Using leaf collection and street cleaning to reduce nutrients in urban stormwater

Science Center Objects

Decaying organic materials, like leaf litter, can release excess nutrients into local streams and lakes, causing eutrophication and algal blooms. To determine if a municipal leaf collection and street cleaning program can reduce nutrients in stormwater runoff, the USGS measured phosphorus and nitrogen in stormwater from residential areas in Madison, Wis.

Graph showing total phosphorus before and after the implementation of a leaf collection and street sweeping program

This graph shows a significant reduction in total phosphorus in a residential neighborhood in Madison, Wis., after the implementation of a timely leaf collection and street cleaning program (Selbig, 2016).

HIGHLIGHTS

  • Leaves are a significant source of phosphorus to urban stormwater.
  • Nearly 60 percent of the annual phosphorus yield can come from leaf litter in the fall.
  • Timely removal of leaf litter from streets can reduce phosphorus load by 80 percent.
  • Leaf removal is one of a few options available to reduce dissolved phosphorus.

 

Photos of leaf litter and organic detritus in streets and storm drains

Leaf litter and organic detritus in streets and storm drains in Madison, Wis.

BACKGROUND

Excessive amounts of nutrients (phosphorus and nitrogen) in stormwater runoff can accelerate the effects of eutrophication in urban streams and lakes, leading to algal blooms which can block sunlight for aquatic plants, clog the gills of fish, reduce levels of dissolved oxygen, and produce toxins that are harmful if ingested. While the sources of nutrients to urban stormwater are many, a primary contributor is often decaying organic materials like leaves. This "organic detritus" can be especially significant in urban areas with dense overhead tree canopy. Rain and melting snow percolates through freshly fallen or already decaying leaves on streets and washes into storm drains that are typically routed directly to receiving water bodies. The leaves continue to decay, releasing excess phosphorus and nitrogen into local streams and lakes.

A city-wide leaf collection and street cleaning program removes leaves and other organic detritus from the street surface, thereby preventing phosphorus from becoming entrained in stormwater. The goal of this study was to determine if, and how much, removing leaves and other organic detritus from streets could reduce nutrient contributions to local water bodies. This knowledge could help tailor more targeted municipal leaf collection and street cleaning programs that may improve the quality of urban stormwater. 

 

SITE DESCRIPTION

To determine the potential for a municipal leaf collection and street cleaning program to reduce nutrient concentrations and loads from stormwater runoff, the U.S. Geological Survey measured concentrations of phosphorus and nitrogen in stormwater from two residential areas in Madison, Wisconsin. In the 6.47-hectare control catchment (neighborhood), there was no effort to remove leaf litter and other organic detritus from streets - no street cleaning or leaf collection. In the 1.21-hectare test catchment, leaf litter was removed through a combination of municipal leaf collection, street cleaning, and leaf blowers. (For this study, a catchment is an area whose stormwater runoff can be intercepted before any contributions from external areas.)

Map of the control and test catchments for a leaf collection and street cleaning study in Madison, Wis.

Location of the control and test catchments in Madison, Wis., with spatial coverage of tree canopy over land and street surfaces (Selbig, 2016).

Both catchments were medium-density residential areas with curb and guttered streets approximately 9.75 meters wide. There were no catch basins or other storage areas in the storm drain network of either catchment. The two catchments were located within 1.5 kilometers of each other to help reduce variation in storm rainfall patterns, and were similar in physical characteristics, including land use, street condition, overhead tree canopy, topography, and lot size. Despite differences in drainage area, the percentage of impervious and pervious surfaces was evenly split with slightly more pervious than impervious.

The trees were generally a mix of mature, deciduous hard and softwood species, with more than 70 percent of street trees as Norway maple (Acer platanoides) or Green Ash (Fraxinus pennsylvanica). Leachable phosphorus can vary among different tree species, so the range of nutrient concentrations reported in this study may differ from other areas with different tree species. A greater percentage of area was covered by tree canopy in the test catchment (64 percent) than the control (46 percent). However, the percentage of tree canopy covering streets was the same in each catchment at 17 percent.

The climate in Madison is typical of interior North America, with a large annual temperature range and frequent short-period temperature changes. Accumulation of leaf litter does not typically start until shortly after leaf senescence in late September, and continues through mid to late November.

 

STUDY DESIGN

A paired-catchment design was used to help evaluate the effectiveness of leaf collection. The basis behind a paired-catchment approach is that there is a quantifiable relation between paired water-quality data, and that this relation is valid until a major change is made in one of the catchments. At that time, a new relation will develop, which allows the effects of the change to be quantified. The strength of this approach is that it does not require the assumption that the control and test catchments are statistically the same; however, it does require that the two catchments respond in a predictable manner together, are subject to similar precipitation events, and that their relation remains the same over time except for the influence of the change (in this case, leaf collection).

The calibration phase occurred in the months of May and September through November in 2013 and April through November in 2014. During this time, paired water-quality samples were collected to develop a relation between the control and test catchments without leaf collection or street cleaning. In 2015, the treatment phase (leaf collection and street cleaning) was implemented in the test catchment, while the control remained the same.

 

Photos of leaf collection and street cleaning study sample collection and study area maintenance

Sample collection and study area maintenance.

COLLECTING AND MEASURING NUTRIENT CONCENTRATIONS

A monitoring station was used to measure stormwater flow and collect water samples at the storm sewer outfall of both the control and test catchments. Precipitation data were collected by use of a tipping-bucket rain gage, and there was no dry weather flow in either the test or control storm drain network.

Sample collection was activated by a rise in water level in the storm sewer pipe during a precipitation event. The volume of water passing the monitoring station was measured every minute until a specific threshold was reached. At that point, a depth-integrated sample arm (DISA) collected a discrete water sample. A DISA was used because it collects a water sample from the entire water column - rather than a single, fixed point - limiting any nutrient concentration bias caused by the stratification of solids in storm sewers. This process was repeated until the water level receded below the threshold. The discrete water samples collected over the duration of an event were combined into a single, composite sample. 

Total phosphorus and nitrogen is comprised of two different phases: particulate and dissolved (less than 0.45 micrometers). Dissolved nutrients are more difficult to remove since they are easily entrained in the stormwater, so it's important to identify the portion of nutrient contributions being transported in a dissolved state. For this study, all samples were tested for total and dissolved phosphorus and total and dissolved nitrogen. Additionally, both the load (the amount of nutrients transported in the stormwater) and yield (the amount of nutrients transported in the stormwater divided by the contributing area of the catchment they came from) were computed for each study area, phase, and season. (A yield comparison provides a way to evaluate and relate loads between catchments of different sizes.)

 

Photos of transfer-based leaf collection and street cleaning

Transfer-based leaf collection and street cleaning by the City of Madison, Wis.

LEAF COLLECTION AND STREET CLEANING PRACTICES

In late September to early October of each year, the city of Madison continuously rotates a fleet of leaf collection vehicles to collect and remove leaf litter and other organic detritus from residential areas. Residents are asked to pile their leaves adjacent to the street to limit excess debris in the street gutter. Upon collection, a vehicle equipped with a modified plow will first move any piles of leaves near the curb into the street. The leaves are then pushed into a garbage collection vehicle for removal. A street cleaner services the area within a few days following leaf collection to remove any residual organic debris from the street and gutter.

During April through September, weekly street cleaning was the only form of treatment in the test catchment. Leaf collection, in addition to street cleaning, was done in October and November and occurred approximately every 7 days. The control catchment remained without any leaf collection or street cleaning throughout the entire study period.

Despite weekly municipal operations, an appreciable amount of leaves and other organic debris would accumulate on the street surface in a matter of a few hours to a few days. Therefore, in order to characterize a "best case scenario" for municipal operations, USGS personnel were deployed to the test catchment in October and November during the treatment phase of the study to remove all organic detritus from the street prior to a precipitation event. Field crews used high-powered leaf blowers to transfer all debris from the street to an area that was not in the drainage area to prevent any stormwater contributions. While this extra measure of leaf removal exceeds the capabilities of most municipal leaf collection programs, it sets a benchmark for the greatest potential reduction of nutrients in runoff through removal of leaves and other organic detritus from urban streets with high overhead tree canopy.

A total of 71 paired samples were collected over the study period (40 collected during the calibration phase and 31 collected during the treatment phase). Sample results from the calibration phase were used to establish a relation between the control and test catchments without any leaf collection or street cleaning, while sample results from the treatment phase were used to evaluate any changes in that relation due to the implementation of the leaf collection program.

 

MAJOR FINDINGS

Results from this study confirm that fall leaf litter can be a primary source of nutrients in stormwater, particularly phosphorus. When an active, thorough leaf removal and street cleaning program is in place, total and dissolved phosphorus loads were reduced by 84 and 83 percent, respectively, and total and dissolved nitrogen loads were reduced by 74 and 71 percent, respectively compared to no leaf removal and street cleaning program. The timing of leaf removal is important because of the highly leachable nature of leaves, and significant reductions in loads of the total and dissolved forms of phosphorus and nitrogen can be achieved with removal of leaf litter prior to a precipitation event.

Under non-leaf-removal conditions, fall leaf litter contributed 56 percent of the annual total phosphorus yield (winter excluded) compared to 16 percent when leaf removal was implemented. However, fall leaf litter showed only a minor change for nitrogen, contributing 19 percent of the annual nitrogen yield without leaf removal compared to 16 percent with leaf removal.

Additionally, in the fall, the majority of nutrient concentrations were dissolved, and the removal of leaves accumulated on streets and in piles near the street curb can reduce contributions of nutrients to storm drains. This makes leaf collection one of the few treatment options available that can be effective at reducing the amount of dissolved nutrients in stormwater runoff.

Graphs of log-transformed phosphorus and nitrogen loads from a leaf collection and street cleaning study

​​​​​​​Log-transformed nutrient loads (phosphorus and nitrogen) of paired samples measured in stormwater runoff in the control and test catchments during the calibration and treatment phases (Selbig, 2016).

 

BY STUDY PHASE (CALIBRATION VS. TREATMENT)

  • Regardless of catchment, the overall pattern of mean monthly concentrations was similar during the calibration phase
  • Phosphorus concentrations were generally lowest in summer and highest in fall.
  • Nitrogen concentrations were lowest in summer but highest in spring.

Calibration Phase

Mean monthly concentrations in both the control and test catchments showed little variation from April through September with only minor increases measured during the month of May, likely due to the presence of new blossoms, seeds and pollen from emerging vegetation. As leaves matured by early June, concentrations of phosphorus were lower and remained relatively steady through September.  During this period of leaf maturation, sources other than seeds and leaves, such as street dirt and grass clippings, were likely the primary contributor to phosphorus and other nutrients in runoff. The maximal amounts of nutrients measured in runoff occurred from senescent leaf litter in the fall where appreciable gains in concentration were observed in October, a period of time when the amount of leaf litter intensified. The largest gains were observed in dissolved phosphorus with a 979 and 933 percent increase over average summer (June through September) concentrations in both the control and test catchments, respectively. Dissolved nitrogen similarly increased but to a lesser degree (91 and 18 percent, respectively). As fall advanced, fewer leaves were deposited thus minimizing sources of nutrients. With the exception of nitrogen in the test catchment during the calibration phase, nutrient concentrations in November declined from October levels but remained above those measured during spring and summer months. This pattern was observed in the control catchment for both study phases as well as the test catchment during the calibration phase, indicating the monthly distribution of nutrients without a leaf collection program was generally repeatable both spatially and temporally.

Nitrogen, unlike phosphorus, appeared to be more erratic and less predictable. Similar to phosphorus, mean monthly concentrations of total and dissolved nitrogen were generally lowest during summer months with higher concentrations observed in May and October. In contrast to phosphorus, the magnitude of increase in October was less, with concentrations higher than summer levels but lower than those observed in spring. It is unclear why the month of November showed an increase in total and dissolved nitrogen in the control catchment during the calibration phase when all other instances observed a decrease. One such source may have been the application of lawn fertilizers which contain nitrogen as a nutrient to stimulate root growth. Lawn fertilizer is typically applied in the spring and fall, which coincided with observed increases in concentrations.

Treatment Phase

Mean monthly nutrient concentrations in the control catchment during the treatment phase were not significantly different than the calibration phase since there was no change in leaf collection practices. The test catchment, like the control, also displayed a pattern in nutrient concentrations similar to the calibration phase, but only for the months of April through September. The month of May continued to show slightly higher mean concentrations than the rest of the spring and summer months despite the addition of weekly street cleaning efforts. In contrast, the combination of leaf collection and street cleaning in the months of October and November reduced nutrient concentrations to near summer levels. Compared to the calibration phase, mean October concentrations of total and dissolved phosphorus in the test catchment during the treatment phase decreased by approximately 80 percent.

Graphs showing total and dissolved phosphorus and nitrogen before and after the implementation of a leaf collection program

Mean monthly concentration of nutrients (phosphorus and nitrogen) measured in stormwater runoff in the control and test catchments during the calibration and treatment phases (Selbig, 2016).

 

BY SEASON

For this study, the months of the year were lumped by season in which leaves were either emerging (spring: April-May), mature (summer: June-September), or in recession (fall: October-November). Winter months (December-March) were not monitored as part of this study.

Graph showing total nitrogen and phosphorus yields, by season, for a leaf collection and street cleaning study

Seasonal contribution to annual yields of total phosphorus and total nitrogen (winter excluded) in the control and test catchments during the treatment phase (2015)(Selbig, 2016).

  • The portion of dissolved vs particulate phosphorus shifted from primarily particulate in the spring and summer to dissolved in the fall.
  • During the calibration phase, fall concentrations of dissolved phosphorus in the control and test catchments were 85 percent or more of total phosphorus compared to less than 50 percent in spring or summer. A similar trend was observed during the treatment phase.
  • Like phosphorus, summer concentrations of nitrogen were variable but generally lower than spring and fall. Unlike phosphorus, mean concentrations of nitrogen were highest in the spring and did not have the same spike in fall. Loads of total and dissolved nitrogen were reduced through a combination of leaf removal and street cleaning in spring and fall during the treatment phase. However, the removal of phosphorus was greater than nitrogen. These two observations suggest sources of nitrogen other than leaves and organic detritus may have contributed to what was measured at the outfall.
  • One such source of nitrogen may have been the application of lawn fertilizers which contain nitrogen as a nutrient to stimulate root growth. Lawn fertilizer is typically applied in the spring and fall, which coincided with observed increases in nitrogen concentrations. (Phosphorus from fertilizers was not a concern for this study since the state of Wisconsin enacted a ban on phosphorus in lawn and turf fertilizer in 2009.)

 

CHANGES AS A RESULT OF LEAF COLLECTION AND STREET CLEANING

According to the paired-catchment approach, any change in the relationship that was established between the control and test catchments during the calibration phase of the study can be attributed directly to leaf collection and/or street-cleaning activity. The magnitude of change is a reflection of the amount of organic material removed from streets by the frequency and method of leaf collection and/or street-cleaning. These changes were evaluated using statistical analyses that determine if a change is "significant" (if it's unlikely to be due to chance).

After implementing a street cleaning program, modest reductions were observed in spring for total and dissolved phosphorus and nitrogen. With the exception of total phosphorus, results indicated no significant difference in loads between study phases during summer, meaning any reduction in loads of dissolved phosphorus and total and dissolved nitrogen as a result of street cleaning was negligible. Street cleaning did, however, show some influence on total phosphorus in summer, reducing loads by 36 percent.

Table showing the seasonal mean concentrations of phosphorus and nitrogen from a leaf collection and street cleaning study

Seasonal mean concentrations of phosphorus and nitrogen in the control and test catchment during the calibration and treatment phase (Selbig, 2016).

The addition of leaf collection in the fall significantly reduced loads of all nutrients. Reductions in total and dissolved phosphorus were similar at 84 and 83 percent, respectively. Significant reductions were also observed for total and dissolved nitrogen at 74 and 71 percent, respectively. Fall reductions for both phosphorus and nitrogen were at percentages nearly twice as those observed during spring. The magnitude of the percent change is a reflection of the amount of organic and inorganic material available for wash off during a precipitation event. A larger amount of leaf litter and other organic detritus would produce higher nutrient concentrations. Although this study used methods to remove detritus from streets that are beyond the capabilities of most municipal programs, it represents the upper boundary of achievable reductions in nutrient concentrations, and other municipal leaf collection programs would likely result in lesser reductions.

 

CHANGES IN SEASONAL CONTRIBUTIONS

Without the removal of leaf litter, fall concentrations of phosphorus exceed those in spring or summer. While the 30-year precipitation average shows only 16 percent of annual precipitation occurs typically in October and November, the magnitude of increase in nutrient concentration may produce a greater nutrient load with less stormwater runoff.

Estimates of the total phosphorus yield in 2015 showed that, despite having the fewest number of precipitation events (8), the highest proportion of annual yields in the control catchment occurred during the fall (56 percent). Conversely, spring and summer contributed much lower proportions of annual total phosphorus yield at 14 and 30 percent, respectively, despite having double and triple the number of precipitation events (16 and 25). Comparatively, the inclusion of leaf removal and street cleaning in the test catchment during the fall resulted in only 16 percent of the annual total phosphorus yield, much lower than the control at 56 percent. Little difference in summer yield was observed between the control and test catchments, suggesting the majority of total phosphorus originated from sources other than organic debris on streets and could not be captured by street cleaning. By removing leaf litter in the fall, the seasonal yield of total phosphorus shifted away from being dominated by the decomposition of fall leaf litter to being focused on the frequency of precipitation events. This is an important finding for environmental managers who must evaluate cost-effective strategies to meet pollution reduction goals.

Estimates of total nitrogen yield did not follow the same general pattern as total phosphorus. The differences between mean seasonal concentrations of total nitrogen were not as large as total phosphorus, with fall concentrations only slightly larger than summer but less than spring meaning seasonal yields were mostly controlled by the frequency of precipitation events. Summer, despite having the lowest seasonal mean concentration of total nitrogen, had the largest number of precipitation events (25) which produced the greatest percentage of annual yield (53 percent) in the control catchment. Spring, with fewer precipitation events (16) but a greater seasonal mean concentration, produced 28 percent of the total nitrogen yield in the control catchment followed by fall at 19 percent. Spring and fall percentages were only slightly shifted downward as a result of leaf collection and street cleaning in the test catchment.

 

IMPLICATIONS FOR EFFECTIVE LEAF MANAGEMENT PROGRAMS AND FUTURE STUDY

Understanding the seasonal contribution of nutrients from leaf litter and other organic detritus can help environmental managers assess the most effective way to limit the source and delivery of phosphorus and nitrogen to receiving water bodies. While the results of this study show promising results, leaf management and street cleaning through a municipal program, or combined with modifications to homeowner behaviors, may not necessarily result in similar concentration and load reductions in stormwater runoff. The methods used to remove organic material from streets during this study exceed what most municipal programs are capable of implementing and therefore represent best-case reductions in nutrient concentrations as a result of treatment. 

Research is currently ongoing to quantify the reduction of nutrients in stormwater over a range of commonly used municipal leaf collection programs in Wisconsin. Results from this study will be used to create a framework by which development and adoption of statewide phosphorus credits can be granted to permitted cities that implement a leaf collection program. The results from these evaluations can also be used as incentive for cities to change or amend leaf collection programs as a way to improve the quality of urban stormwater to receiving waters. This research will conclude in 2020.

 

Tree canopy density Leaf collection - method Leaf collection - frequency Street cleaning - method Street cleaning - frequency
High Transfer Biweekly Mechanical broom Biweekly
HighTransfer Biweekly Regenerative air Weekly
High Vacuum Weekly Regenerative air Weekly
Medium Transfer Biweekly Regenerative air Weekly