You Were Collecting Stormwater Samples and What Happened?

On all too many occasions when attempting to collect stormwater runoff samples or data, we have failed to successfully complete the task. Sometimes, the circumstances that created the problem were unforeseeable and, in some instances, were almost unbelievable. We have returned to the office and shared our field stories with colleagues only to be met with “You say what happened when you were collecting samples?” The purpose of this article is to summarize representative problems that we have encountered, along with some “lessons learned.”

Lessons Learned
Here are the key lessons we have learned:

• Make sure that the objectives of the sampling/data collection are well understood by all members of the team—the field staff, the office staff, and the entity paying the bills.
• Adverse field conditions often occur.
• Plan on two or three sampling attempts for every one valid sample. Accordingly, project the anticipated field time (and cost) and then double or triple it.
• Be patient and recognize that after enough trials and with sufficient time, it will be feasible to collect the necessary number of acceptable samples.
• Make sure that the people who are out in the field have been trained and that they have common sense and are resourceful.
• It is often necessary to have two or more people in the field collecting data; safety is of paramount importance.
• Finicky monitoring equipment is a fact of life.
• Diplomacy can be necessary.
• Document problems with field notes and photographs. Making a mistake is not a problem; “burying” it is a major problem.
• Be thorough, thoughtful, and “middle of the road” in your approach to gathering field data, because the data are often related to regulatory compliance and/or legal issues, where the data will be scrutinized.

Engineers and scientists who work with and apply field data are well advised to participate in data-collection activities, to develop perspective on where the data points come from and on the unavoidable uncertainty and variability (to at least some degree) of the data.

Variables in Data Gathering
Numerous public and private entities collectively spend hundreds of millions of dollars annually collecting and analyzing stormwater runoff samples, gathering runoff data, surveying channel morphology changes, sampling aquatic life, etc. The purposes of such monitoring are wide-ranging. Entities may need to collect stormwater-quantity and water chemistry data to satisfy National Pollutant Discharge Elimination System (NPDES) or other permit requirements. Sometimes data are collected purely for scientific purposes. Parties may be interested in learning how effectively best management practices (BMPs) such as wet ponds, wetlands, or swales are removing pollutants from runoff and/or attenuating peak discharges. Other times, there will be a need to characterize runoff quality from a range of land uses. Some studies focus on hydrology, rather than water quality, such as the establishment of a network of flow gauges to determine hydrologic responses to precipitation events from various types of land use, and measuring changes in channel morphology.

The ability of a beaver to repeatedly dam an outflow weir was frustrating and time consuming.

Sampling can be conducted in rural or urban settings. Adjoining property owners may have no interest in the monitoring or may be extremely interested and opinionated. Samples can be “grab” in nature or collected via automated monitoring equipment that enables measurement of flow rate concurrently with sample collection. The data collected may be used for regulatory compliance purposes or purely for diagnostic reasons. The field team may be experienced and “savvy” or highly inexperienced (unfortunately, there is a tendency to send young, untrained staff to the field to collect samples and data). In short, wet-weather monitoring can encompass myriad objectives, field practices, kinds of monitoring equipment and personnel, settings, and field conditions.

Sediment accumulation where intake and transducer are placed creates problems.

Flooded monitoring vault

Fecal Coliform Happens
Regardless of the purpose of the monitoring, a common feature is that people need to go out into the field and actually collect data. The likelihood of encountering problems during the process of obtaining field data is significant; consider the following illustrative experiences of the authors (including the photographs and captions):

1. A primary issue is poorly planned sampler locations. Imagine driving 20 miles to a sampling site (a shopping center) to find a car parked over the door to the sampler—repeatedly. On various occasions we were unable to access a sampler in a utility vault on the edge of a busy retail parking lot due to the cars parked almost on top of it. After many requests, the site owner let us move a shopping cart return rack to the parking space next to the sampler box to keep cars from being parked there. Now, imagine two months later finding the same sampler covered in 5 feet of dense, plowed snow. It took two of us about two hours to dig down to it!
   
   The same retail client also has porous pavement monitoring wells in its parking lot to enable site water balance calculations to be made. Sampling personnel sometimes have no choice but to crawl under cars to gently crack the lid to the well to grab the Barologger device used to record water levels in the infiltration beds. This can be awkward and potentially dangerous when shoppers return to their vehicles (thus emphasizing the importance of field sampling with at least two people).
   
2. Occasionally, difficulties are caused by overly engaged citizens. For example, a gentleman who lived near a stormwater treatment wetland/pond outlet structure repeatedly plugged the water-quality weir/orifice. This raised the normal water surface by 1 foot and disrupted the fundamental design basis of the facility, thus negating the value of any samples that were collected. Every time he blocked the outlet, we removed the blockage, only to have it replaced. This pattern went on for about two months, until we finally left a note attached to the outlet box explaining that the blockage was disrupting the wetland/pond function and leaving our contact information. The person ultimately contacted us, and we met in person. He explained that he felt that aquatic life and wildlife would benefit from the additional 1 foot of water, and he had no concept that the outlet he repeatedly plugged was designed to release the “water-quality capture volume” over 24 hours to promote pollutant removal.
   
   Another time, a woman accidentally broke a rain gauge because she was curious how it worked (or didn’t work). We have been asked by a child standing on a nearby bridge, eyeing us as we mounted a topping bucket rain gauge on a post, what would happen if “hypothetically” someone were to toss a rock into the rain gauge barrel. Similarly, we have seen a child pull the bubbler tubing out of a mounting, and we’ve seen flumes and weirs used for target practice.
   
   One person accused us of violating Section 404 of the federal Clean Water Act by pulling four handfuls of cattails away from a blocked pond outlet. On many occasions, neighbors have complained to us about the unsightly appearance of BMPs, assuming that this is our responsibility.
   
3. Peripheral factors can interfere with sampling. One time a water line break at a nearby fast-food restaurant filled up a bioswale with water. This triggered the outflow sampler and necessitated a trip to the site to rewash the bottles, change the battery, and reset the trigger at a high enough level that it wouldn’t go off again until the line was fixed and the water drained. Sprinklers have been known to cause “false starts” as well.
   
4. The natural world that we are trying to protect via sampling can also be a hindrance. A colony of ants took over one sampler, and a spider built an egg sack in one of the cables in another. Once, high fecal coliform readings led us to discover a family of mice nesting in a storm drain a short distance upslope from the sampler intake.
   
   In another case, a rain gauge at a pond continually suffered unexplained and unexpected clogging, which prevented the samplers from triggering. A little surveillance eventually showed us a red-tailed hawk that routinely sat on the gauge and “fowled” the sampler.
   
   One of the most formidable natural foes encountered during monitoring by one author was Castor canadensis, the American beaver, who decided that a weir constructed to regulate outflows from a constructed wetland made an ideal starting point for dam construction. This, of course, altered outlet hydraulics, modified the water level in the pond and pond function, and resulted in deforestation of the outlet zone. Over the course of several months, breaching of the beaver dam with a pickaxe to restore outflow function was a weekly ritual. Working thigh-deep in pond muck and water, the author ended up losing balance and falling completely into the fetid water on several occasions. Eventually, a device known as the Clemson Beaver Pond Leveler (found through the Clemson Agricultural Extension Service) provided a solution by subverting the dam with a pipe fed by a submerged, cage-protected inlet, which was constructed and installed by a group of graduate students at the University of Virginia.
   
5. Sometimes the weather does not cooperate. In July 2005, all of the storms that occurred in one part of the Denver metro area happened on Saturdays when the lab was closed. Other egregious errors committed by uncooperative storms include these:
   
   a. Being either too small or too large for the sampler to handle
   b. Not occurring at all
   c. Falling with such low intensity that the samplers could not respond to them
   d. Being too short to capture anything but the first flush
   e. Falling in the wrong direction (i.e., horizontally during high winds) so the rain gauges do not read properly
   f. Falling as snow and not rain
   
6. Probably the largest cause of sampling headaches is the equipment itself. Although sampling devices have come a long way over the years, there is still much to be desired in terms of dependability and ease of use, especially in the more economical apparatuses many clients naturally prefer. Representative problems include these:
   
   a. Broken AV sensors
   b. Equipment that will not operate because it is too hot (The stainless steel lids that cover the sampler vaults can get to over a 120°F and cause samplers to malfunction.)
   c. Equipment that will not operate because it is too cold (Late-season sampling can feature air temperatures below 20°F, which may cause the samplers to malfunction.)
   d. Trash and grass clippings that clog sampler inlet tubing
   e. Lightning strikes that kill the power to the samplers for longer than the backup batteries last
   f. Automated storm notifiers that don’t notify
   g. Samplers that don’t trigger when they should
   h. The bubbler module reading incorrect water levels
   
7. Safety must be everyone’s top priority. Our staff have fallen through ice in ponds and rivers; fallen in swift current and on riprap and steep embankments; sampled under “confined space” conditions; been confronted by rattlesnakes, water moccasins, biting insects, and dogs all too often; and sampled industrial sites with suspected hazardous wastes where highly specialized sampling clothing, equipment, and precautions were necessary. Safety is best addressed through proper training and by not sending someone to the field alone, unless the collection is virtually certain to be risk-free.
   
8. Finally, we must admit that there is such a thing as sampling technician error. We have learned from a few mistakes over the years:

Advertisement

a. Sending a 100-pound intern to install a 50-pound sampler battery
b. Using construction contractors who did not fully understand the intent of our design for monitoring station vault and conduit locations, elevations, etc. This resulted in a sampler and battery submerged in water, ruining the battery and causing the sampler to float on its side, spilling the samples.
c. Running into traffic gridlock when trying to get the sample to the lab before it closes or within holding time
d. Realizing, when already in the field and ready to take the sample, that the lab gave us either not enough or the wrong sample bottles
e. Figuring out the perfect adaptation to make the equipment work properly just as the battery in the cordless drill runs out
f. Programming the sampler for inches when it calls for feet
g. Providing the lab with incorrect analytical instructions, such as detection limits that are too high, or requesting total recoverable metals when the client needed dissolved metals data
h. Preparing quality assurance/control plans that were unreadable and overly academic
i. Summarizing the data in databases and spreadsheets incorrectly

Clearly, one cannot plan for every contingency in the monitoring business. However, through careful planning, experience, checking, and collaboration/interaction with others who collect field data, the frequency of the kinds of problems described above can be reduced, and the quality of the data collected can be reasonable.    

Author's Bio: Jonathan Jones, P.E., is chief executive officer, with Wright Water Engineers Inc. in Denver, CO.

Author's Bio: Andrew Earles, Ph.D., P.E., is with Wright Water Engineer Inc. in Denver, CO.

Author's Bio: Sally Kribs is with Wright Water Engineers in Denver.