General Guidelines and Best Practices for Starting Your Own Environmental Monitoring Study
Table of Contents
Determining the Question at Hand
The first step in environmental study design is determining exactly what questions you are seeking to address. In the case of air pollution, the design of a study will be very different if you are trying to assess emissions versus exposures (e.g. investigating if an industrial complex is emitting more pollutants than it is permitted to do, versus assessing human exposure to pollutants in a residential neighborhood), and it will also be quite different if you are trying to assess chronic over-emissions or exposures versus episodic high-emissions activities. In order to design an effective environmental study, define the question as clearly as possible. This document is geared toward environmental study design in cases of air quality concerns, but the concepts can be applied to terrestrial and water quality issues too. Aspects of the study to determine include (but are not limited to):
Is your approach going to be assessing emissions or exposures (aka sources or receptors)?
Are you interested in chronic emissions or exposures, or in episodic, acute emissions or exposures?
Are there multiple potential sources of emissions, or a single emissions source, and would you want or need to distinguish sources?
If you are interested in exposures, are there multiple populations with different exposures or vulnerabilities (e.g. children at a school near an industrial facility, workers at a facility, elderly people in a nearby residential neighborhood, etc)?
Do you know how variable the emissions might be, or how variable the environmental factors might be (e.g. specific activities associated with high emissions, or highly variable wind speeds and directions)?
In addition to these key features moulding a research question, other important factors, which are discussed in more detail in the following sections of this document, include:
What are your human, financial, and physical resources?
What is the geographic scope of your study?
What is your capacity for advocacy, and what kinds of relationships might need to be built?
When designing your environmental monitoring study, it is important to outline a study that will be achievable with your current capacities, or you will need to include capacity-building into the study process. Types of resources that are important to evaluate include:
Human resources: Who are the people who are going to conduct this study? How many people are interested, and how much time can they commit? Can they commit time on a daily basis, or just occasionally? Do they have access to vehicles or other potentially necessary transportation? Do they have health concerns that may be exacerbated if exposed during sample collection? Do you have people who could fulfill different roles, such as people who are detail-oriented, good overall coordinators, charismatic, etc?
Financial resources: How much money do you have to purchase equipment or potential laboratory analyses? How many monitoring stations can you afford to set up and use? How many discrete samples can you analyze? Can you purchase equipment to provide ancillary (yet important) data, such as wind speed and direction or water flux?
Physical resources: Do you have sample collection or monitoring equipment already? Can you build structures to assist in sampling (e.g. platforms or external housing for instruments)? Do you have means of transportation if that is necessary? Do you have access to the locations that you want to monitor? Do you have a map of the area, or could your community create an aerial map that would be useful?
Advocacy resources: Do you have relationships with persons or entities that could assist you in advocating for change? Would it be beneficial to develop relationships with local environmental enforcement agencies or media outlets? Would members of local government participate in this study?
Do you have other resources that haven’t been mentioned but could be useful in designing, conducting, or communicating your study? Understanding the resources you have available to you will help you design a study that is within your capacity to conduct, and help you effectively utilize all the resources you do have available.
The geographic layout of the area you want to monitor will dictate much of your study design. To start, consider how large the geographic area is, and what sort of access there is to specific areas of interest. If you are interested in monitoring emissions from a certain industrial plant, how close to that plant can you get? Would you be able to monitor from at least four directions around the plant, or are there hurdles to navigate like private land and waterways? Are there landforms that might impact airflow, such as hills or valleys? Is there a dominant wind direction, or does it change frequently? Have you observed microclimates, such as areas that remain damp or where fog accumulates, which could impact atmospheric transport of aerosols etc? For studies concerning water contamination, are there confluences of rivers or streams? Are there artesian springs where groundwater emerges at the surface?
Map it Out
Aerial maps can be very helpful in determining where to monitor. Your team can create your own aerial map using Public Lab’s aerial mapping kit [link], or use another map available to you. Maps can be useful to draw out wind directions, identify roads and roadway access (or blockage) to points of interest, identify streams and stream confluences, and to get an overall sense of scope for the project. Topographic maps can also be useful in study design, particularly in areas with complex topography, such as hilly regions.
Upwind and Downwind
When gathering data, it is important to have comparison data of samples that you would expect to be clean compared to the areas of concern. This is especially important when assessing whether or not a specific source (such as an industrial plant) is the source of a contaminant, and is also important when assessing whether a certain community has higher exposure to a contaminant than other communities do.
For air quality studies, the key locations to collect samples are upwind and downwind of an expected emissions source. Since meteorological trajectories are difficult to track or predict, simultaneous data collection from at least two directions (upwind and downwind), and more favorably 4-8 directions, are important. The distance from the source where you would monitor depend on factors such as plume distance (the expected distance the emissions plume would travel), access for the research team, and landform obstructions, such as a hill that could affect the airflow and wind direction or speed.
For water quality studies, key locations to collect samples are upstream and downstream of an expected source, and if possible, in the effluent of interest too. If there are streams or rivers that come together, take samples upstream and downstream of each confluence that affects the stream or river of interest. In water bodies without a consistent direction of flow, such as a lake, take samples in a pattern that would include the most and least affected areas, such as a radial pattern emanating from the pollution source, or a point-to-point transect of the lake. If there is a density gradient (such as in a saltwater wedge), or there is a subaqueous pollution source, it may be useful to collect samples at various depths.
For exposure studies, include neighborhoods that are relatively similar, but have a different relationship to an expected source of pollution. For example, include neighborhoods that are upwind and downwind of an industrial complex, or towns with similar environmental variables but with different water sources. The upwind, or likely to be less exposed, neighborhood will act as a sort of “control” or “background” exposure metric in order to assess the increased exposure in communities of concern.
Types of Samples
There are several different kinds of samples you can collect, and each type of sample would provide you with different information. Two major categories of samples are composite samples and grab samples.
Composite samples are collections of discrete samples that are compiled or integrated, or single samples that are collected over a long time (i.e. more than 1 hour -- usually 24 hours). Composite samples can be comprised of random samples, but are more often comprised of single long collections or non-random, strategically integrated discrete samples from specific time intervals or collection locations. Composite samples are useful when trying to evaluate the average or normal conditions, or chronic exposure to a pollutant. Composite samples are used in most permits, such as NPDES permits, and federal monitoring requirements for standards such as the National Ambient Air Quality Standards (NAAQS).
The primary advantage of composite samples is that they can provide information representative of general conditions, and are less prone to erratic sample behavior. The primary disadvantage of composite samples is that they obscure potentially important variations within the composite sample timeframe or scale, such as intense acute exposures to pollutants. If you are interested in investigating whether or not your area is in attainment of NAAQS, composite samples may be more translatable to those standards (though the number of samples and types of instruments will likely differ from federal regulations, and thus should not be expected to be directly applicable). However, if you are interested in evaluating whether or not specific, not-constant activities at an industrial complex have high emissions, then a composite sample would not allow you to discern the impact of a given activity during the course of the day, and would not be the appropriate sample type (see Grab Samples below).
There are passive and active sampling devices to collect composite samples. For particulate matter sampling devices, there are passive monitors that rely on air movement and particle settling rates, including a low-cost monitor -- please see www.publiclab.org/wiki/passive-pm. Most composite sampling devices are active samplers, using pumps to draw air into the monitoring systems -- please see www.publiclab.org/wiki/active-pm.
Grab samples are discrete samples that provide information about a specific time and place; grab samples are like snapshots. Grab samples are usually collected on a short time-frame, generally requiring no more than 15 minutes to collect a single sample, and often less than one minute. Grab samples are useful to demonstrate variations in conditions over a short time frame (e.g. over the course of a single day), to capture information during an unusual or ephemeral event, or for quick check-ups on conditions. Grab samples can either be collected on a randomized schedule or plot, or at specific times and locations to documents certain activities or exposures.
The primary advantage of grab samples is that they can capture information on a much finer timescale than composite samples, and are thus able to document high-emissions events or acute exposures. Additionally, grab samples are typically much less expensive to collect and are often less mechanically complex operations. However, the primary disadvantage of grab samples is that they only provide information about a short window of time, and cannot be used to draw conclusions about average conditions. Randomized grab samples can be used as a first step toward understanding general conditions, and that approach is generally used as a less expensive (and less comprehensive) substitute for composite sampling.
Grab samples are quite common in soil and sediment monitoring, and also in water monitoring. Grab samples are somewhat less common in air monitoring, but include devices using Summa canisters or Tedlar bags, like Global Community Monitor’s Bucket.
In addition to composite and grab samples, continuous monitoring can be an important way to observe and document environmental conditions. Continuous monitoring includes non-destructive techniques that provide real-time measurements.
The primary advantage of continuous monitoring is that it provides comprehensive information about a given parameter in a given location. However, continuous monitoring also provides an often overwhelming amount of data that requires sophisticated data processing techniques. Additionally, there are often technological hurdles to continuous monitoring instrumentation such that the data quality of grab and composite samples is often better than that of continuous monitoring data.
For particulate matter, continuous monitoring can be achieved using optical monitors -- please see www.publiclab.org/wiki/optical-pm. There is a wide range of data quality from various optical particulate matter monitors, and none of them are directly comparable to particulate matter regulations, but some of them can be useful as indicator devices to prompt further investigation. Continuous monitoring can also be used to create an alert system to indicate when a useful grab sample should be collected.
Specific sampling methods will depend on the analyses that will be conducted on the sample, and it is best to consult EPA published methods prior to sample collection. However, there are also some universal best practices involved in sample collection.
For data interpretation and evaluation (see below), sampling methods must provide means to assess the method’s accuracy and precision. For real-time and continuous monitoring, instrument calibration is essential to improve accuracy and provide documentation of the instrument’s performance. For grab and composite samples, the analytical instruments must be calibrated, though that is part of the analytical method rather than the sampling method. For all types of samples, but particularly for grab samples, replication is essential to assess precision and reproducibility. A general rule of thumb is that replicate (and ideally triplicate) samples should be collected for 10% of the total number of samples, and more frequently than that for small sample sets.
In the field, it is important to handle the instruments and sample containers carefully. The inside of sampling containers, lids, and any bottle threads should not be touched. All sampling containers should be cleaned according to protocol, which generally involves rinsing in deionized and organic-free water, solvent or acid, and baking any glass bottles to 450 Celsius. Samples that will be analyzed for heavy metals generally should be collected in acid-cleaned, low-porosity plastic like high density polyethylene (HDPE) or polytetrafluoroethylene (PTFE or Teflon). Samples that will be analyzed for organic analyses should be collected in glass bottles, not plastic. Samples that contain compounds that may be UV-active, such as several organic contaminants, should be collected in amber or opaque bottles.
For outdoor air samples, sample collection should be conducted in an open area, away from any building overhangs or sources of exhaust (unless, of course, measuring that source on purpose), with sample collection devices upwind of where research teams may be standing or working.
For water samples, samples should be collected upstream from any disturbances from the research team. Samples that will be analyzed for volatile organic compounds (VOCs) should be filled beyond capacity, such that there is no headspace. Depending on the analytical method, samples may need to be preserved with solvent or acid in the field.
Samples may need to be analyzed quickly after being collected, especially if they are not preserved. Directly after sample collection, most samples will need to be stored on ice, at temperatures less than 4 Celsius (39 Fahrenheit), and some will need to be frozen at much colder temperatures.
The collection details need to be recorded along with the samples, including:
date and time
depth or height of collection device
environmental conditions (e.g.wind speed, wind direction, air temperature, solar index, water velocity, water temperature)
In addition to these collection condition details, sample sets should be kept together. Sets may include all the samples collected on a given day, or with one calibration. Sample set data should be maintained with the appropriate calibration and other relevant data quality indices (such as field blanks) to make sure that data quality can be assessed.
Interpreting the Data
The first step in data interpretation is assessing the data quality, in particular the method accuracy and precision (which is largely based on the analytical method) and reproducibility, which can be assessed by the differences in replicate samples. Evaluating the accuracy, precision, and reproducibility tells you how well you know whatever data is obtained, how large a difference needs to be before you can be confident that the difference is real, and how close your measured value is to the true value.
Once you have assessed the data quality of your samples, visually plotting the data is useful in revealing the information. Two common useful plots include:
geospatial display of concentration, such as on a map
concentration versus time for one location
When observing the geospatial distribution of your data, important auxiliary information includes wind direction (and its variability), wind speed, flow direction and flux. When observing concentration versus time at a single location, important auxiliary information includes environmental conditions (e.g. solar intensity, rain events, temperature, etc) and activities (e.g. rock crushing or coal transfer).
Assessing how well the measurements can or do represent reality (i.e. data quality), visually displaying the data, and determining correlating or confounding factors, such as environmental conditions, provide the basis for interpreting and communicating your data.
Communicating data to compel others -- industry, government, neighbors -- to change practices or behaviors often requires strategy and persistence. Relationships facilitate communication, and developing relationships at the beginning of your environmental monitoring study can prove to be more effective than developing relationships later in the process, particularly if the other parties could provide support.
Talking with local enforcement agencies prior to starting your monitoring study may engender more receptivity to your data, or result in their support of your study. Particularly with scientists or enforcement agencies, a discussion around data quality expectations is crucial, as most civic technologies (with a notably exception of aerial mapping) do not provide the same data quality as government monitoring equipment, but can be extremely useful in indicating problem areas or prompting official investigation. Clear communication about the study design and objectives will facilitate more favorable working relationships with any parties, and especially enforcement agencies. Engaging local government officials, such as city council members or town boards, early in the study development may inspire participation, deeper understanding of study objectives, and more willingness to listen to the results of your study. Local government offices are often more receptive to community data than larger, state or federal agencies. When working with people with a variety of backgrounds or experiences, such as you would expect in local government, presenting the data in a visually compelling way can be very effective. Additionally, including multiple leverage points or angles in your advocacy work with local government can often yield results. For examples of advocacy strategies, please see www.publiclab.org/wiki/frac-sand-advocacy-strategies.
Engaging the media can also be an effective advocacy strategy, particularly to reach wider audiences and gain more diverse support. For tips on working with the media, please see www.publiclab.org/wiki/create-a-media-campaign.