Case Study #4: Monitoring a regional infiltration system

Emmons and Olivier Resources and the South Washington Watershed District.

The infiltration of runoff water into the ground has been proven to be an effective method of stormwater management in the Minneapolis-St. Paul, Minnesota region. Specifically, the South Washington Watershed District (SWWD - Figure 1) has been using an infiltration system as the backbone of its drainage network and has been monitoring its performance to varying degrees since 1997. Results for two infiltration basins and two infiltration trenches are presented in this case study.

The SWWD is a 54 mi2 suburban watershed located over very sandy and deep glacial outwash. The watershed is rapidly developing with a 2006 population of about 100,000. The northern two-thirds of the watershed is landlocked. The developed portion is internally drained through a series of ponds, wetlands, and lakes terminating in a large natural depression (CD-P85) that acts as an infiltration basin. CD-P85 contains two constructed infiltration trenches (one monitored) and several dry wells (not monitored). The undeveloped portion consists of two large landlocked subwatersheds that drain to natural depressions that act as infiltration basins referred to as CD-P76 and CD-P82. Land uses within these subwatersheds are predominantly row crop agriculture and rural residential. Figure 2 (a-c) is a collection of photos of the facilities located at the three CD-P sites while they are partially full of infiltrating water.

Figure 3 illustrates an SWWD infiltration facility at the Woodbury Math and Science Academy (MSA), which has been monitored since 2001. This infiltration trench measures 15 feet long by 8 feet wide and approximately 13 feet deep and occurs approximately 3 feet off the bottom (on a side slope) of the adjacent basin. The lower portion of the basin was designed as a sedimentation basin to allow sediment to settle prior to runoff reaching the infiltration trench. The basin was planted with deep rooted native vegetation, and has not maintained a permanent pool of water since construction in 1999. The MSA trench drains only 0.4 ha from a portion of the school roof, a swale draining a high traffic county road, and a small amount of open space. Figure 3 illustrates a cross-section view of this facility.

Assessment Goals

This case study presents the approach and results of monitoring natural depressions located within CD-P76 and CD-P82 and also constructed infiltration trenches at CD-P85 and MSA (see Figure 1).

Monitoring methods used to evaluate infiltration and potential impacts to groundwater include measuring continuous water levels with the basin or trench, sampling surface and groundwater for water quality parameters, and measuring water level fluctuations responses in the water table.

Data are collected during spring melt conditions, typically beginning in February. Spring melts within Minnesota typically consist of several minor events during mid-winter, followed by a major solar-driven melt event in mid- to late-March. Data are also collected during the summer and fall season in response to rainfall events.

The goals of this multi-year study have been and continue to be:

  1. assessment of the long-term performance of a regional infiltration system
  2. understanding the physical mechanisms that promote effective infiltration
  3. documentation of the water level and water chemistry changes that occur under infiltration facilities.

Assessment Techniques

To meet the assessment goals, a monitoring approach has evolved to include measurement of both inflow and outflow quantity and quality, as well as the groundwater level fluctuation under the site. MSA measurements focus on the inflow of water and the infiltration rate, but the lack of a monitoring well means that groundwater behavior under the facility is not currently monitored.

Water quality data have also been collected as part of the overall infiltration monitoring program. Surface water samples are analyzed for the four locations in Figure 1 for dissolved heavy metals (cadmium, lead, nickel, manganese, zinc, copper), volatile and total suspended solids, total phosphorus, ortho-phosphate (as phosphorus), total Kjeldahl nitrogen, nitrate plus nitrite as nitrogen, chloride and hardness. Groundwater samples are collected at CD-P82 and at CD-P85 and analyzed for dissolved heavy metals (cadmium, lead, nickel, manganese, zinc, copper), nitrate plus nitrite as nitrogen, and chloride.

Composite grab samples of ponded surface water in the process of infiltrating are collected using three-foot long polyethylene disposal bailers and poured into individual sample bottles. At MSA, composite flow-weighted samples are collected during overflow events using an automatic sampler and flow meter.

Groundwater samples are collected from monitoring wells using a submersible pump, and a minimum of three well volumes were purged prior to sample collection.

Groundwater level data are collected at eight wells throughout the watershed as part of this program. One well is located adjacent to each of the CD-P82 and CD-P76 basins, and six are located adjacent to or near CD-P85. MSA does not currently have a monitoring well. During construction of the new outlet at CD-P85, the MW-4 well was damaged and later abandoned in early spring 2005. Except for well MW-3w, the peak groundwater mound elevations could not be determined without the use of continuous data loggers in the wells. However, for the discussion on ground water mounding, the periodic hand measure well readings were used to identify the highest observed mounding in each well. Water level readings were taken at each well with the use of an electronic water level sounder, and an automatic data logger was installed in the MW-3w well to record continuous groundwater levels.

Assessment Results

Monitoring methods used to evaluate infiltration and potential impacts to groundwater include measuring continuous water levels with the basin or trench, sampling surface and groundwater for water quality parameters, and measuring water level fluctuations responses in the water table.

Data are collected during spring melt conditions, typically beginning in February. Spring melts within Minnesota typically consist of several minor events during mid-winter, followed by a major solar-driven melt event in mid- to late-March. Data are also collected during the summer and fall season in response to rainfall events.

Results – Infiltration Rates

The SWWD drainage system is an ideal geologic system for implementation of infiltration BMPs. The results of five to seven years of monitoring the following sites are summarized in Table 1. Summaries of each facility follow including example data collected during the spring melt of 2005.

CD-P76

This infiltration basin receives runoff from a 480 acre watershed consisting of row crop agriculture and rural residential development. The basin typically fills to three feet in depth, with an aerial extent of 5.7 acres. Soils in the basin consist of sandy loam over medium grained sand deposits. Depth to the water table is greater than 60 feet. There is no groundwater monitoring at this site. Figure 4 illustrates the behavior seen in this basin during snowmelt events in 2005 when water level rose within the basin to a maximum depth of two feet. No outflow occurred during these events.

CD-P82

CD-P82 receives runoff from a 580 acre watershed comprised predominantly of row crop agriculture and horticulture land uses. The basin typically fills to eight feet in depth, with an aerial extent of 7.5 acres. Soils in the basin and watershed consist of silt and sandy loam over medium grained sand deposits. The silt deposits are found at the bottom of the basin within and surrounding the pond. The basin contains a small sump area at the bottom, which has been sealed over time with fine grained particles. Sandy loam is present on the basin sides and throughout the watershed. Depth to the water table is approximately 30 feet. A groundwater mound typically forms beneath this basin during spring melt conditions, as recorded in an on-site well. During 2005, a groundwater mound formed that was 5.2 feet high. This mound formed and receded over a two month period and likely coincided with a regional water table rise during the spring season. Figure 5 illustrates this basin’s behavior during 2005 spring melt conditions, during which no outflow occurred.

CD-P85 (Trench)

CD-P85 is a natural basin within which occur two infiltration trenches and four unmonitored dry wells. Although CD-P85 occurs within a rural, undeveloped area, the largest volumes of water entering the basin are from pumped storage out of the terminal pond in a long chain of urban drainage storage facilities. During spring melt conditions, runoff also enters the basin from the direct drainage area (354 acres) which is comprised predominantly of row crop agriculture land uses. During the spring melt, there is typically very little standing water in the basin. Soils in the watershed consist of very sandy loam over sand and gravel deposits. The basin is a result of a large ice block deposit. At the bottom of the basin, a thick layer of clay is present that is bypassed either laterally into the basin banks or via two infiltration trenches that break through the clay layer. Depth to the water table is typically greater than 50 feet. Six water table wells surround the CD-P85 basin. A groundwater mound does not typically form during spring melt conditions.

Occasionally, runoff enters CD-P85 through a pumped discharge from a large holding pond east of the basin, over a small watershed divide. When this occurs, ponded water depths are as great as 28 feet, with an aerial extent of 25 acres. Groundwater mounding is prevalent during pumped events. The groundwater mound will often intersect the basin.

Two trenches reaching approximately 14 feet into the bottom of the CD-P85 basin were installed during 1999; six years after the basin became operational. These trenches were designed to provide a pathway for ponded water to infiltrate through the clay layer at the bottom of the basin and into the sand material below. The trenches allow for the basin to dry between events, limiting long-term ponded water and allowing the basin and vegetation to recover between events.

Figure 6 illustrates the infiltration behavior of the trench during the 2005 spring melt events. Figure 7 shows the compilation of all data collected since monitoring began. The long-term mean infiltration rate for the trench has diminished slightly over time perhaps due to fines entering the trench, which was placed on the bottom of the basin rather than slightly above that elevation.

MSA Trench

Perhaps the most interesting of all of the SWWD infiltration facilities is the Math and Science Academy (MSA) trench, which has been monitored since 2001. This trench measures 15 feet long by 8 feet wide and approximately 13 feet deep and occurs approximately three feet off of the bottom (on a side slope) of the adjacent basin. The basin was designed as a detention basin to allow sediment to settle prior to runoff reaching the infiltration trench. The basin was planted with deep rooted native vegetation, and has not maintained a permanent pool of water since construction in 1999. Figure 8 illustrates a cross-section view of this facility.

The MSA trench drains only one acre from a portion of the school roof, a swale draining a high traffic county road, and a small amount of open space. Figure 9 shows the numerous snowmelt and rainfall events during spring of 2005 that entered this settling and infiltration facility.

Figure 10 shows how more infiltration is occurring at the same elevation today than earlier years. The rate of infiltration is increasing with each year since monitoring began. Research on the reasons for the improvement has not occurred, but speculation is that it has resulted from very good vegetative growth in the entire basin which has improved conditions through better energy dissipation and solids filtration prior to infiltration, massive root growth downward by the native plants and the insect borings that are visible at the site. The behavior of the MSA Trench is opposite that of the CD-P85 Trench which has decreased in infiltration rate as it ages. The CD-P85 trench is located directly on the bottom of a much larger basin where organic and inorganic fines can gather.

Finally, infiltration of snowmelt runoff into the MSA Trench, although very effective, does not appear to be as high as it is for rainfall events. Figure 11 illustrates the monitored infiltration rates for all 2005 events. Note that Drawdowns 1-7 indicate the snowmelt events in the index box. Two separate trend lines were added to the graphic to generally portray the somewhat reduced snowmelt infiltration events (lower line) beginning at about three feet of water depth.

Results – Water Quality

Water quality data have also been collected as part of the overall infiltration monitoring program. Surface water samples are analyzed for the locations in Table 1 for dissolved heavy metals (cadmium, lead, nickel, manganese, zinc, copper), volatile and total suspended solids, total phosphorus, ortho-phosphate (as phosphorus), total Kjeldahl nitrogen, nitrate plus nitrite as nitrogen, chloride and hardness. Groundwater samples are collected at CD-P82 and at CD-P85 and analyzed for dissolved heavy metals (cadmium, lead, nickel, manganese, zinc, copper), nitrate plus nitrite as nitrogen, and chloride.

Although water quality data have been collected for many years, the complexity of the flow system has complicated analysis and a detailed water quality model of the system has not been prepared. Table 2 contains meltwater pollutant data for the most recent four years for surface water inflow and four shallow groundwater wells situated around CD-P85 (Table 2a) and a single monitoring well at CD-P82 (Table 2b). The data reflect the range of values observed for snowmelt events only as an example of data collected.

The data for CD-P85 show that groundwater pollutant levels are generally consistent with the surface water samples for most of the metals. The exceptions are Mn and Ni which are higher in the groundwater. The consistent manganese (Mn) groundwater violations are reflective of high ambient concentrations in groundwater in Minnesota. Dissolved nickel (Ni) in MW2 and MW3 could be problematic, although similar high levels are not apparent in the surface water inflow. Nitrate and chloride levels are similarly higher than the pollutant levels in surface water flowing into the infiltration system, This behavior is indicative of regional ambient groundwater quality which is high for NO3 from historic agricultural and septic system inputs.

The chloride (Cl) data paint an unclear picture of how chloride-laden water moves into the CD-P85 system throughout the year. The two possibilities are lateral flow from transportation corridors to the north and the higher concentration of Cl in water pumped in during non-melt periods throughout the year. That is, water from the highly urbanized part of the watershed is gradually routed through the stormwater system and into CD-P85, therefore not reflected in surface water snowmelt monitoring data at CD-P85. The Cl levels, although higher than normal, do not violate any water quality standards, but should be a warning that continued high salt use can lead to groundwater accumulations that could become troublesome. Although some of the pollutants are reflective of high ambient groundwater conditions and warrant attention, it does not appear that the CD-P85 infiltration system is negatively impacting groundwater.

CD-P82 monitoring data are collected for surface water and a single well adjacent to the basin. Table 2b shows generally the same kind of metals behavior as CD-P85, with the exception of lower levels of Ni in the groundwater. Nitrate inflows are about the same, but groundwater levels are high enough to have violated standards during one sampling event (February 2002). Chloride levels are higher than naturally occurring groundwater, but not in violation of any standard as yet. As with CD-P85, the impact of the infiltration system on local groundwater does not appear to be significant.

Conclusions and Recommendations

  1. Large-scale infiltration in cold climates can be an effective management practice if soil and geologic conditions are “suitable”, which means adequate soil permeability, dry conditions at the time of freeze-up, and bedrock depths well below (at least six feet) the bottom of the infiltration system. These conditions are common across many parts of the upper Midwest, particularly in glacial outwash plains.
  2. Infiltration rates as high as 1.1 in/hr have been documented for snowmelt events within natural infiltration basins. Rates as high as 4 in/hr are documented for an infiltration trench during spring melt and 6.8 in/hour for a summer rainfall event.
  3. Although not discussed in this study, a maintenance program to assure removal of fine particulate matter is essential to the proper and long-term operation of any infiltration system. Situating infiltration trenches above the bottom of a detention pool and using deep-rooted native vegetation help to minimize the maintenance need for repeated removal of sediment and reduced infiltration. The accumulation of fine-grained material has had an impact on some of the infiltration facilities, but not enough to warrant extensive maintenance after up to 10 years of operation.
  4. Although some elevated pollutants have been detected in groundwater near the infiltration system, monitoring data indicate that infiltrating surface water is not the cause of this problem. The most common violations are for naturally occurring elements in the soils and are commonly found within surface and groundwater in southern Washington County. Chloride does not currently exceed any water quality standards, but is higher than “non-impacted” water and should be a warning for careful management of salt application.

Return to Case Studies.