNb:צWKCXϴ/!3G2'yQ w}ZED-1n]9Ku#+X(ąl}#D]iN仕cw/|*&%sTN~u@*S4Lw'1^Iщ<F Ӱw0/'}{DD_RtpshȎK,gkBR`s Ё"]Ys)k !B9K%-/]ir3HW&P8NKW[lǩxKBe8r}O %!D}^^ϴ%6leOEZtœB*AdEǔukT8{6LpWK=Zp3+~H5Z(C,CxVاLԑ{" W_aTXo+넁TvQBgޫn!);V`1g$ڋ)+{ .uy( h0u ȼL%ZLpG)+ѻ& )3~Fql0J W.[E71X=|Rb>7( /ΠRY㘽EcLyEFWȈ/f4fK^/~/fΣRY{%e^ۑ)W \_9L%%,a/&`(1(L+`@3+sZ0m}~Xdiˊ|yO\m7FޓAhb)̓]"@*xSK>6(Ny{۪~4׶LklI;u{'Xڶ%zj{}P)^ߝQj(=f]Í[D|}ЧI9Ŵ~O.ޒS'f(DG LYUK)WZH-nRKcK F@*[>>#]1 v)z1z04lĶnb`jY%!ehSvyɉHʝ"8>ч|CkE7EwO|%1u$Jq`͆sXi xÌ*^苲jAꭴ j)RWՒ+_Ż=YfD ߓ bZx0Jm;{ם:wz%6&HಲTBF S*r7+He>0Jͤyf=f3Gf2G3j6`Ͻ W Y+|> )f]1@*i62vaݮQ>f#zAw KoWX(! 61| \-N|7z 6ྍ#@?ο`\B M+Q43QRhh"*SW3&aIez-pV 2HeJ2-]刏HdǪ^횗ȜȃyK x$.1NV,EOD[^zO] 44`#~KLpb=]X*|P*M .Q@H'G*FR$îd)xTKMTM&%GZ)X~ȯL"N7:X L(fB3+= ^JYTm hABxEo;}<)'b Q(]*(,âJap8R[³I$!'‰ e0( KpC"C6J2Ti=?l˜{^yIP[XH:]ON2e,\㨝ugXR+nQzodĐWRBwOSQR)ˢdBzӑYqLsI wCM|=izݙO@+(` `;bH>҄y19҅F!6N޵'[pmCm?d1MhLw/;L`i RYnT1ɼy֨it;{C]fW`QJw$\r`aJ>0d -^г|9.kJŴTΩ$T@He2ʶ6ȮŻĚsղNWv:39uB@*(sT6'Fۥ_&[`fHY(TR!x%e&椤"n%(tIiG$ ['Fq/: 4 Tfm2 b+^ms{}J#FJ Q)&.X rfх+1` RْXR= /9V-crrRD*[2XoW4HmfRL=4)><vo[Ŗ( :aNhoWlr*v \s!c # \>.s;#T`p@hf j3meeHe z~VpXdEKjYe@*둿Kz!{|T@*9J Oz0`mnm47DB-ˁTʮA(F+ Ixe!HeUl[(e. R l䈭w`Fn][~Q_AR# Y(IsS7WNzmŋԯ (0AQ,z w1Cms2(֓j\\bI !Z^alRII/^4vCKC̓2;KRF++B-ew@*`3dL0u,ff:Z֩Q*"g %*ҡɺ`v W@GRvx)M!OUW6_Kg HM+0eh^!G.Ӫb/W6g 6`"u+c$N*JAp 6&3-_ɺ`ľ^V(J@oFu~^cx/3 R3':`4s|,{D* mݏHLTS) =LqX^^^c}ğ{VG3?HmV'3iOγbn)L}_4Yg~2 Ļ鋏zVgu_ GSG[ݡ>A3L^"v?FWc3⿯ |w,o/ݲ{7|ZPe͑ĸr޽{_c94s^<݉Ǘۘ|'"އu;6wU5ҍGגE&:x|),9/>N&ӏshl2xD N:|Ӿc/8 đi{c7xy++] ؀*Twh IENDB` [,@, Normal CJKHmH H@H Heading 1$<@&5CJKHOJQJF@F Heading 2$<@&56CJOJQJ@@@ Heading 3$<@& CJOJQJ<A@< Default Paragraph Font,U@, Hyperlink >*ph,@, Header  !, @, Footer  !&)@!& Page Number6@26  Footnote TextCJaJ8&@A8 Footnote ReferenceH*2Ob2 ListHeader 56>*6Or6ParagraphItalic6?7U7Ut  BBSSSV7Y.+9jSX7Y/135783mF$Y7Y02469UnknownwangQMvMM7UXGNPV! OLE_LINK3 OLE_LINK48UARAR8U$ - * 9<!U]?E*75= !!0"7""""#P#W#####$$$$9&B&f&m&&& ''/'7'((v(((())))******++,,,,--..P/V/////U0X0~00000000~11111192?222)3135 55555666677777748:8);3;M;U;??W@_@ AABBBBFF4F>a~>>7UP@G:Times New Roman5VSymbol3& :Arial?5 :Courier New;VWingdings"1hރ&8 C"8 ZK&"0dJSXBR (3) SCOOP V4  STATEMENT OF WORK Joanne Bintz David Forrest Oh+'0 ,8 T ` lx!(3) SCOOP V4 STATEMENT OF WORKros3)  Joanne Bintz Soanoan Normal.dotzDavid Forrest S7viMicrosoft Word 8.0E@PCYO@l@:!H C ՜.+,D՜.+,P  hp  SURAn"JSj !(3) SCOOP V4 STATEMENT OF WORK Title(RZ _PID_GUID _PID_HLINKSAN{0E60FFC6-9999-4B73-9E90-12D746DC183D}A`n]7Y:\CB_inun_vims.png  !"#$%&'()*+,-./0123456789:<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~Root Entry F,Gx6HData ;o1Table=WordDocument"tSummaryInformation(DocumentSummaryInformation8CompObjjObjectPoolx6Hx6H  FMicrosoft Word Document MSWordDocWord.Document.89q