Data Management and Analysis Subcommittee Meeting
April 6,7,8 1998
Minutes from the Fall 1997 annual meeting were approved, no changes.
Everyone agreed that using a special email list for the subcommittee would be a good way to conduct business between the meetings. Luther Smith is to discuss this with NADP coordinator Van Bowersox, to see if the Program Office (PO) can develop the list.
Luther Smith went through a list of several action items from the past, many of them related to the move of the Coordination Office to the Program Office at ISWS. Briefly:
The case of Georgia site 99 was discussed: Unusually high calcium chemistry was identified by data users (after having passed all NADP lab QC and QA procedures). The sample in question was found, upon further discussions and investigation between the CAL and site operator, to be the result of nearby field sprinkler irrigation water being captured by the Aerochem sampler. Site and data base corrections have been made, but this was a case where the one (and possibly a few other samples) significantly effected the annual summary statistics for Ca at the GA99 site for 1996. After discussion of this instance and the broader issue of whether or not changes were needed to NADP protocol, the subcommittee consensus was that in this instance, the system "worked" and that no change in CAL or PO data validation checks was warranted. Action Item: Van Bowersox suggested that he and the PO will perform routine checks of isopleth maps to look for anomalous data. Also, a notice of the GA99 experience will be communicated to network site operators.
>Clyde Sweet gave a brief overview of the Mercury Deposition Network (MDN). Then a breakout group consisting of Clyde Sweet, Bob Brunette, Eric Prestbo, Steve Vermette, Bob Larson and others, met to discuss and agree upon data and sample codes used by the Hg Analytical Lab (HAL) and the Program Office. There was some confusion in regards to the codes during this meeting. The HAL wrote an SOP for and applied the codes which were approved by the DMAS at the Ellicotville meeting, to the 1996 Data. The confusion stemmed from a few ancillary codes which the PO office applied later. After this discussion, it was determined that the Ellicotville approved codes that the HAL used were still applied to the 1996 MDN data, however, the PO office adapted some extra codes.
Invalcode vs Criteria Code:
In order to mitigate any potential confusion, the term "invalcode" should be changed to "criteria code". This is to ensure that the end user does not think that a sample is invalid because the sample received an "invalcode". It is possible for the sample to be perfectly valid however, still have any number of "invalcodes" associated with it. The "invlalcodes" basically describe sample conditions which could lead to an invalid sample. Thus "criteria codes" more accurately describes this type of sample code.
MDN Web Page Notice:
Two notices should be posted with the 1996 Database: (1) That Quality Rating Code "B" samples are valid samples (2) Notice stating that the 1996 data is more accurate than the transition data (this in regards to the data overlap of the MDN transition data and the 1996 MDN data).
The draft codes were then presented to the DMAS, with minor changes. These draft codes will next be compared with AIRMoN and NADP/NTN codes for internal consistency. Clyde Sweet will visit the HAL to finalize within the next few months.
Motion: Mary Ann Allan 2nd: Van
Clyde Sweet will (1) make a flow chart of condensed "end user codes" and show their relationship to the approved MDN Codes (from Ellicotville) (2) Research the NADP and AirMon codes and look for a common code structure that is shared between all 3 networks (3) Put a message on the 1996 data to warn the end user that the codes may be subject to change (4) Clyde will report to the DMAS committee by email with his resolution.
Motion: Van B. 2nd: Gary L.
Moved that the PO will develop standardized coding, to the extent possible, among the NTN and 2 subnetworks and come to the Fall meeting with the plan.
Point of Clarification: MDN MMHg Data
Bob B at Ellicotville proposed to put the MDN Methyl Hg data on the MDN web site which was not approved. Now that the MMHg data is paid through the PO office on contract, this data should now be made available to all end users as was decided by Van.
Motion: Mary Ann 2nd: Van
Bring this issue (MDN MMHg program) to the joint session for a discussion and vote.
NADP Web Site Configuration:
Issue 1: The web site map when showing the lower 48 states does not show (1) Alaska (2) Virgin Islands (3) Puerto Rico (4) American Somoa. Therefore, do site sponsors in these locations feel left out when their locations are not displayed on the map?
Bob G. felt it was feasible to shrink the 97 maps to include Alaska and the Virgin Islands.
Motion: Van 2nd: Gary L.
Every time the lower 48 is displayed on the web, Alaska and the Virgin Islands should also be included. This is all inclusive and applies to any NADP product whether a list, map, or data report.
Issue 2: Older Maps - How far back should NADP make available on Web?
(a) It was decided that they would keep all maps from 1984 to present available on the web.
(b) If users want maps further back than this, they can make a special request.
(c) All analytical data should also be provided to this date.
Issue 3: Output off the Web
Bob G. felt that end user want to be able to:
(a) The should be able to pull off the web high resolution maps
(b) They likely want to tailor the formatting themselves like changing images, colors etc.
Issue 4: Steve V. and Bob B. would like to explore the possibility of regional MDN Isopleths for Mercury deposition.
Discussion: Some felt that there were not enough sites in order to do this. Others felt that higher density regions (Florida, Wisconsin, Maine etc) would be good areas to try this. Could use "Kreeging" to do this rather than the other method.
Review of NADP Products:
Issue 1: What products (maps, annual reports etc) are available?
Suggestions-Get Gary to come up with a list of products that he thinks would be useful. It was decided that currently, any and all products are in complete disarray and that they need to be updated and systematically reviewed.
Mark Nilles volunteered to do a literature search-The committee came up with the following key words to begin with: NADP, Bowersox, Lynch, Vermette. Also suggested to check abstracts, titles, citation index and old list of Journal articles that CSU reviewed.
Issue 2: Review Process
How can the subcommittees continue to conduct this type of review outside of the meetings? It was suggested to use email to do this. Action Item: Bob G. will create a DMAS email address where the committee can continue to communicate and therefore make decision throughout the year.
Issue 3: Special Data Requests
Discussion: Can special data requests be accommodated and who will the contact person be at ISWS that can fulfill these requests? It was decided that Bob G. would be the best person suited for this.
Gary L. needs special access to results and data. Idea: Set up a special account or "port" for large end users/sponsors like Gary. This will enable them to gain access to data whenever they need it (CSU had a server with accounts).
Resolution: Bob G. will ask Van to get his opinion on this subject.
"Internal Nitpickers" (External QA Manager)
Discussion: Need someone out of the software-logic/loop to be an internal "nitpicker" to keep the QA unbiased and on track. The PO is in the process of hiring a QA manager, however, what should be done between now and when this QA manager is hired?
Bob B. explained that he feels that his assistant is currently almost filling this role.
Gary L. also brought up the need for a data management Standard Operating Procedure for all who fulfill this role to follow. This should be written and organized apart from the analytical Lab QA plant.
AIRMON status (Special sampler, data status, annual report)
Rick Artz described a new type of event sampler based on a Carousel design. The tipping bucket triggers the 1st sample onthe Carousel etc. The new collector would utilize a different preservative (Thimol vs. Chloroform). There was interest in an intercomparison between the old AIRMON sampler and this new design.
Ricks Issues with intercomparison:
AIRMON Data and Annual Report:
MDN 1997 Data Turn-Around-Time:
Goal for NADP and all subnetworks to standardize their data turn around times. The HAL will transmit the 1997 data to the PO by the end of May. The HALs goal is to be able to transmit final MDN data to the PO 3 months after the end of each quarter.
Trends On The Network:
Luther made a request for 4 people to join a "Trands on the Network" workshop. A few names were mentioned that were not present at the meeting; Lynch, Buttler, Lindberg, Prestbo. Luther will solicit these people to join.