Exploring New Spatial and Temporal Scales in Ecology

Using Wireless Sensor Arrays

Organizer: John Porter

20 September 2003

New avenues of ecological research are opened by recent developments in wireless networking. Wireless sensor networks have the potential to collect temporally-intensive and spatially-distributed data, thus allowing ecologists to “observe the unobservable”. Such networks also can expand the types of data available to ecologists (including image, video and audio). The goals of this workshop were: 1) To identify scientific research opportunities and areas enabled and opened up by wireless sensor networks; and 2) To exchange information on capabilities, techniques and technologies, and experiences for wireless sensor networks.

The workshop was well-attended with over 38 attendees, eight of which were ILTER participants from outside the U.S.

This report includes a synopsis of the presentations, panel presentations and discussions which ranged from existing, operational, sensor arrays with fewer than 20 sensors to research on advanced arrays that might encompass hundreds to thousands of individual sensors. The presentations document a range of approaches to wireless sensor networks for the study of ecological phenomena using both serial and Internet Protocol (IP)-based systems. They also document the challenges, from power supplies to new approaches to network topologies and the new information management techniques that will be required to “drink from the fire hose” of data that these arrays can generate.

Of special importance was the discussion of “next steps”. These included specific suggestions in the areas of:

• Team building – providing expertise or test equipment for developing wireless networks at more sites,
• Training – developing opportunities for training of ecological researchers in the uses and implementation of wireless sensor arrays,
• Community-based Resources – development of email and web-based resources for exchanges of information about what works, and what doesn’t,
• Publications – sharing of information on the development of wireless sensor arrays for ecological research with the broader scientific community, and
• Partnerships – developing working relationships with the engineering and computer science communities.

Workshop organizers and participants are currently engaged in pursuing these “next steps.” The initial emphasis of our efforts is on “Community-based Resources” through the development of a web site that builds on the presentations at the workshop, and publications, seeking to get the word out that wireless sensor arrays are not just in our future, they are here now!

Synopsis of the Workshop

The workshop was structured along the following lines:

Opening Remarks - John Porter, University of Virginia, and Virginia Coastal Reserve

Keynote Presentations

· Ecology at multiple scales – Tim Kratz, University of Wisconsin, North Temperate Lakes

· Technological Advances in Wireless Sensor Arrays - Bill Michener, University of New Mexico, LTER Network Office

Panel Presentations (Introduced and Moderated – Peter Arzberger, University of California San Diego)

· Acoustic Sensing – Stuart Gage, Michigan State University, Kellogg Biological Station

· Buoy Networks – Paul Hanson, University of Wisconsin, North Temperate Lakes

· Eco-Grid – Fang-Pang Lin & Hen-biau King, National Center for High Performance Computing, Taiwan Forest Research Institute and Taiwan Ecological Research Network

· Barrier Island Wireless Network – John Porter, University of Virginia and Virginia Coastal Reserve

Note:

· The presentations can be seen at the following websites:

o VRC

o PRAGMA

o LTERNET (technology page)

· In addition, the contents of this workshop are being prepared for an article in Biosciences

Thus, the contents of the presentations is not repeated or summarized here.

Discussion:

Open discussion followed these presentations, and took place among the more than 40 participants, from different backgrounds, more than 30 institutions (academic, industry, government), 5 countries (USA, Canada, France, Taiwan, Korea), attended and participated in this workshop.

The discussion focus on Opportunities, Challenges and Actions:

Opportunities: What are some of the opportunities for new science that can be afforded by sensors and sensor networks?

Serendipity: Finding patterns that “usual” means of capturing data could not collect. For example,

• Leaving on a monitor all night will capture the event of an unusual sound at 2 am. (Gage)
• Capturing data during a hurricane from a sensor (Porter)
• Seeing wildlife that a human being would likely make too nervous to capture (Porter)
• Detecting new patterns in the relationship between temperature gradients in lake, time of year (Kratz)

In short, the panelists reinforced the statement from the Sensor Workshop, that sensors and sensor networks could “Reveal previously unobservable phenomena”

In addition, other opportunities were present with implementing sensor networks:

• Working with engineers to design the networks more efficiently.(Hanson)
• Furthermore, working with other experts is important so as NOT to reinvent the wheel.
• With a network in place, one has the opportunity to “stack” different technologies close to each other. (Porter)
• In a recent all user meeting of the High Performance Wireless Research and Education Network (supported by NSF, PI. Hans-Werner Braun), the infrastructure provided by the network allowed for more applications than originally envisions (e.g. education in San Diego County of more than 15 of the native American tribes, and for NEW collaborations among users sharing the network (earthquake scientist and earthquake engineers).
• Comment from audience: This sharing of infrastructure can help amortize the cost over time, and can help spread the maintenance cost over many users, making marginal costs lower.
• Comment regarding infrastructure: Both the sensor net workshop report and the Cyberinfrastructure for Biology stressed the importance of not only creating such infrastructure, but also ensuring its long-term existence. Infrastructure requires long-term commitment to maintain and improve.

With sensor networks there is the opportunity for the public to see what is happening with the environment, not just in ones own country, but also in other countries (ES Kim). Further, these networks are useful for bringing new educational experiences to students of all ages.

Challenges: What are the challenges to using sensor networks?

Sampling:

The issue of sampling (rate, location) was raised. Issues of sampling versus power consumption and bandwidth were discussed. Temperature readings do not require the bandwidth (or data storage) that audio capture (such as by S. Gage). Also, power consumption is greater for the larger volume of data that needs to be transmitted!

There are many research issues associated with sampling, such as how often to capture the phenomena (Hanson). Also, the Center for Embedded Network Sensing (http://www.cens.ucla.edu) is doing a lot of research in the area of sampling.

Other issues:

• If 802.11b does not work well in forests, what are the alternatives for forest research. [J. Zimmerman]

• Comment was that people should go ahead and plan to use sensors, given that technology is changing. Furthermore, there are other technologies (licenced band) that do work in forest areas, but a policy change needs to be made by FCC (and some pressure is being brought to bear on this issue to make communications easier in rural American).

• What is being done for sensors to track animals (given the range of current technology is short)? [B. Parmeter]

• There are other ways of handling some of these issues.

• Have people considered what to do about metadata from streaming sensors, when the senor itself is “smart” (i.e. can change the sampling rate based on external stimulous). [G. Henebry]

• Since so much of ecology often involves counting, are there sensors that have been devised to do that? [J. Goldman]
• Some work has been done to count zooplankton (Michener). Also, there are ways of counting Gypsy Moths (S. Gage)
• There is a great deal of technology being developed by the military (bioterrorism – attack on the food chain).
• There was a statement about “thousands” of sensors.
• There was a reinforcement of the challenge of the lack of common interface to sensors.

Possible Actions: What would be useful for the community to facilitate development of wireless sensor networks?

During the workshop, many of the participants indicated that they were in the process of establishing wireless sensor networks.

BUILD TEAMS: The example of Dave Hughes was mentioned, both in terms of training as well as in terms of hands-on, on-sight consulting. Having a team of individuals that could bring various equipment to sites, and help with experimenting with wireless as well as sensors could be a great advantage to sites trying to ramp up. [Example was given of Dave’s road show of various types of radios that could be brought to a site. Also, J.Porter mentioned a box of parts always ready to go with various pieces of equipment.]

Conduct Training:

- Hand- on training was very valuable. Hughes example, and LTER Network office example.

- Point was made that each installation site is different, thus having experts be able to go around was critical.

o LNO if not swat team, would be good to have someone act as consultant that we could go to

Create Community Based Resources:

- E-mail alias for group of individuals interested in working together on experiences.

- Web site with information

o Caveat here is that since each site is different so this is not a replacement for on-site consultation.

o Also, this site would need to be maintained on a regular and frequent basis, as the technology is changing rapidly.

§ Need to reduce costs by collectively providing taxonomy on types of sensors needed

o Create Record of lessons learned on web: Some of this already http://oldcolo.com. Also, the web has other sites (see Michener talk).

Note: When NEON comes, a coordinating center could fulfill some of this need of keeping and sharing expertise and experiences. Need to involve engineers

Communicate via Publications (for disseminating information but also for consumption of new technologies)

- Create article for BioScience, highlighting the diverse examples that are currently using wireless, lessons learned, and challenges, and common resources.

- IEEE Journals

- Trade journals – telemetry

- Sensor magazine (http://www.sensorsmag.com/)

Create Partnerships:

- Possible partnership with CENS (http://www.cens.ucla.edu), and possible supplement to have LTER as a key outreach.

- Work with engineering and computer science communities. Several examples of partnerships were cited:

o At North Temperate Lakes, there is a nice interaction between ecologies and engineers around issues of energy consumption, sampling scheme and automated discovery of unusual events.

o In Fushan Park there is a collaboration between grid and wireless experts putting in a wireless sensors and the ecologist.

Other Needs:

- The animal as the sensor – sensing physiology etc. via telemetry

- Need to develop list of research (ecology) challenges

- Not just radio – also acoustic modems for underwater