	2004 LTER Information Managers Meeting
		July 28-31, 2004 in Portland, Oregon
	"Enabling the LTER Decade of <Synthesis/>"

Home > Site Bytes

LTER Site: Andrews LTER

Contributor: Don Henshaw and Theresa Valentine (Jul 22, 2004)

Site Byte:

Andrews site byte 2004

Andrews Database and EML status

Don Henshaw and Gody Spycher continue to lead efforts to migrate core and other legacy databases into our SQLServer metadata database. Currently 90 study databases (300 data tables) are online. Migration of remaining databases and the review of all study metadata will be top priority over the next year as the Andrews prepares for an NSF review in 2005. Mark Harmon (AND-LPI) has directed a $25K supplement to Andrews information management for the 5^th consecutive year.

The metadata database has been expanded and modified to accommodate EML. Once study data is in place, and thanks to the programming skills of Zhiqiang Yang (now the Konza data manager), we are able to serve EML dynamically. Yang has described the process we use at Andrews in the Konza site bite, which was originally patterned after CAP's Xanthoria. A Microsoft .NET Framework program extracts metadata for a given dataset from the metadata database into a customized XML template. An XSLT stylesheet then transforms the customized XML into an EML 2.0 compliant document. Our next task will be to bring the generated EML files into compliance with the EML Working Group's Best Practices document. Additionally, FGDC metadata records for spatial data stored as XML in ESRI's ArcCatalog are awaiting a stylesheet to convert to EML 2.0.

ClimDB/HydroDB and other DB's

Thanks to everyone for diligently and continuously updating and adding new data into ClimDB/HydroDB! Suzanne Remillard continues as the database administrator for the combined ClimDB/HydroDB effort, and with help from last year's NSF supplemental funds, Suzanne and Kyle Kotwica merged the project web pages. The public page is now located at http://www.fsl.orst.edu/climhy and the participant page for harvesting data and entering metadata (pw=climhy) is located at http://www.fsl.orst.edu/climhy/harvest/harvest.htm. These new links are/will be available from the LNO intranet page. A data access policy has also been established and available on the ClimDB/HydroDB public web page (comments welcome). Thanks to the continued efforts of Wade Sheldon (GCE), autonomous updates of USGS streamflow data occur on a weekly basis.

Theresa Valentine and John Pennington (graduate student) are working on a project (WatershedDB) to assemble spatial data for participating USFS experimental watersheds participating in ClimDB/HydroDB. The goal is to build an internet accessible application to allow researchers to view and interact with site spatial data layers and link to the data collected and stored in HydroDB. Additionally, plans to build a water chemistry database (ChemDB) as an extension to HydroDB is underway with HBR, SBC, and AND as willing members in developing a prototype. Seed funding for both projects has been provided by USFS Research.

Spatial Data

Theresa Valentine is moving forward with the current strategy to incorporate the HJ Andrews LTER spatial data into our general data catalog. Each GIS data layer has been given a study code and will soon be incorporated with our standard web catalog display and searching tools. Web clients will see a listing of spatial data, along with links to FGDC metadata (in html format), and a link to an ESRI export file stored on the ftp site. Ongoing discussion involves whether study site locations should be converted to GIS layers as most are stored as coordinate data in conventional databases. These databases can be used in ArcGIS without having to store them as coverages, shape files, or feature classes and may be mapped dynamically using tools available in ArcMap. As these databases are usually managed within the database management system, converting them into standard GIS information may cause redundancies which could lead to mismatched versions between the systems.

Project Management

Kari O'Connell, Andrews new site manager, has established a new system to track all the research projects on the Andrews. To clarify the process of setting up a new research project and to create a database of all ongoing research projects on the Andrews, a set of guidelines for conducting research on the Andrews, as well as a research proposal application form, are now available on our web site http://www.fsl.orst.edu/lter/about/research.cfm?topnav=159. This process will allow the capture of abstract-level metadata and GPS coordinate data of study sites at the outset of new research.

LTER Site: Arctic LTER

Contributor: James Laundre (Jul 27, 2004)

Site Byte:

The Arctic LTER was renewed this year for another 6 years. In preparing for the renewal we planned to redesign the ARC web site. However we decided to wait to incorporated EML metadata implementation. We were expecting a conversion tool from David Brinkman in the spring but the delivery date has slipped several times. It now looks like it will be available soon. In the mean while we have been reviewing our metadata for inconsistencies. The variable attribute section remains the most problematic and will require the most work to standardize. Although readable by humans the attribute section varied in the order of the columns and by the type of separators. I am in the process of using a perl scrip to help the clean up of the attributes. When we get the conversion tool we should be able to convert most of our structured ASCII metadata to EML. We also are looking for ways of making the generation of metadata easy for researchers. We feel the generation of metadata should be simple and easy enough that anyone can do it. It should not be considered a chore but rather a tool for research.

This summer record lighting strikes occurred across Alaska resulting in a large number of fires. Smoke from these fires has covered large areas (see http://rapidfire.sci.gsfc.nasa.gov/gallery/?search=alaska&date=) resulting in an increase in the nitrogen deposition measured at Toolik. In early June there was a small tundra fire 10 km east of the Toolik Field Station. Addition Helicopter time was received to initiate studies of this rare event. In mid July lighting struck the Toolik communication tower. This strike took out the fiber optic equipment leaving no Internet or IP phone connections. Replacement equipment was flown in and communications was restore after 5 days of being down.

LTER Site: Baltimore Ecosystem Study

Contributor: Jonathan M. Walsh (Jul 27, 2004)

Site Byte:

NSF Grant Renewal: Baltimore Ecosystem Study Funding Renewed - The National Science Foundation has indicated intent to fund the Baltimore Ecosystem Study for another term.

Online Databases: We continue to use online databases for stream chemistry, bird population surveys and input of metadata.

Online Metadata System: Our metadata repository and reporting system is being improved as follows:

1) We are moving the metadata records out of the "deep web" to make them reachable by url. When the records are only stored in a database, even if the database is online, one must use a middleware application to display or otherwise access them. With the addition of a routine that writes the records to a text file and places them on a web server, the records become reachable by url. One result of this is that they will now be visible to search engines such as google.com.
2) The system will now interface with the BES website directly so one can search the metadata records without redirection to the ORS web site.
3) Metadata stored in the ArcGIS XML format go directly into our metadata RDBMS (ORS) and vice versa.
4) EML 2.0 compliant records will be generated using a script. And CAP's Xanthoria will be used if possible.

Geodatabase: We are building a geodatabase using arc 9 SDE on Linux. The use of the access data file format as opposed to shape files should increase the synergies we have. For example, two people can do two different tasks on the same coverage - a technician could fix the topology on our parcel level database while another does analysis on the data.

EML: We will generate EML from our set of metadata and place it available for Metacat using a VB script. We will use CAP's Xanthoria is possible. We will generate human-readable metadata records using a transformation via an XSLT style sheet.

Other notes:

- We are teaching an advanced spatial analysis course at University of Vermont, and an environmental design course at Columbia. Both use advanced spatial statistical analysis and advanced data visualization techniques.
- We're doing a new flyover of the Gwynns falls watershed using multi spectral imagery - Quickbird or Emerge.
- The City of Baltimore has requested a good deal of our data for strategic planning for the department of recreation and parks.
- We intend to purchase a raid array and a better suite of tape backup software. We have purchased a DVD writer for backup purposes.
- Our publications are now in a SQL server database, generated from Reference Manager, and searchable using an Active Server Pages (ASP) script.
- We received a stipend to further our participation in Climdb and Hydrodb. We now regularly post data to the system.
- An ASP script was created to poll visitors to the BES website website clients requesting data and metadata. It records intended usage information as well as the ip address of the requestor.
- We participate in the Canopy Databank project, a component-oriented database project (http://scidb.evergreen.edu/databank) from Evergreen College in Washington.
- We participate Watershed 263, a Baltimore program involving an urban storm sewer watershed - Evaluation, Technology Transfer and Training - to collect evaluation and research data and develop methods and tools for decision making and management in order to improve the quality of urban land and water resources contributing to urban revitalization.

LTER Site: Central Arizona - Phoenix

Contributor: Corinna Gries and Peter McCartney (Jul 28, 2004)

Site Byte:

CAP successfully renewed its core funding for a second phase. Information management continues to be represented prominently in the over all site leadership design. Beginning with CAP II, Corinna Gries is now the nominal site Data Manager. McCartney continues as a PI of CAP and as Head of the CES Informatics Lab which provides the central home for CAP information management.

After restructuring of the management database and designing of new entry applications was completed the main push over the past year was to populate the metadata portions of it. The long-term monitoring databases are documented in eml and now our main effort is to archive and document the small research datasets that were produced during the first years of CAP. Over 100 datasets derived from CAP research, legacy GIS archives and the GP2100 eAtlas web site are now cataloged in EML and are online.

Our websites and web applications were migrated from a windows server to a Linux server. Therefore, legacy applications were rewritten in jsp. In addition, the CAP website now uses Xanthoria to retrieve data in EML format from our databases and the display is generated with xsl transformations. Development on our website Southwestern Environmental Information Network (SEINet, http://seinet.asu.edu) has reached production stage and it is now fully integrated into the CAP website as our data catalog. Datasets, literature and protocols may be found by using the Xanthoria search engine. The search capabilities for our biological collections in SEINet were expanded. Specifically, more spatial search capabilities were added. The locality input options have been expanded and it now allows the user to enter latitude and longitude for a bounding box. The application to actually draw a bounding box on an existing map as search criteria is currently being developed.

The CES lab has expanded its resources with two Tb of data storage, a new application server, and a new database server to host our biodiversity collections databases. The ASU IT Department's GIS lab was relocated to the Center for Environmental Studies to form a closer partnership with the Informatics lab and has contributed to several web programming efforts as well as the creation of EML metadata for the acquired spatial data.

An new interactive data analysis tool was added to our Ecology Explorers website (http://caplter.asu.edu/explorers). Developed under BDI funding, this wizard uses our Xylopia webservices toolkit and JSP pages to guide students through hypothesis formation and testing with graphic and statistical analysis. Students work with EML documented datasets from our central archive that have been configured to the application using XML. A new proposal in collaboration with Robert Bohanon (NTL) is currently in submission to NSF to further develop this approach for teacher training.

A new project central to CAP was initiated this spring. The 'North Desert Village Experiment' encompasses five identical housing complexes in which landscaping will be installed according to specification that are based on the findings from our extensive survey of the metropolitan area. A comprehensive data management system for this project is currently being developed. It will accommodate and integrate biogeochemistry, biodiversity, NPP, and human behavior research.

ASU recently received funding from NSF to create a Decision Center for Desert Cities. This project will build on the Information Infrastructure created by CAP and other funding from BDI and ITR to develop a coupled data - modeling - visualization platform to support the integration of ecological research and modeling in the policy-making and long-term planning of urban growth in the Southwest.

LTER Site: Florida Coastal Everglades

Contributor: Linda Powell (Jul 08, 2004)

Site Byte:

Improvements to the FCE Data Management and Web Site

Over the past year, Linda Powell (information manager) and Mike Rugge (project manager) have concentrated their efforts on two important components of the Florida Coastal Everglades (FCE) LTER program's information management system: 1) the Oracle9i Database and 2) a web-based project management tool.

Data migration from an MS Access database and flat files to the Oracle9i database is nearly complete. In the near future, FCE web site users (general public and FCE/LTER researchers) will have an option to download queried data from the Oracle9i database in addition to a possible download of the ASCII data files originally submitted by individual scientists. Currently, our updated 'data section' of the FCE website facilitates the exchange of LTER data, as well as, 'Everglades related' ancillary data. Users are able to search both our FCE and FCE ancillary data table of contents by several elements such as keywords, data category, and data originators. The Florida Coastal Everglades LTER Program now has a total of 128 datasets archived and 96 of those data files are publicly available via our website at http://fcelter.fiu.edu/data/.

Project and data management at the FCE LTER is very complex because the study area is roughly 1,400,000 acres and the site research personnel total 114 people to date. There are 21 'official' FCE LTER sampling sites and 260 'related' sampling sites so keeping track of the type of research, parameters, sampling points within a sample site and sampling frequency for each of the current 281 sampling sites is a very difficult task. Additionally, the scientists studying the Florida Everglades have generated numerous amounts historical data over the past 6 decades, some of which the FCE LTER is looking to archive in the information management system. The FCE LTER has added a new web-based interactive mapping application called the 'FCE LTER Interactive Everglades Map' to the FCE website at http://fcelter.fiu.edu/gis/everglades-map/ to facilitate information and project management. Mike Rugge (FCE project manager) was instrumental its design and implementation and is continuously working to integrate new information.

We used the following interactive mapping software and tools:

1) Mapserver - Developed by the University of Minnesota ForNet project and enhanced by the Minnesota Department of Natural Resources (MNDNR), open souce
2) PHP MapScript - Developed by DM Solutions Group, open source
3) ROSA Java Applet - Developed by DM Solutions Group, open source
4) GMap mapping engine (PHP/MapScript version) - Developed by DM Solutions Group, open source.

There are two main components of the application: 1) Interactive mapping and 2) Project and Data management. The interactive mapping feature allows the web user to build maps to their specifications by toggling GIS layers. They have the option to either save their own map creation or immediately print a hard copy. The GIS layers are available for download when the user selects the information button to the right of the GIS layers in the legend. A web user also has the ability to graphically select an area of interest from the map thus pulling up images and details about that site and its related projects, personnel, samplings and publications. If the user knows the name of the particular site of interest, they may also directly choose that site from a drop-down menu whereby they would receive the same information. Our researchers can use this application to help with future experimental designs as they can easily see what parameters are being sampled at a site and which researchers are involved. There is also a direct link to available FCE data and publications corresponding to that site. With information about sampling parameters (names, units, and frequencies) being collected at the start of a project, this tool assists the information manager to better manage the FCE data. In one visit to the web site, the manager can keep track of which data have already been submitted to the system and those data that have not.

EML Implementation

Since the inception of the FCE LTER, metadata has been entered into an Excel metadata template by individual researchers. When researchers submit a data file (.xls format) to the information manager, they also include a corresponding metadata file (.xls format). As was mentioned in Kristin Vanderbilt's site byte for the Sevilleta LTER, Kristin Vanderbilt cooperated with Linda Powell at FCE and programmer Gaurav Gupta at LNO to modify an existing Excel template for EML compliant metadata input and to develop a Java application that will translate it into EML. This project was abandoned by LNO after the departure of Gaurav Gupta in favor of creating a new, web-based data entry application. It isn't clear how long it would take to develop this 'new' tool (Development of this application was placed fourth on the list of priorities for the NIS development team by NISAC) so Mike Rugge (FCE) has recently spent the last week or so on developing a Perl script to convert our Excel metadata template into a valid EML document. Slight adjustments to the 'structure' of the metadata template are necessary and will be completed soon. Hopefully with the revised metadata template complete and the Perl script finished in the near future, the FCE will be able to generate EML. We are more than happy to share our 'EML' metadata solutions with other interested sites.

LTER Site: Georgia Coastal Ecosystems

Contributor: Wade Sheldon (Jul 27, 2004)

Site Byte:

Infrastructure and Data Harvesting

We acquired a 0.5TB network-attached storage server and dedicated data harvesting workstation this year, which significantly improved our IT infrastructure. The storage server allows PIs and project staff to securely store working files and backups centrally for the first time, which addresses a major review criticism and weakness of our highly de-centralized infrastructure. The drives are configured for RAID 5 to provide redundancy and data are backed up to DLT tape regularly. Users can access their directories via SSH2/SFTP and SSL from anywhere, or SMB, AFP and NFS networking protocols within the UGA Marine Sciences subnet.

The dedicated workstation has allowed us to acquire, process and post provisional data from our primary climate station and 2 affiliated USGS stations on an hourly basis, providing near-real-time data to our project participants and affiliates. New routines for automatic generation of time series plots at specified date/time intervals were added to the GCE Data Toolbox (http://gce-lter.marsci.uga.edu/lter/research/tools/data_toolbox.htm), along with routines for generating complete HTML code to render web pages of thumbnails for each plot with links to the full plot. We used this new tool to generate graphical summary pages for all long-term data sets on the GCE Data Portal (http://gce-lter.marsci.uga.edu/portal -- email sheldon@uga.edu for guest credentials), and to provide summary plots for real-time data after each harvest. This feature has been a huge hit with PIs, and many are now using this new thumbnailing feature to summarize their own data or model output.

New toolbox routines were also developed this year for temporal aggregation and resampling (i.e. aggregating on existing or auto-generated date/time component columns at any specified interval, e.g. yearly, monthly, daily, hourly), and auto-generation of data anomaly summary text in metadata based on analysis of flagged and missing values by date range.

Web and Database Enhancements

Adding support for EML dominated most web and database development effort this year (see below), but we did improve integration of our bibliographic database with both our data catalog and personnel database and now display these linkages on the web. For example, citations for all publications referencing each GCE data set are displayed on the corresponding data set details page (with hyperlinks to online documents when available). Personnel database links were used to develop a new automated reprint request service for all journal articles and theses; when a user follows the "Reprint" link and provides basic contact information, an email reprint request is automatically sent to the fist listed GCE-affiliated author (with instructions to pass it to a co-author if they aren't the corresponding author); when no active GCE-affiliated authors are available the request is sent to the IM or UGA Marine Institute librarian for assistance. Reprint requests cc'd to the designated IM and are logged to the database similarly to public data downloads (sharing the same session cookies to provide once-per-session entry of contact info), so we can assess the use of this service. We also provided LNO with a complete description of the URL-API for our bibliographic database web application in order to participate in new harvesting applications in development for the LTER all-site bibliography.

In addition, CSS styles were used to improve display of database-generated tables across the web site, and support for title string searches was added to the interactive filter panel on our data catalog page (http://gce-lter.marsci.uga.edu/lter/asp/db/data_catalog.asp).

EML Progress

In November 2003 we added complete support for dynamic generation of EML 2.0.0 metadata directly from our existing integrated metadata database, allowing us to provide EML on demand for all existing and future GCE data sets via our online data catalog (http://gce-lter.marsci.uga.edu/lter/asp/db/data_catalog.asp). We developed a lookup table and unit dictionary table to map our existing attribute classifications (data storage type, variable type, numerical type, units) to EML measurementScale and units in RDBMS views, so no modifications to our existing metadata content were required. As metadata for new data sets is added to the database a stored procedure automatically adds any unmapped combinations of attribute descriptors to this table so they can be manually fleshed out to provide full EML compliance. Custom units can be defined once and then re-used, and full descriptions in STMML syntax are added as an additionalMetadata section when any custom units are referenced in an EML document. This strategy allows us to continue working with our native metadata templates, database forms and data processing tools to process metadata and data, and produce complete EML and EML-described data on demand as value-added products. We will consider migration to more XML and EML-centric metadata management approaches as EML-based tools and technologies mature.

As part of this process we also added EML export capability to our personnel, geographic and taxonomic databases to support dynamic generation of coverage information for dataset metadata as well as independent EML documents for other purposes (e.g. exchange of species lists, personnel rosters, etc.). For example, we added support for generating custom species list in two different EML implemenations to our interactive taxonomic database application on the web (http://gce-lter.marsci.uga.edu/lter/asp/db/all_species_lists.asp). We also plan to add support for EML Literature to our bibliographic database as time allows.

After completing our EML implementation, we worked closely with Duane Costa at LNO and the Metacat developers at NCEAS this spring to establish automatic harvesting GCE EML documents for inclusion in the KNB Metacat server. This process required significant modifications to our EML implementation to support Metacat ids, versioning, authentication, and search and display characteristics. Consequently, we decided to provide Metacat-optimized EML as a seperate product of our web application (enabled using a "&metacat=yes" query string parameter), that contains these Metacat-specific features as well as alternative data urls for publicly-accessible data sets to enable direct streaming of data from our system (i.e. with transparent access logging) to support EML-mediated data analysis using Keppler and other new tools. We are now dynamically-generating a "harvest list" document based on these Metacat-specific revision numbers and urls, and our documents are harvested and updated in the KNB Metacat server on a weekly basis. This was a highly "educational" process for all concerned, that exposed a number of issues with our EML implementation and also with Metacat itself.

LTER Site: Harvard Forest

Contributor: Emery Boose and Julie Pallant (Jul 29, 2004)

Site Byte:

Computers and Networks

Over the last two years the number of networked computers managed by the Forest has increased to more than seventy, and we've moved from direct ownership to leasing as a more economical way to replace machines every three to four years. A comprehensive backup system was installed that automatically backs up data over the network from individual desktops and laptops to one of our servers. Wireless network access was recently implemented in a portion of our main building, and we expect that coverage will be expanded in the future as the number of laptop users increases. We're also looking into options for extending wireless coverage to some of our field sites.

Harvard Forest Web Page

The Harvard Forest web page was completely redesigned to provide expanded content and a more attractive and engaging interface. This project required about six months, and incorporated extensive input from Forest staff as well as assistance from a professional web designer and a graphics artist. Launch of the new page was concurrent with the move to a local server from its previous location at the LTER Network Office. A new web page was also created for the Fisher Museum and includes an online exhibit of the museum dioramas and online tour of the Sanderson Farm natural history trail.

Information Management

The Harvard Forest Data Archive was also completely redesigned over a period of about six months. Metadata for all datasets was updated by circulating a survey form to researchers. Discovery-level EML was created for each dataset by cutting and pasting from the surveys and other sources into a template using XMLSpy. An XSL style sheet was created to produce an overview web page (with discovery-level information) for each dataset directly from the EML. Work has begun on converting existing entity-level metadata into EML, and a second XSL style sheet was created to produce a data web page (with entity-level information) for each dataset directly from the EML.

Creation of entity-level EML by hand is fairly tedious and we are considering ways to facilitate this process. We are also studying options for long-term management of our metadata / EML and for adding dynamic query capability to our Data Archive.

LTER Site: Hubbard Brook LTER

Contributor: John Campbell (Jul 13, 2004)

Site Byte:

We have made substantial progress in our effort to convert existing Hubbard Brook metadata to EML. We have nearly completed the conversions for our non-geospatial data and now have 90 EML files that will be made public shortly (see http://www.hubbardbrook.org/eml/). We have also developed a stylesheet for our site that enables us to present the EML files in a user-friendly format that is similar to our past metadata format. David Blankman helped initiate the metadata conversion by developing a java-based tool that took a first pass at incorporating much of our existing ASCII metadata into EML. We have also relied heavily on Morpho, particularly for developing data tables. We have found XML Spy to be invaluable and prefer it to Stylus Studio, which we have also experimented with. Our goal is to have all our non-geospatial metadata into EML format by the Annual IM Meeting in July 2004. The next step is to work on converting the geospatial metadata from FGDC format to EML.

Hubbard Brook has had two information management staff changes over the last year. Ellen Denny, formerly of Yale University, relocated to New Hampshire and now works at the USDA Forest Service in Durham. Ellen helps maintain the Hubbard Brook web site and works on other information management issues. Netta Kies was hired for the summer to help with the EML conversions. Netta is a junior at Wesleyan University and has made substantial progress in the relatively short period of time that she has been working with us. We are fortunate to have them both working here and value their contributions. John Campbell continues in his role as Information Manager for Hubbard Brook.

We received a small NSF supplement last year to continue to populate the HydroDB and ClimDB databases. That money enabled us to substantially augment existing HBR data and metadata on the Hydro/ClimDB web site. We have now posted the following data on the Hydro/ClimDB web site: solar radiation, wind speed, relative humidity (1 weather station), min/max/mean air temp (8 weather stations), precipitation amount, and stream discharge (9 watersheds). There have been some preliminary discussions regarding the development of a chemical database, similar to Hydro/ClimDB, and Hubbard Brook has offered to serve as a pilot site for this project. We have also contributed spatial data to a project funded by USDA-FS that will build an internet accessible application aimed at linking spatial data layers to Hydro/ClimDB.

The Hubbard Brook GIS database has been expanded to include 27 coverages (http://www.hubbardbrook.org/research/gis/gis.htm). Additionally, all the metadata has been converted from a crude, text-based ASCII format to FGDC XML format. We are in the process of developing a stylesheet that will allow for a simple conversion from FGDC to EML and hope to post both FGDC and EML metadata on the HBR web site.

A number of interactive web programs have been developed that allow users to calculate and graph forest summary statistics (e.g. species-specific basal area, stem density, biomass) from various HBR vegetation datasets. The options for criteria selection are flexible enough to make this a potentially useful tool for researchers or students to generate summaries for very specific subsets of the data (e.g. a specific set of plots, a specific tree diameter range). The details of this project were published in the Fall 2003 LTER Network Newsletter (http://intranet.lternet.edu/archives/documents/Newsletters/NetworkNews/fall03/fall03_pg04.html).

LTER Site: Jornada Basin

Contributor: Ken Ramsey (Jul 26, 2004)

Site Byte:

Information Management System
The Jornada Basin Information Management System (IMS) databases are being populated with research data and metadata for all Jornada Basin LTER (JRN) research datasets. The IMS has been populated with JRN and USDA-ARS Jornada Experimental Range (JER) climate data collected from 2 weather stations and 68+ rain gauges dating from 1915 to current, GPS infrastructure and research site location data, and remote sensing imagery. All JRN dataset and project documentation are currently being converted to a new format that will support automated parsing of metadata from ASCII text files into SQL Server tables. As the metadata files are converted, the metadata is parsed into the RDBMS. Approximately 25% of JRN metadata have been converted and parsed into the IMS databases. Initial web-based forms have been developed to add additional metadata and to associate people in the RDBMS to datasets and projects in the IMS. Data entry and editing forms have been developed using PHP and Macromedia DreamWeaver using ADODB connections to SQL Server 2000 backend databases. The process of adding metadata other than that stored within documentation files, such as the personnel directory, has begun now that initial forms are available. The web-based database forms will be enhanced this Fall to make the interface easier to navigate and to add a help system and tool tips to help Information Managers and users to access and manage JRN metadata.

GIS and Remote Sensing
The JRN ArcSDE Geodatabase is operational and is available to JRN and JER researchers using ESRI ArcGIS and ArcView clients, ERDAS Imagine clients, and web clients using the ESRI ArcIMS site. The JRN interactive web-based map can be accessed at http://jornada-www.nmsu.edu/maps/JRN_Map with any browser that supports JavaScript. The JRN Geodatabase has been populated with data used to create 24+ published digital maps. These map layers include infrastructure such as pasture fences, roads, dirt tanks, and rain gauges. Recent research map layers include geomorphology map layers for the Jornada Basin that include landforms and parent materials and Desert Project organic carbon and carbonate carbon maps. The Spatial Data Laboratory is currently producing databases for historic brush control treatments. Three new Trimble GeoExplorer XT's sub-meter GPS units have been acquired to augment the 2 existing GPS units. JRN is building databases, forms, and protocols for handling GPS equipment and data.

EML Implementation
The Jornada Basin LTER has not been able to produce EML. JRN is awaiting delivery of an application from the LNO that will allow dynamic generation of EML from the IMS databases. The application would also allow semi-automated creation of EML from GIS metadata stored in ArcSDE. The application would allow users to edit and add additional metadata that cannot be automated prior to generating EML documents for harvest by Metacat. The plan to implement EML at JRN involves restructuring current metadata documentation, adding metadata content, populating and synchronizing site metadata into the IMS databases, and generating EML using the LNO application. Efforts are well underway for converting metadata files and populating the IMS databases. The application being developed by the LNO would use XSLT transformations from SQL Server tables to generate EML instance documents for harvesting into a JRN Metacat node which would then be replicated to the LNO Metacat node. The process of generating and harvesting EML could be automated using stored procedures and server-side scripts to generate new EML versions as underlying metadata contents changes. Metadata content will need to be added to existing documentation in the future in order to generate attribute-level EML. This will require a significant amount of time of both the Information Manager and JRN researchers in completing additional metadata, as well as developing web-based forms to assist with this process.

LTER Site: Kellogg Biological Station

Contributor: Sven Bohm (Jul 26, 2004)

Site Byte:

This year was a year of behind the scenes improvements. I am refactoring the code running the information system and abstracting out the database access code as part of the move from MS Access to MS SQL. The only major new on-line product is the GPS data such as sample points and field corners.

The other notable event this last year was the CPU failure on or main server. Since the system went down during the night (right after the incremental backup) we did not loose any data, but getting the system back on line required almost two days. Half a day to purchase and upgrade disks and memory, half a day to install windows, tomcat, apache, MSSQL server, MS Access, half a day to restore the backup (this was a worst case scenario since the system was due for a full backup the next day, we had to restore the month old backup and then a full month of incremental tapes), and half a day to restore odbc and application settings from memory.

Going through the recovery process has taught me that, while backups are essential they are not enough. On windows systems there is an amazing amount of info stored in the registry that needs to be recreated since the backup program does not back up files that are in use. This affected the setup of the odbc layer, and application parameters such as passwords and user names. I have since gone through the exercise of trying to identify points of failure, and devising, preferably automatic, procedures, to ensure that the relevant recovery data is explicitly available and up to date.

The EML style-sheet was completed last Fall. All of the datasets that are retrieved from the database are available as EML 2.0 documents. On our website they can be retrieved by clicking on the small EML label after the dataset. There are a few legacy datasets that need to be imported into the database yet, but as soon as that is done they will also be available as EML.

The current data delivery path is database -> jsp/java -> kbsxml -> (html | xhtml | eml | kbsxml). With this model adding new products such as EML is relatively easy. The main problems encountered was backfilling the metadata so that the required info was present. I took a few shortcuts in the EML style sheet to reduce the amount of info that was required. For example, we do not capture the concept of a non-numerical domain in our database, and I opted to deliver all non-numerical data as text.

LTER Site: Konza Prairie LTER

Contributor: Jobie Carlisle and Zhiqiang Yang (Jul 07, 2004)

Site Byte:

Wireless Local Area Network (WLAN)

Within the past year and half, a 2.4 GHz 802.11b wide area wireless network was established at the Konza Prairie Biological Station to provide:
. Remote data collection from, and sensor control of, field dataloggers.
. Real time access to Internet resources including on-the-fly GIS mapping services.
. Extension of audio and video capabilities to remote locations.
. Conference and classroom connections in existing indoor facilities and at remote 'field' locations.
Core backbone components include a Cisco 350 bridge that broadcasts a targeted "cloud' through a 14 dB 90o sector antenna located atop a dominant hill just inside the KPBS northwest boundary. The signal view-shed from this point covers over 80% of KPBS. This root access bridge is wired to a second Cisco 350 that is the endpoint to a 5.8-mile point-to-point wireless link to Kansas State University's Computing and Network Services infrastructure. Two 24 dB parabolic antennas focus the interconnecting transmissions. The WLAN has a shared Class C subnet with DHCP IP assignments and is secured using 128-bit WEP encryption keys. The point-to-point link security is enhanced with it's own unique WEP key and MAC address filtering.

Upgrades to KNZ Data Management, Web Site and Data Access

The Konza LTER program is currently completing a redesign of our data management system. The new system is based on MS SQL Server 2000. Most of the data and metadata have been successfully imported into the new database system from the former text based system, and the new system is being beta-tested. The new system will provide greater flexibility for our data management and more options for data access by users. The new data management system will be complemented by a new web site that will provide updated information and provide easy access to Konza LTER data.

EML Implementation

Based on our newly implemented metadata system, we have implemented an EML generation program with Microsoft .NET Framework. From the schema information stored in the metadata database, the program will extract all the metadata for a given dataset in a customized XML format. We designed a XSLT stylesheet, which will then transform the customized XML into EML 2.0 compliant document. In addition, a separate XSLT-FO stylesheet was implemented to generate a PDF document for the metadata. All these functions will be included in our new web site.

LTER Site: LTER Network Office

Contributor: James Brunt (Jul 29, 2004)

Site Byte:

LTER Network Office Informatics Update #13 - LTER Network IM Meeting

New Personnel at the LTER Network Office:

Duane Costa - Analyst / Programmer - Duane started last year as part of the KNB project and transistioned in May to the Network Information System development project.

Jeanine McGann - Web Designer - Jeanine started in January of this year. Jeanine was previously a graduate project assistant with LNO working the KNB project and EML documentation.

Michelle Murillo - Data Manager - Michelle started in December, replacing Troy Maddux. In addition to database management Michelle serves as backup to Greg Shore on Linux Administration.

Mark Servilla - Network Information System Lead Scientist - Mark started in May and is already taking intellectual ownership of the Network Information System Effort.

Thomas McOwiti - Public Information Representative - McOwiti, as he prefers to be called, replaces Patty Sprott doing PR for the LTER Network.

New Initiatives:

NBII/LTER Network Office Metadata Standards Initiative - As the culminating point of three years worth of meetings with the USGS/NBII we have submitted a proposal under a new Memorandum of Understanding with NBII that will place a highly-skilled technical position in the network office later this fall. This persons duties will be to complete the FGDC/BDP crosswalk to EML and provide a web-based tool to effect these translations, to help individual sites to implement metadata standards towards becoming part of a larger network, and to host a workshop to discuss interoperability between FGDC/BDP and EML.

Network Information System Advisory Committee and Strategic Planning - The LTER Network Office, LTER Network Information Managers, LTER Network PI's have been involved in an intensive planning effort towards the development of a Network Information System. Part of this effort includes the prioritization of LNO NIS efforts, the discovery of new technologies, the adoption of an LTERwide Collaborative Web Portal, and the planning for global IT infrastructure.

New Facilities - We are now at the end of a very long process to get new space for the LTER Network Office on the main campus of UNM. Our new offices are located on the 3rd floor of a remodeled building in the central part of campus. Being located on campus will be much better because we will be co-located, close to colleagues, close to food and supply outlets, we will have access to many more meeting facilities including the newly remodeled Student Union Building and a dedicated and specially designed training facility and access-grid conference room. LNO staff sucessfully transistioned all the cyberinfrastructure with a minimum of downtime. The new facility has a dedicated, climate controlled server room and data storage room. The training lab will be outfitted over the next several months to be completed in time for the OBFS/RDIFS training in October.

New Equipment - Last year we upgraded our Intel-based servers for local file system storage and our database server. This year we've replaced our web server and the metacat/srb/ldap server with intel-based systems running Redhat Advanced Server. The new web server is up and running and the new metacat server will be coming along shortly.

Ongoing Efforts:

Network Database Integration - We've been actively engaged in the network database integration project. Marshall White is the lead developer on this project backed up by Michelle Murillo. The first requirements of unifying with a common ID structure, a one login/one password security model and meeting the old system functionality is complete. Phase II will add many new features including the new and easier ways for sites to provide updates - Phase III will enable support for EML. The new design is modular, written in PHP+ and perl with objects being designed for reuse. This effort is also closely linked to the RDIFS project (described below) which extends these features to the OBFS. Documents related to this effort can be viewed at: http://intranet.lternet.edu/projects/informatics/
Contact: Marshall White (mwhite@lternet.edu)

Strategic Planning - The LTER Network Office is in the final stage of the process of developing a strategic plan. This plan, called for in the new cooperative agreement, will help us better meet the needs of the Network through goal-oriented decision making and priority setting. Part of the ongoing strategic planning process will be the development of a system of evaluation and feedback for the LTER Network Office by sites, PI's, and scientists. Strategic planning will also determine the level of emphasis and effort applied to informatics and IT. Contact: Robert Waide (rwaide@LTERnet.edu).

Science Environment for Ecological Knowledge (SEEK) - we're completing the second year of this NSF-funded five-year collaboration between the LNO, the San Diego Supercomputer Center at UCSD, the Natural History Museum and Biodiversity Research Center at the University of Kansas, Arizona State University, National Center for Ecological Analysis and Synthesis at UCSB and a number of other collaborators. This large ITR project is focused on developing new technologies for the integration, analysis, and synthesis of heterogeneous ecological data. The goals of SEEK are to make fundamental improvements in how researchers can gain global access to ecological data and information, rapidly locate and utilize distributed computational services, and exercise powerful new methods for capturing, reproducing, and extending the analysis process itself. In addition to administrative and logistical aspects of the project LNO is responsible for coordinating training for new investigators. See: http://seek.ecoinformatics.org
Contact: Bill Michener (wmichener@LTERnet.edu).

Spatial Data Workbench (SDWB) - we're in the final year of a collaborative project between the LTER Network Office and the San Diego Supercomputer Center (SDSC), as part of the National Partnership for Advanced Computational Infrastructure (NPACI) Earth System Science initiative. Data are made available through a Storage Resource Broker (SRB) at San Diego that provides for the storage and dissemination of large LTER spatial datasets. The system contains airborne hyperspectral (AVIRIS) data for 1997 through 2001, for the Sevilleta, Jornada, Konza, Harvard Forest and Andrews LTER sites, and Landsat Thematic Mapper images, GFL reconnaissance data and current MODIS data for all of the LTER sites. This infrastructure is being mirrored at the LNO where these and other image types will also be uploaded and the metadata registered in the LNO metacat. See: http://sdw.sdsc.edu/ Contact: John Vande Castle (jvc@LTERnet.edu).

Resource Discovery Initiative for Field Stations (RDIFS) - a Research Coordination Network Grant from NSF to develop the informatics framework that can facilitate storage, discovery, and access to the strategic environmental information resources that are collectively held at North American biological field stations now in its second year. LTER Network Office activities focus on the development of databases that will support discovery of biological data and information resources at field stations and training in informatics. The biological field stations are being included in the network database integration project which will include a Site Characteristics Database for North American Field Stations; a Bibliography of North American Field Station Publications; and a Database of Field Station QA/QC and Standard Methods. Education this year will principally focus on the annual intensive (2-week) training workshops in ecological informatics that are developed around a series of course modules - this year the first week is devoted to data and information management while the second week is an intensive course in remote sensing and GIS. Contact: Bill Michener (wmichener@LTERnet.edu)

Knowledge Network for Biocomplexity (KNB) - is a collaborative project between the LTER Network Office, the National Center for Ecological Analysis and Synthesis, and the San Diego Super Computer Center funded by the Knowledge and Distributed Intelligence (KDI) program at NSF. Now in it's fourth and final year , the intent is to bring closure to the project by hosting a workshop in September to orient practicing scientists to the tools developed. LNO effort this year focused on the completion and deployment of the metacat harvester. See: http://knb.ecoinformatics.org Contact: Matt Jones (jones@NCEAS.ucsb.edu).

Computing and Communication Infrastructure services in support of the LTER Network:
Integrated Website Design and Planning - Marshall White has now worked with several sites to facilitate the development of website plans that are functional and maintainable in a network context. I've watched this process and think it is really an excellent exercise for any site considering redesigning their web presence. Contact: Marshall White (mwhite@LTERnet.edu)

Centralized Personnel Database -Web interfaces to this database allow searching and quick access to people in the network and provide easy update for individuals and data managers. Templates can be provided to sites for including the search and update routines in site web sites eliminating the need for sites to maintain their own personnel database. Contact: Marshall White (mwhite@LTERnet.edu)

Individual Mail Alias - provides for each LTER person an @LTERnet.edu alias, for e.g., mharmon@lternet,edu, spickett@lternet.edu. This provides a convenient way to correspond with anyone in the LTER Network. The use of this service as steadily increased over the last 10 years.

Mail List Management - this function is migrating to a group management module of the integrated databases - it will still provide the capability to create and manage site, committee, or special interest mailing lists, @LTERnet.edu, based on the personnel database but will be more extensible and implements the mailing list as a function of a group.

LTER All Site Bibliography - While it's primary function is documenting LTER research this database of LTER publications from 24 LTER sites is a searchable resource to find and download information about published information from the LTER Network. The LTER Network Office has migrated this data to MySQL , completed new interfaces, and is nearing completion of integration into the web framework. With features like Endnote import and EML export, this service can completely take the place of maintaining a local bibliography and can be accessed via the web interfaces or through any Endnote or other scientific bibliography client.

Domain Name Aliases / DNS services- Individual sites can get an lternet domain web address for their web site that uses their 3-letter abbreviation. For example:
Palmer pal.lternet.edu
Luquillo luq.lternet.edu
Hubbard Brook hbr.lternet.edu
Coweeta cwt.lternet.edu
This has been done for all sites and for those without virtual web server capability redirects have been put in place. It moves us toward a network identity and provides the DNS infrastructure necessary for a more integrated web presence. This also allows individual web sites to be indexed in search engines so they can all be searched at once. In addition, sites customized domain names, for e.g., sevilleta.org, obfs.org, can be hosted.

Web/Database Hosting - Sites that do not have the resources or do not wish to host their own web site can use the LTER Network servers to host their web site. Several sites are doing this already as the costs of personnel time for managing computer systems has increased. The extensions available include PHP, MySQL, postnuke, perl, and other open source approaches to web applications.

New LNO Phone Numbers:
Andrea Briscoe - 505.277.2503
Brooke Brooks - 505.277.2534
James Brunt - 505.277.2535
Duane Costa - 505.277.2708
Fax Machine - 505.277.2541
Subhek Garg - 505.277.2750
Pam Griego - 505.277.2551
Jeanine McGann - 505.277.2756
Thomas McOwiti - 505.277.2638
Bill Michener - 505.277.2769
Michelle Murillo - 505.277.2578
Sonia Ortega - 505.277. 2592
Deana Pennington - 505.277.2595
Katy Perry - 505.277.2597
Samantha Romanello - 505.277.2607
Greg Shore - 505.277.2620
Steve Tekell - 505.277.2625
John Vande Castle - 505.277.2634
Robert Waide - 505.277.2649
Marshall White - 505.277.2667

LTER Site: Luquillo LTER

Contributor: Eda C. Melendez-Colom (Jul 23, 2004)

Site Byte:

Since LUQ LTER 3's review in June 2003, IM at Luquillo has had two main goals: to continue tightening the collaboration with the LUQ's investigator, and to migrate our metadata to EML.

The former was favored by an ongoing activity that I got involved almost two years ago with the Evergreen State College Canopy Database Project under the direction of Judy Cushing. Producing two sets of visualizations for 3 censuses of a grid project that started in 1990 allowed the investigators to learn more about their 16-ha. plot. Also, I had the opportunity to work with the Canopy Database team and to participate actively in the analysis process of the data with the LUQ investigators. Two posters were produced in collaboration by the three groups of the scientific community: scientists, researchers in information technology, and information managers. I am pretty sure that the benefits of these activities will continue in the future.

While I was able to migrate some of my metadata in one of the workshops that we participated, the migration of all of LUQ's metadata has turned out to be a more time consuming task. Two main situations are being taken care. Although metadata online seems to be satisfactory for the user that have accessed them, they were not ready to be incorporated into an EML database. The structure and the format of the metadata database are being modified, and the quality of the metadata is being enhanced. Another important feature is the format and structure of the data files or databases per se. One definitively cannot give structure to the metadata if the data files' structures are poor or non-existent.

By participating in the web services workshop this year, I realized that databasing the web should be also part of LUQ IM current efforts. In a web site like the LTER's this process has to start with the scientific databases. Many of the LUQ's data files lacked structure, or their structure is or was not appropriate for a database. So we started, and are about 2/3 done, with transforming the structure of many of our almost 100 online databases. At the same time that the data is transformed, the metadata is transformed so it will be transferable to EML. David Blankman assisted me in transforming some of the ASCII-format metadata into databases, but these have to be revised too, and finally incorporated into the metadata database.

I believe that, although LUQ has taken the longest path, the benefits will be evident in the future and it will finally give a "common data management framework" to LUQ's Information System.

LTER Site: Niwot Ridge LTER

Contributor: Todd Ackerman (Jul 26, 2004)

Site Byte:

The Niwot Ridge LTER site submitted our proposal this year. Information Management was given some suggestions for improvement- mostly based on our out-dated website (it is functional, but not very pretty). Work has begun on the redesign of the site, however we have recently learned of the LNO helping some sites with website design and would like to inquire about obtaining help.

EML conversion is nearly finished for the existing flat file datasets. The original plan of exporting our native XML from the relational database was abandoned because we never received the XSLT's from the network office in order to do the conversion. Therefore a perl script has been written to query the database and output the metadata as EML. The last step that is currently being worked on is to deal with transforming the existing attribute information into the EML measurementScale section. Steps have been taken to start on this, but will probably not be completed until shortly after the IM meeting in Portland. Spatial data still needs conversion to EML.
Plans for the local use of EML are unclear at this time. Our current process of querying of local metadata works well and the PI's are happy with its function. EML will be available for download if the user needs it, and can be loaded into whatever catalog the Network wishes.

Last year we set up wireless connections to several of our meteorological towers. This year, we have set up an automatic import into the relational database of climate data. The data is imported on a weekly basis so the researchers can query more recent climate data. Public access has yet to be set up for this data, due to the metadata not being up to par yet, and the fact that the server it resides on is rather slow. The tool has been made available to the climatologist which is useful for the QA/QC process. The query page can be viewed at http://culter.colorado.edu/exec/Database/database_query.pl. We would like to get this these loggers connected to ClimDB and HydroDB soon. We are currently calibrating the wireless logger at the Albion Town Site for its input into HydroDB.

Last year we had been using ESRI's Metadata Explorer to serve our GIS data. That product proved to be unreliable, so we had to rethink the process of serving GIS data. We have two methods of serving the data, the first is an ArcMap project available at: http://dbkong.colorado.edu/Website/NiwotRidgeMap/NRM_instrux.html, and the second is a page generated by SQL server: http://culter.colorado.edu/exec/Database/gis_layer_query.pl.

At Niwot Ridge, all of the data collected passes through the information management lab in its most raw form. We have been redesigning/updating some of the archaic data entry procedures. Previously when data had been received via hard copy data forms, it was entered into a DOS based software that was confusing to the user, difficult to move around in, and difficult to develop new forms. We have redesigned the data entry forms in Visual Basic so that the form on the screen mimics the hard copy data sheet, so the entry process is more straight-forward. These new forms use re-key verification and lookup tables to reduce human error of the data entry process.

LTER Site: North Temperate Lakes

Contributor: Barbara Benson, Dave Balsiger, and Jonathan Chipman (Jul 23, 2004)

Site Byte:

We have designed and implemented an upgraded information system for near real-time data from instrumented buoys to accommodate the increase in the number of instrumented buoys. The sensors measure meteorological variables such as air temperature and wind speed and lake parameters including water temperature at depth and dissolved oxygen. The new system includes automatic uploads to the Oracle database, automated range checking for quality assurance based on Oracle triggers, and near real-time graphical displays to assist in buoy maintenance by the field technician. The data are uploaded to the database every hour and can be queried dynamically from the North Temperate Lakes (NTL) website.

NTL is collaborating with scientists in Taiwan to study lake metabolism using instrumented buoys and wireless connectivity to share near real-time data from field sensors on two sides of the Pacific Ocean. This collaboration involves scientists at NTL, the University of California San Diego, San Diego Supercomputer Center, the Taiwan Forestry Research Institute, the Taiwan National Center for High-Performance Computing, and Academia Sinica in Taiwan. NTL scientists, a technician and a graduate student have visited the field site in Taiwan and helped with installation of the instruments, and scientists from Taiwan have visited the NTL site. We have videoconferences regularly that have involved multiple sites in the U. S. and Taiwan using Polycom equipment. Proposals have been submitted to extend the lake metabolism network to additional countries.

We have had further successful experience with using the dynamic database program we originally created to permit dynamic querying of the NTL Oracle database from the NTL website. The designed flexibility of the program has enabled us to easily provide comparable dynamic queries for other projects.

During the past year, we have been working to enhance our spatial data management system. We began by cleaning up and improving the existing, static "spatial data catalog" by updating many data sets and standardizing coordinate systems and file formats. All vector data, and some raster data, are now being compiled in a geodatabase, to facilitate data management and data sharing. During the coming months we plan on establishing a set of three ArcIMS map-serving applications for our web site (covering NTL's northern and southern lakes regions, plus a statewide-to-regional scale version).

One spatial-data milestone was reached on June 3, when our fully-automated MODIS image processor downloaded and archived its 1000th MODIS satellite image. (On a daily basis, this system automatically retrieves Level-1B MODIS imagery from our EOS Direct Broadcast reception facility, extracts the region of interest for NTL-LTER, and performs a variety of geometric and radiometric enhancements.) While much of the May-June period was quite cloudy, June 3rd turned out to be a spectacularly clear day, and this 1000th image was almost completely cloud-free.

NTL has been active during the past year in network-level information management planning activities. Barbara Benson is one of the PI's on the network planning proposal submitted to NSF in spring 2004 ("Preparing the LTER Network for collaborative science, education and synthesis: A planning proposal). NSF has approved this proposal for funding. Tim Kratz and Barbara Benson are serving as members of the Network Information System Advisory Committee, a standing committee that conducts strategic planning and evaluation for the LTER Network Information System. This committee met in March and June 2004 to create a strategic plan for the Network Information System and engage in interactions to further collaboration with NCSA and SDSC.

The NTL book, Long-Term Dynamics of Lakes in the Landscape, in the LTER Oxford series has recently been sent to the publisher. Barbara is one of the book editors, and the book contains a chapter on NTL and LTER information management.

EML Status

EML documents for on-line NTL data are created by a java application and are available on the NTL website. The application is driven by metadata in the NTL Oracle database. Metadata are entered directly into the database by information management staff. The published NTL documents are not yet valid EML as we still need to populate some required elements. We have created a script that dynamically generates the harvest document for the central LTER Metacat. There are data versioning issues yet to be addressed.

LTER Site: Palmer Station

Contributor: Karen S. Baker (Aug 29, 2004)

Site Byte:

PALMER INFORMATION MANAGEMENT supported the annual January cruise and the seasonal October to March field work at Palmer Station, Antarctica. The second year of the 3rd Palmer LTER proposal cycle initiated technology review process and infrastructure redesign in order to support transitions to new approaches to networking as well as to data and metadata structures.

LTER NETWORK ACTIVITY included participation in LTER Information Manger Committee activities and annual meeting (Aug04), a web services workshop (Jan04) at SDSC, and an LTER IM Executive Committee Meeting (Jan04) in San Diego. Contributions were made and support provided for the semi-annual Databits Newsletters. Investigations of collaborative tools as part of a web site upgrade are underway and synergize with ongoing efforts at the LTER Network Office, at other LTER sites, and at Scripps Institution of Oceanography (SIO).

SITE DATA MANAGEMENT activities included annual updates to the Palmer LTER data repository, document archives, and to the site bibliography. In preparing for redesign of the Palmer LTER data system, a workshop (Dec03) on Data Management Exchange provided input on the Joint Global Ocean Flux Studies (JGOFS) Data Management System, recognized as one of the oceanographic community standards. Installation occurred this year at the request of the Palmer scientific members familiar with its user interface. Additional analysis packages as well as community exchange protocols are also being investigated as part of efforts to collaborate with community partners. METADATA STATUS: Palmer LTER data formats and metadata tags are being updated and a set of XML dictionaries developed and populated as the set of local SIO, JGOFS, and LTER Ecological Metadata Language requirements inform development of new procedures for data quality control and normalization.

COMPUTATIONAL INFRASTRUCTURE for the Palmer LTER within the Institute for Computational Earth System Science (ICESS) at UCSB has continued without major change except for upgraded security vigilance. Partnership with the Integrative Oceanographic Division at SIO has added support for Mac platforms and Linux as well as renewed focus on open source software (Apache, MySQL and Php). Installation this year of a new storage and web server system provides a contemporary foundation from which to build infrastructure and collaboration with the new California Current Ecosystem LTER (CCE) site located at SIO.

Ongoing Palmer LTER IM INTERDISCIPLINARY RESEARCH, in collaboration with Science & Technology Studies, bridges environmental, information, and social sciences. In addition, collaboration with education/outreach endeavors is in the midst of a two year cycle to design and prototype a conceptual framework for site outreach supported by information management. This work has been presented to the LTER community via talks and written communications. Papers were presented at the Hawaii International Conference on System Science (Jan04) and the Participatory Design Conference (Aug04) and a poster presented at the LTER All Scientists Meeting.

LTER Site: Plum Island Ecosystem

Contributor: Hap Garritt (Jul 27, 2004)

Site Byte:

PIE Site Byte 2004

The Plum Island Ecosystems (PIE) LTER received renewal acceptance this past Spring, 2004. Review comments regarding data management focused on the need to integrate the various GIS datasets into the PIE database. Three groups associated with PIE LTER conduct GIS research, University of New Hampshire (UNH), Marine Biological Lab (MBL) and Clark University. Two of the groups, UNH and MBL use ArcGIS/ArcView software while Clark University uses IDRISI software that it developed and updates based upon feedback from researchers. During the PIE May 2004 ASM meeting, a working group discussed PIE GIS, focusing on availability of on line GIS datasets, GIS metadata and conversion of metadata to EML 2.0.0. There are compatibility issues between ArcGIS/ArcView and IDRISI, which we are addressing. We reviewed the number of data layers that have been developed, 29 (with some redundancy) and decided that each group should submit their data layers to the PIE IM. The data layers will be reviewed for redundancy, incorporated into the PIE database and made available through a server at MBL. Due to the large size of the GIS data layers, an additional hard disk drive will be purchased. To avoid redundancy and conserve disk space, data layers of USGS Ortho photos, USGS topographic maps and NOAA charts will be made available using hot map links to the State of Massachusetts GIS web site http://www.mass.gov/mgis/, which maintains and is the original source for extensive GIS information for Massachusetts. GIS metadata (FGDC) was reviewed during our ASM meeting and a minimal standard subset was agreed upon. We are in the process of determining how best to convert the FGDC format to EML 2.0.0. An initial conversion, thanks to the use of a script from Peter McCartney (esri2eml.xsl), Stylus Studio and the Morpho validator was successful but a very time consuming iterative process.

EML 2.0.0 conversion for text-based files has just begun recently as we received the EML conversion software from David Blankman the end of June 2004. XML files have been generated from metadata files given to David back in 2002 but the software does not work with newer metadata files. We are currently working on how to transform our existing attribute information into the EML attributeList measurementScale section of EML 2.0.0. I am also wondering if it would be easier to re-enter all of our metadata if only there was a simpler/down scaled version of a Morpho type program. At PIE, to minimize loss of metadata information, we prefer to have the researchers and associated research assistants enter their own metadata but that requires a fairly simple template. We are very interested in exploring various options for simplifying metadata entry and conversion to EML 2.0.0 including the Excel spreadsheet templates other LTER sites are using or the Web based data entry application that the LNO is considering.

LTER Site: Santa Barbara Coastal

Contributor: Margaret O'Brien (Jul 27, 2004)

Site Byte:

Santa Barbara Coastal had a change in information management staff during the last year. Wei Yee Luan left the project in mid-2003 and Chris Jones served as acting information manager for the remainder of the year. In January 2004, Margaret O'Brien assumed the role, relocating from the Institute for Computational Earth System Science, also at UC Santa Barbara. SBC-LTER has continued its partnership with another coastal ecology research program (PISCO) to leverage expertise in scientific data management, EML implementation and web development. Improvements to our website include dynamic generation of our personnel page from our LDAP database and tools to facilitate scheduling and user account management. Since any interdisciplinary project employs data from divergent sources, assisting scientists with access and data manipulation continues to be a high priority.

Since SBC has not yet established a relational database, we are considering the practicality of using EML as our primary format for storage of both meta and scientific data. Our core research activities are now cataloged in 20 basic data packages, each containing one to several data tables ranging in size up to 30mb. Currently, these packages contain data though 2002 and their tables will be appended with new data as necessary using perl scripts being developed here at SBC. About 70% of SBC's core metadata is now stored as EML2.0, and complies with the anticipated EML Best Practices Level 4. It is anticipated that update of the remaining 30%(in EML2beta6) will be completed this summer. Data packages can be viewed at http://data.piscoweb.org/catalog. EML2.0 metadata files and data tables also reside at http://sbc.lternet.edu/external/EML/Data_packages. XSLT stylesheets for viewing packages on our own website are in progress. The next step will be to insert data tables into the database (Postgres), and begin programming tools for searching tables.

In addition, we are now storing taxonomic species lists and our customized list of units in EML. A literature database in XML (based on the eml-citation schema) is in progress. SBC's data manager participated in the Network-wide EML Best Pracitices working group, which met in May of this year.

LTER Site: Sevilleta LTER

Contributor: Kristin Vanderbilt (Jul 06, 2004)

Site Byte:

The Sevilleta unveiled a new website in January 2004 (http://sev.lternet.edu). Sevilleta IM personnel Kristin Vanderbilt, Mike Friggens, and Harsha Belludi collaborated with Marshall White from LNO to design and implement this new site in the PostNuke content management system. In addition to a fresh look and feel, the new web site is more intuitive to use and has much more dynamically generated content. Renee Brown, Sevilleta system administrator, has also established an intranet, which is a new feature of the Sevilleta web site. Web-based data entry programs being developed by Harsha will be accessible from the intranet, as will WebMail, a new email client made available to Sevilleta associates in July 2004.

Hardware upgrades at the Sevilleta included the addition of a SUN A1000 hardware RAID to the SUN E450 server, adding approximately 400GB of much needed space. The E450's memory was also doubled, amounting to 2048MB. The Sevilleta network's Dell PE2300 server was replaced by a Dell PE2650 as part of the SIMS server infrastructure upgrade. A PV775N (Network Attached Storage or NAS unit) was also included as part of this upgrade, and together the PE2650 and PV775N provided an additional 382GB of space, primarily for the use of backing up Solaris, Windows, MacOS, and Linux clients.

A wireless backbone has been established at Mckenzie Flats on the east side of the Sevilleta National Wildlife Refuge to aid in electronic data collection. David Hughes, wireless guru from Old Colorado City Communications, visited the Sevilleta in April to locate sites to place equipment and then returned to orchestrate the installation in early June. Many Sevilletans enjoyed the fine view from the helicopter as it carried people and equipment to the top of the granite escarpment of the Los Pinos Mountains overlooking Mckenzie flats. The so-called backbone of the system consists of two Trangolink10 EXT Bridge Radios which are 5.8 ghz spread spectrum radios. One of these is situated on the roof of the Field Station conference/lab building and the second is sited on the top of the granite escarpment that borders the east side of Mckenzie Flats. This is a distance of 25 km. The radio on the escarpment is linked to two Smartbridge Access Points. These two Access Points are 802.11b radios (a standard wireless Ethernet networking technology) and are arrayed to cover all of Mckenzie Flats. During this first phase of the installation the goal was to install radios and the hardware that would allow access to Campbell dataloggers at four sites on Mckenzie Flats. These four sites were selected to explore the feasibility of connecting over various distances through various Access Points to one or two dataloggers of different vintages (CR7 CR10, CR10X CR23X). Data collected include water and carbon dioxide fluxes, soil temperature and moisture, and weather data. This wireless system should allow anyone with Campbell access software and the IP addresses of these stations to call up any of theses stations and see data in real time over the Internet from anywhere that has Internet access.

IM Kristin Vanderbilt participates in Network Level activities by serving on the IM Executive Committee. She also collaborated with Longjiang Ding and Peter Arzberger at the San Diego Supercomputer Center to organize the January 2004 web services training workshop for LTER information managers and international IT personnel. She contributes to the SEEK (Science Environment for Ecological Knowledge) project by co-teaching workshops for post-docs in ecology, and also co-teaches information management workshops for personnel from the Organization of Biological Field Stations (OBFS) in support of an RCN proposal with LNO.

EML Status:

LNO personnel determined in late 2003 that extant Sevilleta metadata are unstructured enough so as to be impossible to parse into EML. Janine McGann of LNO therefore cut and pasted metadata from a few Sevilleta datasets into EML to serve as examples for converting the rest of the legacy metadata. A few legacy datasets at the Sevilleta have since been agonizingly converted into attribute-level EML using Morpho. The rest of the legacy data will be converted into discovery-level EML this fall as time permits.

To capture new metadata, Kristin Vanderbilt cooperated with Linda Powell at FCE and programmer Gaurav Gupta at LNO to devise an Excel template for metadata input and a Java application that will translate it into EML. This project was abandoned by LNO after the departure of Gaurav Gupta in favor of creating a new, web-based data entry application. The capabilities needed from this tool were discussed by James Brunt, Mark Servilla, Linda Powell and Kristin Vanderbilt in early June. Development of this application was placed fourth on the list of priorities for the NIS development team by NISAC. The Sevilleta still lacks a clear plan for collecting new metadata in a way that makes it easy to translate into EML.

LTER Site: Shortgrass Steppe

Contributor: Nicole Kaplan and Bob Flynn (Jul 07, 2004)

Site Byte:

The Information Management Team (Nicole Kaplan, Bob Flynn, and John Norman) at the Shortgrass Steppe LTER has focused on updating IT equipment and systems, as well as data and information to support local and Network-wide research. In addition, the team is strengthening partnerships with other LTER and ecological research sites to leverage expertise and experience in database management, web development, and EML implementation. We have supplied our field station and technical staff with new desktops, laptops and handhelds to serve a growing number of scientists and students in residence. Mobile equipment and a distributed network of data loggers are used to collect field data from study sites in the field. There is connectivity from the field station to the new SGS-LTER Information Network Server on campus so data can be captured and backed-up in the system within twenty-four hours of collection and entry. John and Bob are revising legacy quality control programs originally composed in Fortran, for many of our data tables that support vegetation research and relate to over 60% of our studies. John has expanded this suite of programs, which we call the Matrix, to process meteorological data for submission to CLIMdb (http://www.fsl.orst.edu/climdb/). The SGS-LTER IM Team will continue to work with researchers to development tools that will more efficiently process and publish their data with high integrity. Bob and John have organized the GIS library and are developing search tools for spatial data discovery.

Nicole continues to represent the IM Team and work closely with researchers and students to update the SGS-LTER and network-wide databases. Nine new data tables relevant to prairie dog and regional biogeochemical research were published on the SGS-LTER website (http://sgs.cnr.colostate.edu/Data/AcquisitionPlcy.htm). CLIMdb also was updated with SGS-LTER meteorological data from two weather stations. Nicole continues to investigate and acquire more robust metadata content for legacy data sets to archive and create documents in Ecological Metadata Language (EML). The IM Team has tested several tools that have been developed within the ecoinformatics community to create EML, including Morpho 1.5 (NCEAS), an EML MS Excel Spreadsheet Template (FCE), and an EML Relational Database Management System (RDBMS) conceptual model (Blankman, LNO). David Blankman and Marshall White from the LTER Network Office (LNO) visited the SGS-LTER IM office in October of 2003 to help support IM Team efforts to generate EML and improve the SGS-LTER website by making information more accessible and manageable. LNO staff and SGS-LTER IM Team members communicated the importance of creating EML, improving access to information on LTER web sites, and partnering with sister sites and LNO staff to leverage expertise across the ecoinformatics community.

The IM Team has focused their efforts to generate EML on redesigning and retrofitting the existing RDBMS to support EML implementation. The current RDBMS contains the necessary metadata to support the implementation of the "datatable" module of EML. However, these metadata are not related or normalized within the RDBMS. In the past, the SGS-LTER RDBMS was designed to support the logistical procedures necessary to approve applications to conduct experiments and manage project level information over time. Currently, this project level and dataset level information is published discretely using functions in the database and website. The IM Team is working to establish new relationships between the tables that contain the necessary information to create a valid EML document. The IM Team is exploring ways to generate XML from the more integrated RDBMS design and then convert this XML to EML using style sheet tools. In addition, Bob is focusing on GIS metadata documentation procedures and tools to generate EML.
Nicole continues to collaborate with Judy Cushing and others (The Evergreen State College), Kristin Vanderbilt (SEV), Ken Ramsey (JRN), Eda Meléndez-Colom (LUQ), Jonathon Walsh (BES) on the Canopy Databank Project (http://canopy.evergreen.edu/). They will be presenting a poster at the Ecological Society of America 2004 meeting in Portland, entitled Database Tools for Ecological Data Integration and Synthesis. One focus of their work has been designing databases to integrate and synthesize annual aboveground primary production data from three grassland LTER sites that are employing different methodologies to measure productivity. Nicole stays closely connected to the LTER IM community as a member of the IM Executive Committee. She also participated in workshops, including the use of Web Services to integrate datasets, sponsored by the LNO and San Diego Supercomputer Center, and co-authored an article about the workshop in the Spring 2004 LTER Network Newsletter (http://intranet.lternet.edu/archives/documents/Newsletters/NetworkNews/spring04/spring04_pg06.htm).

LTER Site: Virginia Coast Reserve

Contributor: John Porter (Jul 27, 2004)

Site Byte:

The year got off to a bang when, during the LTER All-Sci. meeting, Hurricane Isabel paid a call to the VCR/LTER. Fortunately, despite extensive flooding, we only had one of our three meteorological stations destroyed. The hurricane moved enough inland that although record flooding was observed, like our primary research site on Hog Island, most of the barrier islands showed few direct effects of the storm. However, there were exceptions. Myrtle Island, which has been retreating in recent years, showed substantial changes after the storm and has been the subject of repeated faunal surveys looking at the impact of the extensive flooding. See http://ecocam.evsc.virginia.edu/image_selector_isabel_index.html to view images of Hurricane Isabel.

On the information-management side, we continue to develop our wireless network on the islands. Our wireless Ethernet extends over 30 km, from our lab throughout Hog Island and adjoining islands. A series of web cameras are capturing imagery of the islands, focusing on Fiddler Crabs, bird nesting colonies and general landscape changes. The high flow of images (at least hourly for each camera, with some every 5 min.) has posed some organizational challenges. Over 200,000 images have now been entered into a MYSQL database, indexed by location, date and time. The metadata for the images is stored in the database, but the images themselves are stored in a conventional file structure, indexed by date and time. PERL programs query the database and assemble the relevant images and PHP-enabled pages create HTML for the retrieved images, automatically.

We have been doing substantial work on "community" web sites using the PostNuke Content Management System (CMS). PostNuke sites store all the web site content in a database (again, a MySQL database) and include numerous tools for getting input from users. All administration is web-based. We have not adopted PostNuke for our primary LTER web site because of concerns about the difficulty retrieving archival content, however, it is perfect for special purpose sites where archival issues are not critical. Examples of such sites include a clearinghouse for information on wireless networks (http://www.vcrlter.virginia.edu/~jhp7e/wireless), the Mid-Atlantic Region Ecological Observatory (http://mareo.org) and even a revamped web site for the Long-Term Studies Section of the ESA (http://www.esa.org/longterm).

We have expanded our use of the Mapserver web-based mapping system, which has grown greatly in capabilities. It remains a "programming-based" system, with text configuration files for maps and hand-coded HTML templates, but if GIS shapefiles already exist, an Internet Map is only an hour or so away once you get the knack.

Our highest outstanding priority is to move from generating obsolete EML 1.4 metadata to EML 2.0 metadata. The biggest challenge has been finding a block of dedicated time to focus on writing the PERL DBI code to extract the metadata from our old MiniSQL database (which alas, lacks fully functional JDBC and ODBC drivers) to generate EML. Plans to generate EML 2.0 prior to this year's IM meeting were foiled when, during time scheduled for working on EML, we experienced a major disk failure that took down a critical system while a replacement disk was being sent from the manufacturer. Fortunately we were able to restore all critical files but the failure came at a bad time vis-a-vis EML production. Nonetheless, we still hope to have all metadata available as EML by September 1, 2004.

21-Sep-2004

Web design by Wade Sheldon, Georgia Coastal Ecosystems LTER

webmaster

2004 LTER Site Bytes

Andrews site byte 2004

Andrews Database and EML status

ClimDB/HydroDB and other DB's

Spatial Data

Project Management