Sensor Site and Platform Selection
back to EnviroSensing Cluster main page
The primary editors for this page may be contacted for questions, comments, or help with content additions.
Scotty Strachan - Department of Geography, University of Nevada, Reno - scotty at dayhike.net
Adam Kennedy - Andrews Forest LTER - adam.kennedy at oregonstate.edu
Selection of exactly where and how to acquire data via in-situ sensing efforts is a crucial point in the science process where environmental research is concerned. Decisions made when choosing sites, sensor packages, and support infrastructure in turn place boundaries on what the final science deliverables can be. Data types, quantity, and quality are more or less set in stone during this process. Initial costs, timeframes, and sustainability are also determined by these choices. Selections need to be made based on the desired science products, but also in consideration of a wide array of variables ranging from land ownership and access to equipment budget and maintenance capability to previous research and construction/deconstruction logistics. Setting up terrestrial sensing systems is a major infrastructure commitment with budgetary and environmental concerns, and every effort towards maintaining a robust, low-impact, and long-term data stream should be made. Because each region possesses unique geography, there is no “one size fits all” solution. Instead, a series of considered decisions need to be made, with the goals and capabilities of the research team clearly defined.
Identifying both the deployment strategy (site, process) as well as the physical hardware (sensor platforms and support infrastructure) for environmental sensing is usually a daunting task. Very often the decision tree for choosing exact locations and deployment schemes is dependent on interacting factors (such as permitting/geography/access; Fig. 2). There is also a vast array of possible sensor/hardware packages available for a multitude of science applications.
It is critical that Principal Investigators (PI’s), logistical techs, and sensor specialists work together to develop specific deployment plans and alternatives, ideally in the pre-proposal stage. Planning topics must by needs include science objectives, operating budgets, proposed locations, seasonal weather patterns, power sources, communications options, land ownership, distance from managing institutions, available personnel/expertise, and potential expansion/future-proofing. All of these categories are equally critical for discussion as proposed instrumentation projects move towards implementation.
Site visits, permit/agreement negotiations, equipment specifications, and deployment timeline need to be initiated concurrently and naturally are very interdependent (Fig. 2). The P.I., together with the technical personnel, should identify likely site candidates for the deployment based on local topography, permit/agreement availability, logistical access, and availability of services (such as power and communications). Portions of the plan (such as some purchasing decisions) should remain flexible until the precise sites, permits/agreements, and data flow plan have been positively determined.
Environmental conditions have considerable bearing on science application, platform design, construction logistics, access restrictions, equipment reliability, and maintenance cost/longevity. Conditions for in situ sensing can vary tremendously from region to region, and so siting and equipment must be selected accordingly on a case-by-case basis.
- Local topographic variables include: northern versus southern exposure, which can affect snow persistence and hours of direct sunlight; and valley/sink versus ridgeline settings, which can affect what type of daily temperature cycles and winds are encountered. The differences in airflow, wind exposure, cold sinks, snow drifts, sky exposure for solar panels, and possible radio/communications pathways are all important variables when selecting a site and what type of equipment will be deployed.
- Dominant vegetation conditions and potential long-term growth can alter sensor readings via shading effects, affecting temperature, radiation, and snow-related measurements. Radio communications are also affected by vegetation, with most microwave frequencies used by high-speed data radios being strongly attenuated by trees and brush. Vegetation can also be a long-term hazard in the forms of fire fuels and deadfalls.
- Visibility and the visual impact of deployments should be considered for both security and aesthetic considerations. Sometimes reduction of visual impact is required by landowners, but in general it is simply good practice. Metal structures can be camouflaged with paint to reduce visibility, structure heights may be reduced to blend with vegetation, and ground disturbance can be kept to a minimum to avoid biasing certain types of measurements and erosion.
- Dominant weather conditions determine what levels of seasonal access are available, what structural designs should be used, and what sort of equipment should be purchased. Extreme temperatures, tropical storms, lightning, snow depth, riming/ice, UV exposure, high humidity, wind speeds, salt water exposure, flooding, and stream depth variation are all examples of conditions which will influence design and deployment plans.
- Wildlife can provide hazard considerations or be affected by proposed deployments. Bird perching and flight paths, cattle, soil invertebrates, rodents, and large mammals can all disturb or be affected by a monitoring installation. Landowners will have either regulation or preferences concerning these factors, and proactive steps on the part of the science team to minimize opportunity for animal conflicts are suggested.
- Sensitivity to local political and social issues also need to be considered, as in the end, these data are really supposed to serve the local populations as well as the scientist and funding agencies.
- Site security is of course a primary concern when engineering a deployment. Hazards to sensors include natural disturbance/disasters such as wildfire, flooding, extreme winds, and mass wasting. Planners should be aware of all these possibilities and at least examine the likelihoods of these event at sites which have been evaluated from the scientific point of view. Human theft/vandalism is another potential cause of sensor disturbance or failure. While remote deployments are nearly impossible to secure physically, measures such as camouflaging, informative signs, fencing, and lockboxes may be employed to mitigate hostile or irresponsible passers-by.
Locations for in-situ sensing must be accessed for survey, construction, and maintenance over the life of the project. Seasonal conditions, roads, and topography determine what types of access may be used during different times of the year. Categorical considerations include:
- Vehicular access. Commercial vehicle/equipment, 2WD auto, 4x4 truck, ATV, snow machine, boat, helicopter.
- Non-motorized access. Hiking, skiing, pack animals, snowshoeing.
- Access improvements. Road building, trail building, trail demarcation, safety rails, harness anchor points.
- Seasonal access. Define access by Spring/Summer/Fall/Winter seasons. This is directly related to local weather/topographical conditions.
- Construction access. Heavy equipment, special equipment, heavy loads, and heavy foot traffic are all likely possibilities depending on monitoring design.
- Minimal impact considerations. Can traffic/access be directed in a way to minimize environmental impact (e.g. erosion, vegetation)? Solutions include boardwalks, bridges, raised steps, delineated pathways.
Science platform selection
Once the science objectives and siting conditions are known, an itemized list of sensor and support system platforms/hardware may be assembled that best fits the application and budget. Primary considerations include reliability, comparability with other similar field systems, technological (e.g. programming) requirements, budget, and system flexibility (upgrades, expansion, telemetry options, etc.). Accuracy, precision, and expected period of use prior to calibration or replacement may also be a consideration. In some fields of study, there are only one or two alternatives to choose from in terms of scientific instrumentation, whereas in others there can be an impossible array of choices. Often options can be narrowed by researching what equipment/standards are used by existing installations to which comparability is desired. Once a science data acquisition platform and sensor array is chosen, remaining support systems are then designed around this core equipment.
Support system specification
The subsystems of structure, electrical power supply, and data communications are all designed to best support projected science platforms on the sites chosen. While some vendors offer “all-in-one” packages supplied with standard instrumentation, it behooves the research team to assess whether these solutions are adequate for their chosen site and objective. Quite often multiple science packages are deployed and require design of “purpose-built” systems sourced from multiple vendors.
- Structures – these are the building-blocks of any remote data acquisition site, including tripods, towers, poles, buoys, solar panel racks, storage boxes, fencing, concrete pads, and the like. Quite often a single tripod or tower does not have adequate space or structural integrity to support all of the sensors, antennas, solar panels, batteries, and other items, so a typical site design incorporates multiple structural components.
- Power generation and storage – for sustaining long-term reliable data streams, power independence is critical. Stations should be capable of generating and storing their own power locally, as well as taking advantage of any grid or other power available within budget and design restrictions. Because the majority of related electronics are ultimately powered by DC voltage, having a generation system and DC battery bank for every site (and sometimes discrete subsystems) is recommended to minimize data gaps. If available, tying grid power as an alternate DC charge source is preferred. Independent generation sources are most commonly PV arrays (solar), wind, or water turbines. For reasons of cost, reliability, and maintenance issues, PV (solar) is recommended as the primary on-site generation source if environmental conditions allow. Incorporating simplicity, redundancy, and excess capacity is important for long-term reliability.
- Data communications – Use of real-time data transmission (in addition to local storage capacity) is strongly desirable to enable remote system health monitoring, troubleshooting, and minimize data gaps. This is usually performed using in situ radio communications (whether vendor-specific or building a general-purpose field IP network. Communications systems need to be robust, secure, and have low power requirements. (Refer to “Data Acquisition and Transmission” Best Practices for further detail.)
- Construction details – When selecting and designing the science packages and support systems, several details of construction should be considered when providing specifications for and purchasing hardware. Wiring management should involve protective conduit and pull/storage enclosures to avoid exposure to damage and seasonal degradation. Wire lengths, enclosure sizes, and mounting locations should be planned for accordingly. Anchoring for support structure should be designed to withstand worst-case weather/environmental conditions. Use of corrosion-resistant metals for structure and hardware such as galvanized steel and aluminum will greatly reduce failure or ongoing maintenance problems.
The site layout at first seems trivial, but in fact can be very important where interaction of the various subsystems are concerned, influencing reliability and data quality. Deployment design of science sensor packages to optimize measurement quality should be determined first, followed by placement of support systems and additional structure. Solar arrays need to be optimized for sun exposure, minimal shading, and snow-shed angles. The impacts of site structure on measurements (such as wind eddies, incoming/outgoing radiation, camera viewsheds, or precipitation catch zones) need to be considered as well as impacts to visibility if in a frequently-visited area. Power and data cable runs should be protected and kept as short as possible, voltage drop over long runs can be a consideration in layout and design. Stipulations in site permits may be drivers of site layout and construction. Once the site layout is designed and mapped, specification of construction materials, sensor cables, and other supplies may be optimized.
Selection of deployment sites, science platforms, and support systems are interacting processes which can require some iteration before arriving at the final determinations. Unless the science objective is extremely narrow or exceptional in nature, it is unlikely that any one of these decisions can be made in a vacuum without considering the others. With this in mind, the following overarching recommendations can be emphasized:
- P.I. consultation with system/hardware/construction specialists while in the proposal phase will minimize budget surprises or platform compromise later in the process.
- Data quality and longevity should be the ultimate goals when designing the deployment. Making choices for more robust and widely-used core systems and sensors will ensure that data comparability is maximized and hardware problems corrupting data or creating gaps are minimized. Purchase of reliable and known equipment is not as expensive as repairing/replacing equipment halfway through the study or losing valuable data.
- When data quality and continuity is paramount, use of replicate sensors or stations may be required.
- Planning for real-time connectivity is crucial for reducing field maintenance time and data gaps.
- Optimal siting for science objectives can often be impeded by permit requirements and landowner preferences, so starting the conversation with the landowners early in the process may improve the chance of getting the locations/deployment types that are desired.
- Standardizing sensor and support hardware, software/programming, and structural designs across multiple sites minimizes maintenance issues as well as construction costs and design time.
- Assessing access capabilities to the sites will allow for planning of emergency maintenance access, procedures, and costs.
- Overbuilding structure, power capacity, and site infrastructure (cabling, networking, etc.) will prevent problems in the case of unforeseen events or site expansion.
Common Points of Failure
- Power problems are one of the most frequent causes of total system failure. Battery fatigue, loose connections, and electrical shorts need to be anticipated and prevented where possible. Power systems need to be protected, over-engineered, and replicated wherever possible.
- Temperature extremes of heat or cold can cause electronic or mechanical failure of individual sensors and systems. Insulating enclosures, ventilating enclosures (active or passive), and placement of equipment in sheltered zones can help alleviate these problems.
- Humidity and condensation can be a serious issue for electronics longevity and circuit performance (including accuracy). In zones of high average humidity, sealing enclosures and providing some means of reducing humidity (i.e. desiccant packets) is desirable.
- Sensors can be disrupted by wildlife actions and fouling. Hardening of sensor systems (e.g., armoring cables, fences) can help with some problems. Near-real time data feeds allow rapid detections of problems that do occur.
- Lightning hits or near-misses is a common problem with exposed or mountainous sites. Extensive grounding (exposed copper wire network) and use of surge protection throughout the power system and at ends of long power and data cable runs will compartmentalize the site electrically and protect as many components as possible.
- Lack of data storage replication can cause loss of data - by incorporating high capacity storage on-site (datalogger) as well as off-site (database), this problem can be mitigated.
- Personnel turnover coupled with lack of process and hardware documentation can lead to data discontinuity or equipment failure (see Sensor Management and Tracking for additional details).
- NevCAN Transects or Walker Basin (Scotty) --- To be completed, will include a station design and systems, maintenance/access plan, data flow, and some photos/diagrams.
- Andrews Research Sites (Adam) --- To be completed
- Sevilleta - Renee to complete with multiple case study examples