Home

  Agenda

  Register

  Participants

  References

  Site Bytes

  LTER Home

  ASM 2003

2003 LTER Site Byte

LTER Site: North Temperate Lakes

Contributor: Benson, Balsiger, Chipman, and Hanson (Sep 17, 2003)

Site Byte:

Our site conducted a pilot project this summer to investigate (1) the logistics and usefulness of collecting very high temporal frequency data (1-2 minute) at instrumented buoys on a small bog lake that is one of our primary study lakes, (2) feature extraction in diel dissolved oxygen (DO) for the purpose of discriminating between signals due to biology and physical mixing, (3) detection of anomalous lake thermal profiles for the purpose of correlating them with weather events. We collaborated with Dr. Yu Hen Hu and Dr. Chin Wu from the College of Engineering at the University of Wisconsin-Madison to use both data mining and modeling to understand internal waves and to detect patterns in diel dissolved oxygen and anomalous variation in thermal structure. For the Trout Bog data, analyses were performed by converting time series to the frequency domain and then conducting a spectral analysis. We discovered that waves internal to Trout Bog propagate at a periodicity of about 40 minutes, suggesting that wind and solar radiation drive internal mixing. For the diel DO data, San Diego Supercomputer Center collaborators applied support vector machine algorithms to more than 140 days of data and achieved an accuracy of 85% in identifying biological signal in the data. For the detection of anomalous events, time series data were run through a low pass filter, and the subsequent trend data were removed from the time series. The remaining data, which contained high-frequency variations from the mean, were analyzed for outliers, which were discovered to occur approximately once every couple of weeks. The sensor data from some of the buoys were uploaded via spread-spectrum wireless transceivers to LoggerNet software (Campbell Scientific) residing on the file server at the field station. Current IM development is underway to make our information system for these instrumented buoys more streamlined and scalable. We plan to deploy additional buoys on our study lake during the coming year

We have upgraded our software that produces metadata in EML for online data in the NTL-LTER data catalog in order to be compliant with EML2. Now we are in the process of populating the metadata tables in our Oracle database for some of the information that was not previously part of the NTL metadata.

We have added some additional data sets to ClimDB that capture the historic meteorological data for our site that starts in 1869. Our site is also participating in HydroDB.

We are currently working on improving the management of spatial data at NTL, with a focus on four specific objectives: (1) implementing a new spatial database; (2) facilitating the integration of spatial and non-spatial data; (3) providing EML-compliant metadata for all spatial data; and (4) developing an Internet map server application for ready access to spatial data via a web browser. The architecture we are developing to meet these objectives is based on the combination of an Oracle relational database with an ArcSDE front-end to facilitate management and analysis of geographic data. While NTL has a long legacy of data management via an Oracle relational database, we are customizing a second Oracle database for efficient handling of spatial data. Working with a geographic information system (e.g., ArcGIS or ArcView), end-users are able to connect to both spatial and non-spatial databases, employ spatial analysis tools for queries and manipulation of NTL spatial data sets, and seamlessly link to non-spatial data, such as in-lake chemistry measurements, based on a common field such as lake ID.

We have begun populating the spatial database with a variety of NTL vector data sets and satellite images, to provide test data for the next steps in this “renovation” of our spatial data management infrastructure. We are currently generating XML metadata files as each spatial data set is added to the database, and will be making them EML-compliant. While we have not yet begun formal work on an ArcIMS application to permit web browser-based access to the spatial database, we expect to implement that during the coming year. The ArcIMS system, as envisioned, would allow users to readily query and map all publicly-accessible NTL spatial data, and would also provide a “spatial” interface for accessing non-spatial NTL data where such data are linked to specific lakes or other geographic features, e.g. the lake chemistry example cited above. It would also permit the provision of metadata via the new Metadata service included in ArcIMS v.4.0.1.

Last fall we had our first annual information management “retreat”. Information management staff, the site manager, the field station manager, GIS specialists, field technicians, the computer administrator, and the chemical technician all participated in at least a portion of this day long meeting held at our field station in northern Wisconsin. We all appreciated the opportunity to review the broad spectrum of IM related activities and to identify key activities/projects for NTL IM.

The NTL LTER site book, Dynamics of Lakes in the Landscape (Magnuson, Kratz, and Benson, eds.) is scheduled to go to the publisher in December 2003. There is a chapter on information management entitled “Breaking the Data Barrier: Research Facilitation through Information Management” by Benson, Hanson, Chipman, and Bowser.