A Sustainable Infrastructure for Real-Time and Long-Term Environmental Monitoring

The Santa Rosa Island is part of the Channel Islands of California. It is located thirty miles south west of the city of Santa Barbara in Santa Barbara County in Southern Californi, fifty miles west from the campus of the California State University Channel Islands (CI) in Camarillo, Ventura County, and a hundred miles west of Los Angeles.bieszczad_wmnc2015_submission_fig1

The Channel Islands are home to many unique and understudied communities of plants and animals. For example, there are some plant species on Santa Rosa Island that only exist there. There is the endangered Island Fox whose behavior and migratory pattern data have only recently started to be collected by the National Park Service (NPS). There are three local marine preserves around the island that provide insight about marine biodiversity away from human contact [1].

California State University Channel Islands has established the Santa Rosa Island Research Station (SRIRS) to provide and support research, education, and outreach activities. Monitoring the environment and wildlife is part of the efforts to restore — and then preserve — the island that was destroyed by years of uncontrolled farming exploitation. When the island was turned in completely to the National Park Service in 2011, its native environment was dramatically depleted [1]. The goal had been set to remove the numerous non-native plants and animals that were endangering the very survival of the native fauna and flora that were threatened with extinction. The CI’s SRIRS actively participate in these efforts.


Environmental monitoring is paramount for any restoration and preservation activity. The NPS is a federal institution that has a mandate to conduct — and sponsor — surveys on the island which is part of a national park. All data collection is currently done by hand: human surveyors follow appropriate protocols that define the routines governing such activities. That method of data collection has a lot of shortcomings including high costs, limited coverage capabilities, error- proneness, and excessive time requirements. With the very limited federal funding there is a need for a more cost-effective approach.

The CI Rainbow Research Project had been established within the Computer Science Program of the California State University Channel Islands, and in collaboration with the Santa Rosa Island Research Station, to overcome these shortcomings. The project goal is to design — and ultimately deploy on the island — an infrastructure that will support real-time and long- term automated collection of environmental data. In other words, the goal is to allow any researcher who wants to conduct studies on the island (e.g., animal tracking, climate changes, hydrology, marine life, etc.) to conveniently collect data using the CI Rainbow infrastructure, and then to access and analyze the data from the CI Rainbow Data Collection and Analysis Center.

The following objectives had been specified to attain the overall goal of the research efforts of the project:

  • design, develop, and test hardware and software to record data remotely,
  • ensure that the infrastructure is able to accommodate a variety of sensors, data, and analytical tools,
  • ensure infrastructure longevity by employing flexible and expandable architecture,
  • ensure data longevity using extensible database storage solutions,
  • create an easy-to-access intuitive data analysis and visualization center,
  • create a comprehensive infrastructure Operation, Administration, Management, and Provisioning (OAM&P) facility,
  • ensure that data collected is accurate and secure, and
  • ensure that the remote sensor and network power needs are sustainable off the electric grid.

In an excellent and very comprehensive survey [2], a variety of aspects of wireless sensor networks are reviewed. Using the nomenclature from that survey, the CI Rainbow terrestrial infrastructure is not a typical wireless sensor network (WSN), since [2] points out that typical WSNs have no, or little infrastructure. It is also a well structured WSN at least as far as the core transport network is concerned; the most remote parts of the infrastructure allow for certain degree of ad hoc organization. Although our application is similar to some reviewed in [2], it seems to be unique as a non-specific, general, and open platform for environmental monitoring in harsh terrain. It also incorporates data analysis and visualization tools that normally are not considered part of a WSN.

Another interesting survey [3] reviews a number of Environmental Sensor Network (ESN) that not necessarily are WSNs. That survey is geared towards Earth Sciences.

The authors of [4] attempt to evaluate power demands for sustainable WSNs. Our approach to this problem has more brute force nature; we aim at satisfying average requirements, and augment that with implementing a power-failure recovery mechanism.

Consequently, most of the solutions presented here are our own inventions rather than copies. Many comments in [2] and [3] concern the sensors themselves, and these play a secondary role in this report as we firmly focus on the infrastructure.

Exploratory Infrastructure

Due to its remoteness and lack of utilities as well as its protective status, developing such a complex system on the Santa Rosa Island is not feasible. Yet due to the same reasons, the ultimate deployment is realizable only if the solutions being put in place are well-tested and with a minimal exposure to errors. That paradox yields a need to develop and test a prototype of the infrastructure in a more accessible and convenient to develop and test environment. This environment needs characterization by similar climate, topology, and fauna and flora as those on the Santa Rosa Island.

Fortunately, quite recently CI has acquired a nearby wildlife park at the foothills of Santa Monica Mountains that is only fifty miles away from the island and has analogous natural conditions. In particular, the park is similar to the Santa Rosa Island in the following respects:

  • similar topology that includes mountains and valleys,
  • environmental conditions; e.g., the park is subjected to draught most of the year, and to frequent marine layer,
  • similar plant diversity,
  • similar wildlife,
  • while much more accessible than the island, the park is still remote enough to test various aspect of the infrastructure; for example, the lack of utilities forces the use of renewable energy,
  • due to the topology, the line-of-sight between the park and the campus that hosts the CI Rainbow data center is obfuscated by hills; that enforces the need for a sequence of relay points that must pass data.

bieszczad_wmnc2015_submission_fig4Therefore, we elected to conduct our research and development efforts — including initial deployment and testing — in the the rugged CI Park location. The exploratory CI Rainbow infrastructure stretches from the data center located on the CI campus to sensors in the CI Park. This particular deployment of the infrastructure supports three sensory wireless clouds that cover most of the park. Each cloud is capable of supporting numerous heterogenous sensors. The data is sent over short and long haul links to the data center. The communication is bidirectional, so that commands can be dispatched from the control center at the campus to any component of the infrastructure.

Infrastructure Architecture

The CI Rainbow infrastructure is a complex system consisting of numerous software and hardware components that constitute the three main parts:

  • exploratory wireless network,
  • application and data center, and
  • sensory node platform.

The CI Rainbow network carries the data from sensors to the data center and commands from users and administrators in the opposite direction. The furthest network access points support sensory clouds.

The CI Rainbow Data Collection and Analysis Center is the heart of the system that resides in the university data center located on the CI campus. The servers in the center support all functionality of the system. Firstly, users can provision sensors tying specific sensors to the database entries, so when these sensors are deployed in the field the data can start to be collected automatically. It is worth noting that these are users, and not administrators, who manage sensors including their provisioning and deployment. Using another Web application, users can access, analyze, and visualize the sensory data collected in the database. System administrators can also use the backend to manage every aspect of the infrastructure.

The sensory node platform is based on a general purpose microcomputer capable of supporting a variety of sensors. When deployed, the data from the sensors is packaged on the platform and shipped over the network, starting with the cloud access point, and ending at the data center.

Support for Professional Sensory Stations

The latest activity in the project is interfacing with professional grade sensors that we were lucky to acquire: WeatherHawk 610 for weather monitoring and EXO2 Water Sonde for evaluating water quality.

WeatherHawk 610 Weather Station

Thi is a serial (RS232) connection weather station. It measures:

  • wind speed & wind direction with an acoustic array,
  • precipitation (rain and snow) with an upward looking radar,
  • barometric pressure,
  • air temperature,
  • relative humidity, and
  • solar radiation.







EXO2 Water Sonde

This is a multi-parameter 6-port water quality sonde with anti-fouling wiper for oceanographic, estuarine, or surface water applications.

We have sensors ports (sensors) for measuring:

  • Wiped Conductivity / Temperature For long-term monitoring
  • Total Algae
  • Turbidity
  • pH/ORP
  • Dissolved Oxygen Optical

To connect the sonde to a PC (RaspberryPI) we also need:

The EXO DCP Signal Output Adapter 2.0 and 100 meters-long EXO Flying Lead Cable allow for remote underwater data collection.


  1. “Channel Islands Interpretive Guide Santa Rosa Island”, National Park Service, http://www.nps.gov/chis/planyourvisit/upload/Santa-Rosa-Island-Interpretive-Guide-2014-low-res.pdf
  2. J. K. Harta and K. Martinez, “Environmental Sensor Networks: A revolution in the earth system science?”, in Earth-Science Reviews Volume 78, Issues 3–4, October 2006, Pages 177–191.
  3. J. Yick , B. Mukherjee , and D. Ghosal, “Wireless sensor network survey”, in Computer Networks Volume 52, Issue 12, 22 August 2008, Pages 2292–2330.
  4. G. Anastasia, M. Conti, M. Di Francesco, and A. Passarella, “Energy conservation in wireless sensor networks: A survey”, in Ad Hoc Networks, Volume 7, Issue 3, May 2009, Pages 537–568.