The Gulf of the Farallones National Marine Sanctuary is one of the purest and most beautiful oceanic regions in North America.
No one but bona fide researchers can set foot on the pristine Farallones Islands, which are the centerpiece of the 1,200 square miles of open ocean, coastal waters, bays and estuaries protected in the sanctuary. Its waters extend from Sonoma County down to the southern end of San Mateo County.
Humpback whales and endangered blue whales feed and frolic here. The islands are breeding sites for myriad seabirds, while the waters are spawning grounds and nurseries for fish and shellfish.
At least 36 species of marine mammals have been observed here, while 25 endangered and threatened species live within its borders, according to the Farallones Marine Sanctuary Association.
Monitoring The Water
But this fragile ecosystem is contiguous to some of the most heavily populated and industrial areas of the nation, from the chip factories of Silicon Valley and industrial farms of the Central Valley to the gridlocked streets of San Francisco.
Toby Garfield spends his days watching over these waters. As an associate professor of oceanography for San Francisco State University (SFSU), he’s in charge of a system of floating sensors that constantly monitor the health of this stretch of the ocean. The sensor network was placed by students working at the Romberg Tiburon Center, an SFSU marine field station just north of San Francisco in Marin County.
Using sensors to monitor the waters off the Farallones Islands.
Source: Gulf of the Farallones National Marine Sanctuary
“Basic parameters for water quality are temperature and salinity,” Garfield said. “If salinity is low, we know the seawater is diluted by a freshwater discharge. From that we know the density, and then we know a little about which way it’s moving.”
Other sensors can measure the water’s nutrient load to tell whether drainage came from farms, as well as how much sediment is in the water. Too much sediment keeps sunlight from penetrating the depths, which kills the algae that sustains larger water creatures.
If conditions worsen, the marine ecologists and oceanographers at the center alert the regulatory agencies in charge of monitoring farm and industrial operations.
Data Acquisition on Land
But in the past, researchers had to take a boat out to each sensor bobbing in the water and manually siphon its data to a laptop, and then return to shore for uploading to the main computer system.
Now, thanks to a project done in partnership with Agilent Technologies
and Sun Microsystems
, researchers can obtain real-time access to the data without leaving shore.
The Networked Bay Environmental Assessment Monitoring Systems (NetBEAMS) broadcasts environmental data over inexpensive cell phones using Java open source software.
“The real value of NetBEAMS is you do away with the computer part of it and replace it with a telephone,” Garfield said. “The phone interrogates the hardware, collects the data and phones it off to a computer.”
The NetBEAMS application polls a network of sensors placed in different parts of the San Francisco Bay. The sensors provide information on water depth, temperature, salt content and algae growth, and then transmit the data via cell phone to a database that uploads the information to the Web.
San Francisco State has to constantly reinvent how it trains students for the job market, said Dragutin Petkovic, chair of the Computer Science Department.
“It used to be purely technical,” he said. “Learn Java programming five years ago, and you got a job paying $70,000 when you graduated.” Today, he said, hiring managers expect students to have a master’s degree and to have done innovative work with industry partners.
“The way we educate them is by doing projects like this,” Petkovic said. “It’s much better for students to be given a real problem and work with people outside their field. And we knew sensor arrays are a sexy thing. You always need to have some really good product to push the technology. So I spoke with Toby Garfield and said, ‘Let’s find a good application where we have a need, and the computer science students can work with them to solve the need.'”
Putting Java to Work
Meanwhile, Jim Wright, a senior solutions architect with Sun, had his own agenda. He’d been working on Industrial Java, looking at how to move it closer to the factory floor. He was focused on how to build Java networks that monitor and remotely control equipment, machines and the environment.
“Rather than work inside the Java community process, I wanted to see if we could put together software and get folks to use it,” Wright said. He identified existing standards that could be amplified, such as the IEEE 1451 standard for plug-and-play sensors and actuators, created a Java environment supporting it, and put it up on Java.Net to test the waters.
SFSU was a great match. “They had an institution, the Romberg Tiburon Center, were well-funded to do oceanographic research in San Francisco Bay, but they were having problems actually collecting the data,” Wright said.
The software development was done on the weekends, under extreme programming conditions, Wright said, with folks from Sun, Agilent and SFSU coding for 48 hours straight.
They designed general sensor node software that had self-descriptors for the sensor, actuator and data, as well as a networking protocol. Agilent provided the generic platform and the software that enabled the sensors to collect and forward the oceanographic data used by NetBEAMS.
Next was figuring out a simple, inexpensive way to transmit the data. There are dedicated networks used to transmit sensor data, but they’re expensive. The team hit on the idea of using consumer cell phones.
“I can go down to Cingular and get a family plan for a whole bunch of phones, and it’s quite inexpensive,” Wright said. “We scooted around the bay testing the signal quality. Cell coverage is superb along the shoreline and all across the bay.”
Cell phones typically have a serial datalink, as do sensors, so they simply loaded the sensor software into the cell phones, then hooked the phones to the sensor arrays mounted on buoys.
“The cell phones call home to upload the data, and we distribute it over the Net using conventional techniques,” Wright said.
Garfield said open source code and the OSS development model let the team do things cheaper and better.
“In oceanography, the market for any instrument is very small,” he said. “There are people who build proprietary software, and it sucks usually.”
The community development process will let oceanographers everywhere improve the application and add enhancements to suit their needs.
The software will find its way into commercial use, as well. Both Sun and Agilent are looking at using similar sensor networks to provide tsunami alerts and to maximize vineyard production.