Skip to content



California Coastline Data and Analysis
Accessible on the Web

Patrick Mantey, Dean of Engineering, Professor of Computer Engineering, UC Santa Cruz
Darrell Long, Associate Professor, Department of Computer Science, UC Santa Cruz
Alex Pang, Associate Professor, Department of Computer Science, UC Santa Cruz
Eric Rosen, Qualcomm, Inc.

Patrick Mantey began with a vision of a tool to help forecast environmental changes and make decisions on environmental issues. In particular, he was interested in a tool that could take advantage of the ocean and weather measurements collected by hundreds of instruments around Monterey Bay. The bay is a coastal region not only close to UC Santa Cruz where Mantey works, but also important for its rich biological life and its impact on the neighboring San Francisco Bay area. REINAS is the tool that Mantey and his colleagues built over the past six years to realize that vision.

"There are vast amounts of data in the description of a single environmental occurrence, be it a flood or a heat wave or another El Niño-type event," said Mantey, who is the dean of Engineering and Jack Baskin Professor of Computer Engineering at UC Santa Cruz. "From a system like REINAS [for Real-time Environmental Information Network and Analysis System], which captures and makes accessible real-time data from an array of instruments off the California coast, forecasters and policy makers can get the data they need, displayed in a visual format, and use that information to drive decision-making."

REINAS has found a following among a range of users, including surfers, sailors, and fishermen of the Monterey Bay area who use the system to inform their recreational plans. Although the database is sophisticated, a goal of the REINAS team has been to make the data it contains easily accessible so that this population of "casual users," as well as environmental scientists without a computing background, can use of the information.

"An important aspect of REINAS is hiding the complexity of the database from the user," said Darrell Long, associate professor of computer science at UC Santa Cruz and REINAS's lead for database development. "They shouldn't have to be a computer scientist to use the system." The REINAS project forms the core of a collaboration between NPACI's Earth Systems Science, Interaction Environments, and Data-intensive Computing thrust areas.



Real-Time Coastal Data AcquisitonFigure 1: Real-Time Coastal Data Acquisition

A model of upper atmosphere air temperature splits a column of air into layers, then colors each layer. The color range represents the discrepancy between the forecast model and the resampled data model. Red areas indicate minimal discrepancy, while white represents high discrepancy and may lead a scientist to refine the calculations used to generate a value for that point.


One prototype interface on the REINAS Web site allows the casual user to query a variety of the system's instruments and get back real-time or historical data. It is returned as a graph that maps changes in, for example, wind or air temperature or precipitation over the period of time the user indicates. "When this work is completed, the user will be able to control the colors, glyphs, averaging methods, and other parameters used to display their results," said Alex Pang, associate professor of computer science at UC Santa Cruz who led the creation of the REINAS visualization tools.

Environmental scientists, who use REINAS to observe, monitor, and analyze regional ocean and weather phenomena from their desktops, may make use of the suite of visualization tools attached to the system. Many of these are available on the Web through a software release that allows users to download applications to their own machine.

"Modeling software is an important component of REINAS," Mantey said. "It can, through interpolation, help environmental scientists fill in the spaces in forecast models due to gaps in instrumentation ranges. Comparing a forecast model to a model created from resampled data--that represents ground-truth at the same point in time or place predicted by the forecast--helps the scientist gauge the reliability of their forecast algorithms." (Figure 1)

Reliable forecasting has many applications for decision-makers. Recently, the team at UC Santa Cruz discussed using REINAS with the Federal Aviation Administration, who see predictive modeling based on data from the San Francisco Bay Area as a possible means of solving the age-old flight problem for the San Francisco airport--fog. Modeling patterns and running those measurements forward into the future may indicate when windows open in the fog, allowing proactive flight scheduling for inbound and outbound planes.

Earth Systems Science Prototype

Figure 2: Earth Systems Science Prototype

This image demonstrates a technique for using arrow glyphs to show uncertainty from simulations and observations in winds and currents. REINAS researchers at UC Santa Cruz, in collaboration with NPACI partners at UC San Diego's Scripps Institution of Oceanography and SDSC in the Interaction Environments and Data-intensive Computing thrust areas are developing a prototype system to allow Earth systems scientists to incorporate real-time field data into running simulations of global ocean-atmosphere systems and regional bay models, thereby improving the accuracy of simulation results.


"We are working with the Interaction Environments thrust area of NPACI on improving REINAS' end-user interfaces," Pang says. "The intention is to refine the tools and then make them available to users in other disciplines by incorporating the suite into other applications or amending them to different sources of data."

Long has also been working with the Data-intensive Computing thrust area to refine data manipulation, including the loading of data from distributed instruments to disparate databases. In the future, these disparate databases may include a database currently in demonstration mode at SDSC. "With NPACI, we're working to link the instrument and database nodes, as well as users running visualization applications, via a logical network connected by the Internet that looks seamless to the user," Long said.

The REINAS architecture can be divided into two major components: the suite of visualization tools on the one hand, and on the other, the network of instrument and database nodes that fuel the system. The emphasis on providing real-time data presents a challenge to constructing the instrument nodes. Measurements received by these geographically disparate nodes is transferred to a database node via the Internet. The instrument nodes, therefore, must be tightly coupled with the Internet's standard protocols. On the receiving end, the main REINAS database is currently located at UC Santa Cruz, but groundwork was recently laid for a second database to be connected to the system at SDSC, and a third may soon be located at the Oregon Graduate Institute.

REINAS instrumentation focuses on data from surface meteorological stations, Coastal Ocean Dynamics Application Radars, and vertical wind-profilers, which cumulatively require the system to manipulate scalar and vector data in point, linear, gridded, and higher-dimensional structures. Work is in progress to include other instrumentation, such as satellite imagery and remote camera instruments.

Managing input data falls to a stream-server that makes a disk copy at the same time as a loader interprets the measurement data and inserts it as appropriate into the database. "About 90 percent of database activity is devoted to writing new data into the database," Long says. "Separate inserts are received from several hundred sensors every few seconds. Buffering the data through the disk-based transaction log mediates the stress this places on the system and insures that REINAS is available to users, even in the midst of peak activity." A violent storm, for example, might generate rapid inserts that a scientist would be particularly interested in, so precautions are taken to protect such valuable observational time.

However, just in case a scientist missed such a storm, REINAS supports retrospective study through rich metadata describing the sensor-produced data pedigree. Metadata tags include geographic location, data type, source type, system, and human activity associated with the data. "In this way," Mantey says, "the scientist can assess how valuable a segment of data is and whether it fuses with other data, dependent upon the conditions under which it was received and handled."

REINAS has undergone many changes since the project began in 1992 with funding from the Office of Naval Research, and Mantey expects many improvements and changes in the future. "Through our interaction with NPACI," he says, "we hope to create a seamless picture of the entire western coast of the United States, built upon distributed data sets located up and down the coast and available to all types of users over the Internet through easily accessible visualizations." --AF