- clearwatervic.com.au · i preface this report investigates the effectiveness of street sweeping...

EFFECTIVENESS OF STREET SWEEPING FORSTORMWATER POLLUTION CONTROL

TECHNICAL REPORTReport 99/8December 1999

T.A. Walker and T.H.F. Wong

C O O P E R A T I V E R E S E A R C H C E N T R E F O R C A T C H M E N T H Y D R O L O G Y

i

Preface

This report investigates the effectiveness of streetsweeping as a stormwater pollution source controlmeasure. The Cooperative Research Centre forCatchment Hydrology (CRCCH) Project U1 (Grosspollutant management and urban pollution controlponds) focuses on ways to improve the quality ofstormwater runoff. The project covered means toreduce gross pollutants both before and after theyentered the piped stormwater drainage system. Thisreport describes a scoping study to assess theefficiency of Australian street sweeping practices inthe removal of pollutants from street surfaces. Thisstudy has provided information on the effectivenessof street sweeping, currently practiced, in thecollection of pollutants across the range of particlesizes representative of a street surface load.

It is a pleasure to acknowledge the contribution ofTracey Walker and Tony Wong to the UrbanHydrology Program. This work has providedimportant insights into the limited role streetsweeping plays in improving stormwater quality.

Tom McMahonProgram Leader, Urban HydrologyCooperative Research Centre for Catchment Hydrology

COOPERAT IVE RESEARCH CENTRE FOR CATCHMENT HYDROLOGY

Effectiveness ofStreet Sweepingfor StormwaterPollution Control

T.A. Walker and T.H.F. Wong

Cooperative Research Centre for Catchment Hydrology

December, 1999


ii

Executive Summary

Street cleansing is a common (and expensive)practice undertaken by most urban municipalitieswith annual expenditure by a municipality oftenexceeding one million dollars. Street sweeping,essentially the operation of large trucks for cleaningstreet surfaces, is primarily performed for aestheticpurposes. It is, often perceived to lead toimprovements in the environmental conditions ofurban waterways by preventing pollutants depositedon street surfaces from reaching the stormwatersystem. There is, however, little available evidence toquantify the extent to which street sweeping canimprove stormwater quality. This report investigatesthe effectiveness of street sweeping for stormwaterquality improvement.

The effectiveness of street sweeping for stormwaterpollution control is examined for two types ofpollutants, gross pollutants (> 5 mm) and sediment(including associated pollutants). The researchliterature on street cleaning indicates a general dearthof studies that address the issues of gross pollutantmanagement. Most studies predominantly examinethe effectiveness of street sweeping for sediment andassociated contaminant removal. This study looks atthe effectiveness of street sweeping for grosspollutants using the results of Australian field studies,while sediment and other suspended solid removal isinvestigated with interpretation of results fromoverseas studies.

Experimental studies overseas found street sweepingto be highly effective in the removal of large solidsgreater than 2 millimetres under test conditions.However, field conditions are expected tosignificantly reduce the efficiency of solid removalbecause of limitations with sweeper access to sourceareas (mainly due to street design and car parking),sweeping mechanisms used and operator skills. Fieldstudies undertaken by the Cooperative ResearchCentre for Catchment Hydrology (CRCCH) inAustralia found significant stormwater gross pollutantloads generated from source areas in spite of a dailystreet sweeping regime.

An earlier CRCCH study, involving analysis of grosspollutant loads from a 50 hectare urban catchment ofmixed residential, commercial and industrial land-use, found a clear relationship between the grosspollutant load in the stormwater system and themagnitude of the storm event. The shapes of thecurves relating gross pollutant load to event rainfalland runoff were found to be monotonically increasingand representable by a logarithmic function. Theshape of these curves suggests that the limitingmechanism affecting the amount of gross pollutantsentering the stormwater system is rainfall dependent(ie. the available energy to re-mobilise and transportdeposited gross pollutants on street surfaces) ratherthan being source limiting (ie. the amount of availablegross pollutants deposited on street surfaces).

Overseas studies indicate that street sweeping isrelatively ineffective at reducing the street surfaceload of fine particles (below 125 µm). The particlesize distribution of suspended solids conveyed instormwater in Australian conditions typically rangefrom 1 µm to 400 µm with approximately 70% of theparticles smaller than 125 µm. Therefore, streetsweeping as it is currently practiced cannot beexpected to be effective in the reduction of suspendedsolids and associated trace metals and nutrientconcentrations in stormwater.

The study concludes that the performance of streetsweeping for stormwater pollutant control is limitedand must be accompanied by structural pollutanttreatment measures to effectively reduce thedischarge of gross and sediment associated pollutantsin stormwater. The incremental benefits in increasingthe frequency of street sweeping beyond what isrequired to meet street aesthetic criterion is expectedto be small in relation to water quality improvements.As a result, there seems little benefit in conducting anin-depth field-based study into the effectiveness ofstreet sweeping for stormwater pollution control.


iii

Acknowledgements

The authors of this report would like to acknowledgethe continued assistance and support provided by theMoreland City Council, particularly the cleansingmanager, Les Horvath for sharing his knowledge ofstreet cleaning operations. Many thanks to all thosewho were interviewed for Council information,especially Peter Barton from the City of GreaterDandenong who also arranged participation in anearly morning street sweeping program.

In addition, Ranger Kidwell-Ross from AmericanStreet Sweeper magazine is thanked for his help anddirection, and Roger Sutherland for forwardingcurrent literature from America. The authors alsowish to acknowledge in particular Dr Robin Allisonfor his advice and discerning suggestions, and DrFrancis Chiew for his comments and input.

iv


Preface i

Executive Summary iii

Acknowledgements iv

1 Introduction 1

2 Background 3

2.1 Street Sweeping Pollutant Removal Monitoring 3

2.2 Modelling Sweeper Pollutant Removal Efficiencies 3

2.3 Factors Influencing Street Sweeping Effectiveness 4

3 Methodology 5

4 Melbourne Street Sweeping Practices 7

4.1 Street Sweeping Operations 7

4.2 Target Pollutants 7

4.3 Contracts and Sweeping Frequency 7

4.4 Council Perspective of Effectiveness 9

5 Sweeping Mechanism 11

5.1 Types of Sweeping Mechanisms 11

5.2 Sweeper Effectiveness 11

6 Pollutant Types 13

6.1 Gross Pollutants 13

6.2 Sediment and other Suspended Solids 15

6.3 Contaminants Associated with Sediment 16

6.4 Australian Conditions 19

7 Sweeping Frequency and Timing 21

7.1 Sweeping Frequency and Rainfall Patterns 21

7.2 Inter-Event Dry Period 21

7.3 Street Sweeping Timing 24


v

8 Gross Pollutant Wash-off Characteristics 25

8.1 Gross Pollutant Load Generation 25

8.2 Influence of Catchment Land-use 28

9 Discussion 35

9.1 Gross Pollutant Load and Rainfall

Depth Relationship 35

9.2 Impacts of Street Sweeping on Gross

Pollutant Loads 35

9.3 Supply Limiting Condition 36

9.4 Street Sweeping Efficiency Issues 36

10 Conclusions 39

References 41

1 Introduction

This report presents the findings of an investigationon the effectiveness of current Australian streetsweeping practices in the collection of pollutantsacross the typical range of particle sizes found onstreet surfaces. The study was initiated to define andscope a further more-detailed field-based study toquantify the effectiveness of current street sweepingpractices as an at-source stormwater pollutionmanagement measure. The term street sweeping isused here to describe essentially the operation of largetrucks to remove deposited litter and debris from thekerb and channel of major roadways, streets, andcarparks. The study examines the effectiveness ofstreet sweeping practices to remove pollutants of twotypes:- (i) gross pollutant and litter removal and (ii)sediment and associated contaminant removal.

Over the past decade there has been an increase in themanagement of urban stormwater to protect urbanwaterways and receiving waters. These initiativeshave, in part, resulted from community awareness ofenvironmental impacts of urban stormwater pollutionand their expectation that urban aquatic ecosystemsshould be protected from further environmentaldegradation.

Pollutants generated from urban land-use activitiesare transported by stormwater to urban receivingwaters. Pollutants washed off street surfaces includegross pollutants, sediment and associated metals,nutrients, hydrocarbons and dissolved pollutants.Increased volumes of stormwater runoff anddischarge rates resulting from increased impervioussurface areas and hydraulically efficient drainageinfrastructure throughout urban catchments havemeant that the transport of urban pollutants toreceiving waters is particularly efficient.

Most urban metropolitan councils perform cleansingof streets and similar impervious surfaces. This iscommonly for the purpose of controlling grosspollutants, particularly litter, to maintain a level ofstreet cleanliness and aesthetic quality. The focus onenvironmental issues is growing and local authoritiesare now considering street sweeping as a beneficial

at-source method for reducing the amount of streetborne pollutants entering the stormwater system. Theactual contribution of street sweeping to theabatement of stormwater pollution is however notwell understood. The objectives of street sweepingfor street aesthetics and stormwater pollution controlare very different, with the former placing particularemphasis on the visual impact of environmentalpollution while the latter encompasses a much widerrange of pollutant types and sizes. Despite streetsweeping being widely considered an at-sourcestormwater pollution control method its effectivenessis unknown.

This report undertakes an interpretation of relevantstreet sweeping literature, research and survey results.The background to street sweeping operations,focusing on the effectiveness of sweeping for removalof street surface pollutants, is established in Section 2.The methodology undertaken for this investigation isdiscussed in Section 3. Results from a survey of 21Melbourne Metropolitan councils on street sweepingpractices are assessed in Section 4, to establish anunderstanding of current operations, target pollutantsand sweeping frequencies. The different types ofstreet sweeping mechanisms and their measuredeffectiveness are examined in Section 5. Pollutanttypes found on street surfaces are reviewed in Section6, including an analysis of Australian sedimentcharacteristics to assess the influence of streetsweeping practices on fine particulates and associatedcontaminants.

Inter-event dry periods can influence street sweepingeffectiveness and these are determined usingAustralian rainfall statistics in Section 7, andcompared with current sweeping frequency andtiming information. Section 8 examines field data todetermine gross pollutant load generation and theinfluence of catchment land-use and associatedsweeping frequency on pollutant load. The impactstreet sweeping has on gross pollutant loads enteringthe stormwater drainage system is discussed inSection 9, highlighting important issues affectingcurrent sweeping efficiencies. Section 10 concludeswith a summary of specific observations from each ofthe sections of the report from which the effectivenessof street sweeping as a stormwater pollution controlmethod is assessed.


1


2

sweeping can significantly reduce pollutant washofffrom urban streets due to the improved efficiencies ofnewer technologies now employed to conduct streetsweeping in some American states. Theirinvestigations showed that when street sweepingmechanisms and programs are designed to removefiner particles (ie. small-micron surface cleaners ortandem sweeping) it can benefit stormwater runoffquality.

2.2 Modelling Sweeper Pollutant Removal Efficiencies

Sweeping technologies with the ability to effectivelyremove accumulated sediments, including fineparticles, may significantly increase the efficiency ofsweeping for the removal of a variety of stormwaterpollutants. Sutherland and Jelen (1993) described theuse of a calibrated version of the Simplified ParticleTransport Model (SIMPTM) as being able toaccurately simulate the complicated interaction ofaccumulation, washoff, and street sweeper removalthat occurs over a time period. For varying streetsweeping operations Sutherland and Jelen (1997)employed the SIMPTM to predict the average annualexpected reduction in total suspended solids (TSS) attwo sites in Portland, Oregon. Sweepers used in theirsimulations included the NURP era broom sweeper, amechanical broom sweeper, a tandem operationinvolving a mechanical broom followed by a vacuumsweeper and a newer technology, the small-micronsweeper. The predicted reductions in TSS showedthat all of the newer street sweeping technologies aresignificantly more effective than the NURP era broomsweeper. It was further concluded that new streetsweeping technologies designed for effective removalof fine particles, are capable of removing significantsediment loads and associated pollutants from urbanstreet surfaces.

In a further study Sutherland and Jelen (1998)compared the new small-micron street sweepingtechnology to wet vaults, a widely used stormwaterquality treatment method. The ability of the small-micron street sweeper to achieve significantreductions in urban pollutant washoff led Sutherlandand Jelen to consider it an effective Best ManagementPractice (BMP) for stormwater pollution control.


3

2 Background

2.1 Street Sweeping Pollutant RemovalMonitoring

The role and usefulness of street sweepers to controlstreet surface pollutants was first investigated in thelate 1950’s and early 1960’s by the United StatesEnvironmental Protection Agency (US-EPA) and itsassociated researchers. Many of the US-EPA’sNational Urban Runoff Program (NURP) studiesmeasured the efficiency of street sweeping as astormwater pollution control method with particularemphasis placed on sediment and sediment-boundcontaminants.

Since the late 70’s studies have measured streetsweeping effectiveness in terms of the reduction inend-of-pipe runoff pollution concentrations and loadsrather than assessing the effectiveness of specificequipment. Sartor and Boyd (1972) found sweepingschedules based on a seven day cycle to be almosttotally ineffective while daily sweeping was shown topotentially have a high level of pollutant removal forlarger sized pollutants typical of street surfacematerial (Sartor and Gaboury, 1984). Pitt andShawley (1982) and Bannerrnan et al. (1983)concluded that only minor benefits to stormwaterquality are provided by street sweeping practices.However, Terstrierp et al. (1982) and Pitt andBissonette, (1984) demonstrated that street sweepingcollects significant amounts of particles, for selectparticle size ranges, from street surfaces. The overallconclusion reached by the US-EPA, was that, as awater quality best management practice, streetsweeping did not appear to be effective at reducingend-of-pipe urban runoff pollutant loads.

Subsequent investigations into the effectiveness ofstreet sweeper mechanisms for water qualityimprovement report findings that vary to thosepresented in the conclusions of the earlier NURPstudies. Alter (1995) and Sutherland and Jelen(1996b) assert that the NURP studies concluded thatstreet sweeping is largely ineffective, because thesweepers used at the time of these studies were notable to effectively remove very fine accumulatedsediments which are often highly contaminated.Sutherland and Jelen (1996a) suggest that street

2.3. Factors Influencing Street SweepingEffectiveness

The pollutant reduction effectiveness of any streetsweeping operation is dependent on the equipmentused and the environmental and geographicconditions (eg. wind and presence of parkedvehicles). Unless other influential factors (such asstreet parking) are addressed, the efficiency ofindividual sweeping mechanisms can be a relativelyinsignificant factor in the overall effectiveness ofstreet sweeping operations. It is anticipated that theeffectiveness of street sweeping programs dependmore on factors such as land-use activities, the inter-event dry period, street sweeping frequency andtiming, access to source areas and sweeper operationthan the actual street sweeping mechanism. Thesefactors all influence the deposition, accumulation andremoval rates of pollutants on street surfaces.Physical features such as the degree of catchmentimperviousness and the hydraulic characteristics ofstreet surfaces can also influence the effectiveness ofstreet sweeping. These factors require considerationbefore a thorough assessment of street sweepingefficiency for stormwater pollution control can beachieved.


4

other major capital cities in Australia. Melbourneinter-event periods were compared to the surveyedresults of typical sweeping frequency and timing toinvestigate likely sweeper performance. Thisinformation facilitates a “hydrological basis” forselecting a street sweeping frequency that wouldoptimise gross pollutant removal.

The study also examines data obtained from fieldstudies previously undertaken by the CRC forCatchment Hydrology and others to investigate theeffectiveness of street sweeping on litter and grosspollutant removal. Gross pollutant load data gatheredat 192 side entry pit traps (SEPTs - baskets fitted intoroadside stormwater entry pits) in the suburb ofCoburg in Melbourne by Allison et al. (1998) weregrouped according to the street sweeping frequenciesin their respective streets. Similar data are availableat two further study catchments in the suburbs ofCarnegie and McKinnon in Melbourne (Hall andPhillips, 1997). The load data captured by the SEPTsduring a typical street sweeping program are used toevaluate the amount of gross pollutants typicallyentering the stormwater system under normalMelbourne street sweeping frequencies andconditions. While it was not possible to compute ameasure of pollutant removal efficiency owing to aninability to account for pollutants by-passing theSEPTs, the data nevertheless provided an insight onwhat might be the expected gross pollutant exportload from streets that are swept at regular intervals.


5

3 Methodology

This study assesses the effectiveness of streetsweeping for stormwater pollution control by:

� reviewing previous studies on sweeperperformances and street pollutant characteristics,

� reviewing objectives for street sweepingoperations (eg. aesthetic),

� considering rainfall distributions with streetsweeping frequency and timing to investigatelikely sweeper performance,

� examining field data from an earlier CRC studyand others on gross pollutants,

� investigating the potential effects of changingstreet sweeping regimes on the gross pollutantloads in stormwater.

This study interprets available Australian andoverseas field data on the measured efficiencies ofstreet sweeping and street surface sediments. Variousstudies describing the particle size distribution ofsediment loads were also collated to provide aninsight into the particle size distribution pattern ofsuspended solids typical of street surface runoff.Some significant overseas studies on the partitioningof sediment sizes and the contaminant associations(eg. metals and nutrients) with each particle sizepartition were used to assess the pollutants likely tobe discharged into the stormwater system from streetsurfaces. Information regarding street sweepingefficiencies and sediment contaminant associationsfrom these studies are combined with data onAustralian stormwater suspended solidscharacteristics to enable an assessment of streetsweeping practices on removal of fine particulateassociated pollutants.

A survey of street sweeping practices amongstmunicipalities in Melbourne was carried out toexamine current sweeping objectives, procedures andmechanisms in these municipalities. This survey wasalso used to determine the perceived effectiveness ofstreet sweeping in maintaining a certain standard ofstreet aesthetics. Australian rainfall distributionswere then examined and used to assess typicalstatistics of inter-event dry periods for Melbourne and


6

4 Melbourne Street Sweeping Practices

4.1 Street Sweeping Operations

The responsibility of keeping urban streets clean,commonly by sweeping road surfaces with largevacuum trucks is an operation carried out by localgovernment. A survey of 21 Melbourne metropolitancouncils was performed to determine the motivationfor the large expenditure on street cleaning. Theresults indicated that street sweeping is primarilyundertaken for aesthetic purposes in response tocommunity expectations. Table 4.1 summarises thestreet sweeping practices of 21 municipalities inMelbourne.

4.2 Target Pollutants

Street cleansing programs are generally designed toconcentrate on collecting human derived litter toaddress the obvious visual impacts. However, duringautumn, organic matter becomes a focus and thesweeping frequency is altered to reduce the safetyhazard associated with decomposing leaf litter onstreet surfaces and to reduce drain blockages. Streetsurface sediment collection was not identified as amajor issue when designing street sweepingprograms.

Street cleansing programs involve what is oftentermed ‘building line to building line’ cleansing,incorporating footpath cleaning, and the standard kerband channel street sweeping where it is apparent alarge proportion of litter accumulates. This requires acombination of cleansing methods and equipment forthe successful removal of such pollutants. Australianstreets are cleaned customarily with large truckmechanical broom and vacuum systems. However, itis becoming common practice to operate smallerbroom and vacuum sweepers designed for cleansingareas inaccessible to the traditional larger plants. Themost commonly used sweepers are the regenerativeair model, for both large truck and small plantsystems.


7

4.3 Contracts and Sweeping Frequency

Under new competitive tendering legislation, thebidding process for street cleansing contractsestablishes a requirement for operators to becomevery competitive. Contractor performance ismeasured against output based specifications set bythe council. This means the council stipulates a set ofcleanliness requirements they wish to achieve with astreet cleansing program but not the frequency oroperation methods used. Street sweeping practicestherefore differ considerably between Melbournemetropolitan councils. Street sweeping frequenciescan range from every two weeks to every six weeks inresidential areas and from daily to every two weeks incommercial areas. Shopping centres and commercialareas are swept more frequently, typically rangingfrom once or twice a day in busy areas and once ortwice a week in less popular areas. Street sweepingfrequencies for residential areas range from once aweek for highly populated areas to every six weeks inless populated areas.

CO

OPE

RA

TIVE

RE

SE

AR

CH

CE

NTR

E FO

R C

ATC

HM

ENT H

YD

RO

LOG

Y

8

Table 4.1 Street Sweeping Practices for Melbourne Municipalities

COUNCIL PURPOSE TARGET CONTRACT FREQUENCY SWEEPING COUNCILPOLLUTANT MECHANISM PERSPECTIVE

Bayside:Hobsons Bay Aesthetic Litter / Leaves Internal (3-5yrs) 1 day 4 weeks Regenerative EffectivePort Phillip H&S / SW / CD Litter / Leaves Internal (3-5yrs) 1 day 2 weeks Regenerative Effective

Bayside SW / aesthetics Litter / Leaves Internal (3-5yrs) 1 day 3 weeks Regenerative EffectiveKingston SW / aesthetics Litter / Leaves External (3-5yrs) 1 day 5 weeks Regenerative Effective

Inner City:Banyule

Boroondara

Glen EiraManninghamWhitehorseStonnington

Moonee ValleyMelbourne CityMaryibynong

MonashMoreland

Outer City:Brimbank

HumeGreater Dandenong

Knox CityMoroondahNillumbik

Note: Councils not listed were conducting tender negotiations for street sweeping practices during the time of the survey.H & S = Health and SafetySW = Stormwater QualityCD = Community Demand

Amenity / SWAesthetics / H&S / SW

CDAmenity / SW / CD

Aesthetics / SWAmenity / Aesthetics

SWCD / amenity

CD / aestheticsAesthetic / CD / SWCD / aesthetics / SW

Litter / LeavesLitter / Leaves

Litter / LeavesLitter / LeavesLitter / LeavesLitter / LeavesLitter / LeavesLitter / LeavesLitter / Leaves

LitterLitter / Leaves

Internal (3-5yrs)Internal (3-5yrs)

External (3-5yrs)Internal (3-5yrs)Internal (3-5yrs)Internal (3-5yrs)Internal (3-5yrs)External (3yrs)

Internal (3-5yrs)Internal (3-5yrs)Internal (3-5yrs)

2 weeks3-7 days

1-3 days1 day1 day1 day1 day1 day1 day1 day1 day

5 weeks4 weeks

4 weeks 6 weeks 3 weeks

1-2 weeks6 weeks2 weeks2 weeks6 weeks2 weeks

RegenerativeRegenerative

RegenerativeRegenerativeRegenerativeRegenerativeRegenerativeRegenerativeRegenerativeRegenerativeRegenerative

EffectiveEffective

Not EffectiveEffectiveEffectiveEffectiveEffective EffectiveEffective

Not EffectiveEffective

Litter / LeavesLitter

Litter / LeavesLitter / Leaves

LitterLitter / Leaves

SW / CDAmenity / SWCD / amenity

SW / CDCD / amenity / SW

Aesthetic / SW

Internal (3-5yrs)Internal (3-5yrs)Internal (3yrs)

Internal (3-5yrs)Internal (3-5yrs)Internal (3yrs)

1 day1 day1 day2 days1 day

1-2 weeks

5 weeks4 weeks17 days5 weeks21 days4 weeks

RegenerativeRegenerativeRegenerativeRegenerativeRegenerativeRegenerative

EffectiveEffectiveEffectiveEffectiveEffectiveEffective

Commercial Residential


9

4.4 Council Perspective of Effectiveness

All but two councils indicated that street sweeping asit is currently practiced was an effective way ofcollecting litter. Numerous councils stated that streetsweeping aided in the prevention of litter entering thestormwater system and therefore reduced theoccurrence of stormwater pollution and drainblockage but had no data to validate theseobservations. Several councils regarded streetsweeping as effective only when practiced inconjunction with other source pollution controlmethods such as bins, side entry pit traps and othergross pollutant traps.

Overall the survey indicated a general satisfactionwith the effectiveness of street sweeping in collectinghuman derived litter and organic matter (grosspollutants) for aesthetic objectives. However, there islittle quantitative information for councils to assessthe effectiveness of street sweeping practices onstormwater pollution reduction. Throughout theliterature there are many suggestions that streetsweeping can have an effect on stormwater qualityalthough the degree to which this practice is effectiveis unknown.

The assessment of the effectiveness of streetsweeping in stormwater pollution control rather thanjust aesthetic requirements will need a detailedanalysis of the following major influencing factors.

� street sweeping mechanism

� pollutant types (from sediment and associatedcontaminants to gross pollutants)

� sweeping frequency & timing

� pollutant load wash-off characteristics

Each one of these factors is examined in detail in thefollowing sections of this report.


10


11

5 Street Sweeping Mechanisms

5.1 Types of Sweeping Mechanisms

Types of street sweeping mechanisms commonlyutilised in Australian practice include:

1. Mechanical broom sweepers involving a numberof rotating brushes sweeping litter into acollection chamber;

2. Mechanical broom and vacuum systems involvingthe combination of rotating brushes and a vacuumto remove street litter;

3. Regenerative air sweepers which are likemechanical vacuum sweepers but use recirculatedair to blast the pavement, dislodging litter beforeit is swept by rotating brushes towards a vacuumfor pick-up. This sweeper also uses water spraysfor dust suppression,

4. Small-micron surface sweepers which combinerotating brooms enclosed in a powerful vacuumhead in a single unit, performing a drysweeping/vacuuming operation. A powerful fanpulls debris and air into a containment chamberbefore the air is finally passed through a series offilters to capture small micron material.

5.2 Sweeper Effectiveness

Pitt and Bissonnette (1984) found following a periodof street sweeping trials that street sweepingequipment was unable to remove particles from thestreet surface unless the loadings were greater than acertain threshold amount. This value was found to bethree times higher for a mechanical broom cleaner,most referred to in the US-EPA’s NURP studies,compared to the regenerative air street sweeper trialedfor a comparison in a study by Pitt and Bissonnette(1984). The study found the regenerative air vacuumsweeper to exhibit a substantially better performancethan the regular mechanical street sweeper, especiallyfor the smaller particle sizes. Such findings haveprogressively led to the mechanical broom methodbeing replaced by the vacuum system method forstreet sweeping practices. The removal effectivenessdata for the smallest particle sizes (less than 125 µm)between the two methods of street sweeping washowever found to be inconclusive.

The regenerative air vacuum sweeper (Figure 5.1) is acommon mechanism used for street sweeping inAustralia. The recirculating air cycle tends toimprove the effectiveness of sweepers for the removalof heavy debris but is less effective for removing finesediment. The air blast is able to dislodge heavier

Figure 5.1 Australian streets are cleaned with large truck vacuum sweepers

materials and propel them into the vacuum airflowhowever finer materials often remain uncollected (Pittand Bisonnette,1984). Fine particles may becomeairborne as a result of the air blast and take some timeto settle back onto the road surface or may be leftbehind on the street surface.

The most recent technology to be employed for streetsweeping is a highly effective, vacuum-assisted drysweeper (the small-micron surface sweeper)originally developed and manufactured by EnviroWhirl Technologies Inc in the United States ofAmerica. The sweeper was originally developed forthe containment of spilled coal dust along railwaytracks. This system is reported to be extremelyeffective in removing fine street surface sedimentsand preventing their escape into the air by filtering airemissions down to sizes as small as 4 µm. Sutherlandand Jelen (1997) described this system as having anadvanced ability, when compared to other sweepingmechanisms, to remove a broad range of particlesfrom road surfaces down to sub micron particulates.The small-micron surface cleaning technology hasbeen shown by Sutherland and Jelen (1997) to havetotal removal efficiencies ranging from 70% forparticles less than 63 µm up to 96% for street surfacepollutants larger than 6370 µm.

Despite there being new street sweeping technologiesreported to be more efficient, most municipalities andprivate street sweeping companies in Australiacontinue to use the mechanical broom andregenerative air vacuum street sweepers. This isbecause of the high capital costs of newertechnologies and their limited availability on theAustralian market.

Street Sweeping Mechanism:

� Mechanical and regenerative air street sweeping equipment requires a minimum threshold load ofsediment on the street surface before they become effective.

� The threshold load can be three times higher for the mechanical sweeper compared to the regenerative airsystem.

� Overall the regenerative air sweeper exhibits a substantially better performance than the regularmechanical sweeper.

� Street sweeping technology is developing and improving to remove finer street surface particles for avariety of street surface loads,


12


13

pollutants transported by stormwater to include litter(predominantly paper and plastics) and vegetation(leaves and twigs) as shown in Figure 6.1. Organicmatter comprised the largest proportion by mass ofthe collected gross pollutants and therefore should bea major consideration in street cleaning programs.The data was based on field monitoring of grosspollutants retained in a Continuous DeflectiveSeparation (CDS) unit treating a catchment area of 50hectares in Coburg, an inner city suburb ofMelbourne.

Only a small number of investigations have examinedstreet sweeping effectiveness on gross pollutantremoval. Nilson et al. (1997) conducted aninvestigation into source control of gross pollutants inAdelaide and attempted to assess the efficiency ofstreet sweeping for gross pollutant removal instormwater. This study sought to quantify the amountof gross pollutants entering the drainage network inthree similar streets swept at different intervals.Catch baskets in side entry pits were used to collectgross pollutants which were not otherwise collectedby the sweeper for a street swept every day, once aweek, and not at all. Trapped pollutants in these

6 Pollutant Types

The effectiveness of street sweeping to removepollutants, across the typical range of particle sizesfound on street surfaces, has not yet been successfullyquantified for Australian conditions. The examinationof street sweeping effectiveness in the present studyfocuses on two pollutant types:- (i) gross pollutantsand litter and (ii) sediment and associatedcontaminants. Gross pollutants have been defined asany solids that are retained by a 5 mm mesh screen byAllison et al. (1998) and this definition is adoptedhere. Solids washed off street surfaces which aresmaller than 5 mm and not considered to be grosspollutants include a proportion of litter and organicmatter but are predominantly sediment particles,typically between the course sand to fine silt range,and sediment associated contaminants.

6.1 Gross Pollutants

Allison et al. (1997a) undertook an investigation intothe types of gross pollutants derived from an urbancatchment. The study found typical urban gross

Figure 6.1 Composition of Gross Pollutants by Mass (Allison et al., 1998)

Vegetation

OthersMetalsCommercialPlastic

PersonalPaper

PersonalPlastic


14

baskets were removed and quantified weekly duringthe study.

The results of the study by Nilson et al. (1997) showlittle correlation between the frequency of sweeping,rainfall or wind-run in the catchment with the grosspollutant load collected in the catch baskets. Thestudy provided little conclusive information on theeffectiveness of street sweeping with respect to grosspollutants. The study found that typically, asignificant amount of gross pollutants were mobilisedinto the stormwater system from the street duringbursts of rain, wind or both, irrespective of the natureof the street sweeping program implemented. Theseresults suggest the amount of gross pollutants or streetsurface load does not limit the amount transportedinto the stormwater system regardless of the streetsweeping frequency.

The observed composition of the gross pollutantmaterial collected by Nilson et al. (1997) wasconsistent with other studies conducted by Sartor andBoyd (1972), O’Brien (1994) and Allison and Chiew(1995), where gross pollutant loads measured in drymass comprised approximately 70-90% organicmatter, and 10-30% litter.

Broad-based investigations into street sweepingconducted by the US-EPA suggest that streetsweeping efficiency increases with particle size.Sartor and Boyd (1972) found sweeper efficiency tobe nearly 80% for the collection of particles greaterthan 2 mill imetres under ‘test’ conditions (ie.sweeping more frequently than the occurrence ofrainfall events and effective use of parkingrestrictions). Ideal street cleaning conditions areunlikely to occur during normal street sweepingoperations, and sweeper efficiencies for collectinggross pollutants would be expected to be considerablylower than the recorded 80% despite anyimprovements gained through refinements ofequipment since the study. In practice, theeffectiveness of street sweeping for gross pollutantremoval is influenced by a number of factorsincluding: access to the street load, operator skills andsweeping speed, sweeping mechanism, time of daysweeping is conducted and weather conditions.

Gross Pollutants:

� Typical urban gross pollutants transported by stormwater include litter (predominantly paper and plastics)and vegetation (leaves and twigs).

� Significant amounts of gross pollutants are mobilised into the stormwater system during bursts of rain,wind or both.

� There is little correlation between the frequency of sweeping and the transport of gross pollutants into thestormwater system.

� Street sweeping efficiency increases with particle size.

� Sweeper efficiency can be up to nearly 80% for particles greater than 2 millimetres under ‘test’ conditions(ie. Sweeping more frequently than the occurrence of rainfall events and effective use of parkingrestrictions).


15

6.2 Sediment and Other Suspended Solids

Street sweeping performance for smaller streetsurface particles depends considerably on the type ofstreet sweeper used and also conditions such as thecharacter of the street surface (texture, condition andtype), street dirt characteristics (loadings and particlesizes), and other environmental factors (Pitt andBissonnette, 1984).

Sartor and Boyd (1972) found the removalefficiencies of sediment by conventional streetsweepers to be dependent upon the particle size rangeof the street surface loads as shown in Figure 6.2.Mechanical sweeper efficiency was found to begenerally low for fine material. This finding wassupported by two further studies conducted by Benderand Terstriep (1984) and Pitt and Bissonnette (1984),who reported that the proportion of the total streetload smaller than 300 µm was less affected by streetsweeping. Pitt and Bissonnette (1984) alsodemonstrated that no effective removal was evidentfor street dirt particles smaller than about 125 µm forthe regenerative air sweeper.

Mechanical broom sweepers are found to be effectiveat collecting larger particles but less effective thanregenerative-air vacuum sweepers in removing thesmaller particles (Pitt and Shawley, 1982). Theregenerative air vacuum sweeper, although regardedas more effective at collecting smaller particle sizesdoes not successfully control or remove fine particles.

Problems are encountered with water-based dustsuppression methods as they tend to resuspend thesmall micron particles and their associated attachedpollutants, forming a slurry which either fills thecracks in the pavement or is discharged into thestormwater system. Similarly, fine particles caneasily escape collection when they are re-mobilisedinto the air by the pavement blast used by theregenerative air sweeper to dislodge larger materials.

Studies by Pitt and Sutherland (1982) indicated that asignificant proportion of the larger dirt particle sizespicked up by street sweepers are not easilytransported by rain and that removal of these particlestends to expose the smaller sheltered particles. Thesesmaller particles exposed by street sweeping are thenmore readily mobilised and transported into thestormwater drainage system during rainfall events.The small-micron surface sweeper sweeps dry, withno water being used, and thus overcomes problemsassociated with resuspension of fine particulates andassociated pollutants by dust suppression sprays.These machines util ise strong vacuums incombination with uniquely-designed main and gutterbrooms. The air filtration system, enables smallerparticles to be removed from the street surface withthe return of clean air to the atmosphere (ie. filtersparticles down to 2.9 microns). This relatively newtechnology is regarded to be a high-efficiencysweeper (Sutherland et al., 1998).

0

2 0

4 0

6 0

8 0

1 0 0

0 5 0 0 1 0 0 0 1 5 0 0 2 0 0 0 2 5 0 0

Particle Size (microns)

Figure 6.2 Street sweeping efficiency as a function of particle size (Sartor and Boyd, 1972)

Swee

per E

ffici

ency

(%)

Particle Size (µm)


16

The removal performance of street sweepers forsediment has been often determined from samplingaccumulated street dirt before and after sweeping hasbeen conducted. Initial street surface conditions areestablished and the street swept at a specified speedof 7-8 kilometres per hour before it is sampled toestablish the residual condition. The differencebetween initial and residual loadings by specificparticle size defines the removal performance ofstreet sweeping operations. It was concluded fromthis method that sweeping removes little, if any,material below a certain threshold. This thresholdload was found to vary by particle size range. Aseries of mathematical equations developed by Pitt(1979) to describe this removal performance havebeen recently calibrated and employed by Sutherlandand Jelen (1996a and 1997) to evaluate and comparethe removal performance of numerous streetsweeping technologies.

Sutherland and Jelen (1997), using their SimplifiedParticle Transport Model, tested the removalperformance of the small-micron sweeper, along witha regenerative air vacuum sweeper, a mechanicalbroom sweeper, and a tandem operation that involveda single pass by a mechanical broom followed by avacuum sweeper. The small-micron sweeper wasshown to be the most efficient, with average totalremoval efficiencies of 70% for particles less than 63µm and between 77% and 96% for particle sizesranging from 125 µm to larger than 6370 µm. The

small-micron sweeper demonstrated an ability toefficiently remove particles without any thresholdlevel unlike the other sweepers tested. Theregenerative air sweeper was shown to be the secondmost efficient with overall removal efficienciescalculated to range from 32% for less than 63 µmrange to 100% for larger particles between 600 and2000 µm. However, the removal efficiency of theregenerative air sweeper for particles between 250and 2000 µm can drop to zero, due to the necessity oflarge threshold loads for particles within this sizerange. The tandem operation and mechanical broomsweeper were found to be the least efficient despitesome recorded high efficiencies. This can be mainlyattributed to the high threshold loads required bythese operations before any significant sedimentremoval is recorded.

6.3 Contaminants Associated with Sediment

It is well recognised that a significant amount ofmetals and nutrients are transported as sediment-bound contaminants. Many investigations have foundthe concentration of sediment-bound contaminants tovary with particle size, with high concentrations ofcontaminants attached to the finer particles (Sartor &Gaboury 1984, Sartor & Boyd 1972). Hvitved-Jacobsen et al. (1991, 1994) investigated road runoffpollutant characteristics and found 60-80% ofphosphorous, 30-40% of zinc, 70-80% of lead, 30-40% copper and about 55% of total nitrogen in road

Sediment and Other Suspended Solids:

� The removal efficiency of sediment and other fine organic particles by conventional street sweepers wasfound to be dependent upon a threshold level of load on the surface and the particle size range of thesurface loads.

� Material smaller than 300 µm was less affected by street sweeping.

� No effective removal (>50% removal efficiency) was evident for particle sizes smaller than 125 um forconventional street sweepers (excluding the new small-micron surface cleaning technology).


17

runoff to be associated with particulates. While mostparticulate matter found on street surfaces is in thefractions of sand and gravel. Approximately 6% ofparticles are in the silt and clay soil size and theywere found to contain over half the phosphorous andsome 25 percent of other pollutants, as indicated inTable 6.1, adapted by Shaver (1996) from results ofSartor et al. (1974).

Many other investigations have found theconcentrations of sediment-bound contaminants instreet dirt to be associated with the fine particle sizefraction. Pitt & Amy (1973), NCDNRCD (1993) andWoodward-Clyde (1994) have all shown that higherconcentrations of pollutants such as heavy metals areassociated with the smallest particle size fractions ofurban dust and dirt. These data indicate that almosthalf of the heavy metals (represented by copper, leadand zinc) found on street sediments are associatedwith particles of 60 to 200 µm in size and 75% areassociated with particles finer than 500 µm in size.Dempsey et al. (1993) undertook an analysis ofparticle size distributions for urban dust and dirt, andpartitioning of contaminants into a number of sizefractions to determine the concentrations ofcontaminants in each particle size range. Results

show the highest recorded concentrations of Cu, Znand TP to be associated with sand particles between74 and 250 µm in size.

Colwill et al. (1984) found 70% of oil andapproximately 85% of polycyclic aromatichydrocarbon (PAH) to be associated with solids in thestormwater. That study demonstrated that over aperiod of dry weather conditions, increasingproportions of oil become solid associated where thehighest oil content was found in sediments of 200 to400 µm in size.

Sansalone et al. (1997), Fergusson and Ryan (1984),Baker (1980) and Wilber and Hunter (1979) allreported that heavy metal concentrations increasewith decreasing particle size. Results presented bySansalone et al. (1997) from particle size distributionand metal analysis indicate that zinc, copper and leadconcentrations increase with decreasing particle sizeor, equivalently, increasing specific surface area. Theabsorption of contaminants to particles is oftenregarded as being directly related to the surface areaper unit mass available for ion absorption. Measuredspecific surface area results presented by Sansalone etal. (1997) indicated that the assumption of smoothspherical particles to estimate available surface area

Particle Size (Microns)

Pollutant <43 43 - 104 104 - 246 246 - 840 840 - 2000 >2000

Total Solids 5.9 9.7 27.8 24.6 7.6 24.4

Volitile Solids 25.6 17.9 16.1 12.0 17.4 11.0

COD 22.7 45.0 12.4 13.0 4.5 2.4

BOD 24.3 17.3 15.2 15.7 20.1 7.4

TKN 18.7 19.6 20.2 20.0 11.6 9.9

Phosphates 56.2 29.6 6.4 6.9 0.9 0.0

All Toxic Metals 27.8 - 23.5 14.9 17.5 16.3

(Source: Shaver; 1990; adapted from Sartor, Boyd, and Agardy, 1974)

Table 6.1 Percentage of Street Pollutants in Various Particle Size Ranges

Particle Size (µm)


18

grossly underestimated the actual available surfacearea of particulates transported in stormwater.Specific surface area values were found to deviatefrom the monotonic pattern expected for sphericalparticles. Particles in the mid-range to coarser end(100 to 1000 µm) of the distribution were shown tocontribute a larger surface area than would normallybe expected.

The sediment binding behaviour of other toxicantssuch as polychlorinated biphenyls (PCB’s) andpolycyclic aromatic hyrdrocarbons (PAH’s) isdifferent to that of heavy metals. Schorer (1997)reported PCB’s and PAH’s to have no correlation withparticle size distribution or surface area but ratherwith the abundance of organic material. Resultsindicated that the organic material content in differentparticle size fractions was bimodally distributed withmaximum measurements recorded for fine silt (2 - 6.3µm) and fine sand fractions (63 - 200 µm).Concentrations of PAH’s would therefore be expectedto be attached to these particle size fractions.

A substantial database, identifying particle sizedistributions and other parameters that relate to

reactivity and mobility of contaminants, has resultedfrom data collected by a number of US-EPA studies.However, to date only limited information regardingthe physical and chemical characteristics of urbanstormwater runoff are available for Australianconditions. Results from an investigation by Mannand Hammerschmid (1989) on urban runoff from twocatchments in the Hawkesbury/Nepean basinindicated the existence of high correlations betweentotal suspended solids (TSS) with total phosphorus(TP), total kjeldahl nitrogen (TKN) and chemicaloxygen demand (COD). Ball et al. (1995) found thatTSS and TP show similar characteristics andcorrelations to other overseas studies.

In relation to street sweeper effectiveness, theassociation of pollutants with sediment, particularlythe finer fractions, would suggest street sweepingneeds to remove these particles in order to provideeffective stormwater pollution control. However,street sweeping has to date been found to be generallyeffective only for material larger than 300 µm (seesection 6.2).

Contaminants Associated with Sediment:

� Significant amounts of metals and nutrients are transported as sediment-bound contaminants.

� Most of the total mass of contaminants is associated with the fine particles.

� Conventional street sweeping is generally ineffective at removing particles smaller than 300 µm andtherefore will not effectively reduce the export of sediment-bound contaminants such as nutrients, metalsand PAHs.


19

6.4 Australian Conditions

Various studies undertaken by the US-EPA found themajor constituents in street dirt to be consistentlyinorganic, mineral-like matter, similar to commonsand and silt. This could be due to the fact that manyof the US-EPA studies were conducted in cities whereapplications of screened sands are made to roadsurfaces. Street surface particulate matter has beendescribed as having particle sizes ranging from about3000 to 74 µm and less (Sartor and Gaboury, 1984).

A collation of reported particle size distributioncurves for solids found on street surfaces and in streetsurface and highway runoff is shown in Figure 6.3.The collection of 20 particle size distribution curvespresented in Figure 6.3 are derived from samplingsolids from street surfaces and suspended sedimentcollected in road runoff from a number of overseasand Australian catchments.

It is evident from Figure 6.3 that despite the overseasdata being collected from a variety of sources,locations and by various methods, they show aconsistent distribution ranging from approximately 10µm to approximately 10,000 µm. The particle sizedistributions derived from sampled road runoff fromtwo Australian sites, one as part of an ongoing CRCproject and the other by Ball and Abustan (1995), arealso presented and appear to fall outside the range ofthe particle size distribution curves of the overseascatchments. The Australian data range from 2 µm toapproximately 500 µm. There may be a number ofpossible explanations for this observed finer particlesize distribution including differences in sampling

and analysis techniques. However, it should be notedthat the particle size distributions derived fromoverseas catchments were based on a variety ofsampling and analysis techniques. The upper particlesize limit can influence the position of the derivedparticle size distribution curve. Adjustments (Lloydand Wong, 1999) to the overseas data to eliminateparticles larger than 600 µm, to allow a common basisfor comparison of these curves, still showed theAustralian data sets to exhibit finer particle sizecharacteristics. The significantly different particlesize distribution of the Australian catchments mayindicate fundamental differences in catchmentcharacteristics.

The Australian sampled road runoff data displays asignificantly finer particle size distribution, with agreater percentage of particles less than 125 µm (upto 70%). Although only based on sampling at twosites, the inefficiencies of street sweeping inremoving particles less than 125 µm would result inlittle reduction of up to 70% of the particles found inrunoff in these Australian catchments. The difficultyfor Australian street sweeping is the fine nature of thesediment found on roads. Up to 70% of particlesfound on street surfaces are less than 125 µmcompared to 20% for overseas road runoff data. Theinefficiencies of street sweeping in the reduction ofsediment-bound pollutants entering the stormwatersystem is therefore expected to have more severeimplications under typical Australian conditions.

Removal of Sediment and Associated Contaminant:

� Limited sampling of sediment in street runoff in Australia indicates that 70% of particles are less than 125 µm compared to 20% for overseas data.

� The fine sediments found on Australian streets would suggest that conventional street sweeping will havea minimal effect on sediments and associated contaminants reaching stormwater systems.

CO

OPE

RA

TIV

E R

ES

EA

RC

H C

EN

TRE

FO

R CA

TCH

MEN

T H

YD

RO

LOG

Y

20

Figure 6.3 Particle Size Distribution of Suspended Solids in Road Runoff


21

7 Street Sweeping Frequency And Timing

7.1 Sweeping Frequency and Rainfall Patterns

Sartor and Gaboury (1984) concluded that thedominant influence on the effectiveness of streetsweeping appears to be time intervals, ie. therelationship between the average interval betweenstorm events (a function of local meteorologicalconditions) and the frequency at which streets areswept. Street sweeping operations are typicallyprogrammed for a fixed interval (eg. swept once perweek). If the average time between rainfall events ismuch less than the sweeping interval, then much ofthe street surface load could be washed away bystorm runoff, hence, making street sweepingrelatively ineffective. In this context, analysis ofrainfall statistics is important in the design of streetsweeping programs to ensure street sweeping iscompatible with the frequency of storm events andtherefore optimise the effectiveness of streetsweeping for removal of stormwater pollutants.

Generally street sweeping frequencies are determinedaccording to land-use. Street sweeping frequencies,practiced by Melbourne metropolitan municipalities,generally range between daily sweeping for busycommercial areas and every six weeks for residentialareas. The sweeping frequency in the CBD ofMelbourne could however involve numerous sweepsthroughout the day. Councils ordinarily stipulatesweeping specifications for the purpose of meetingcommunity demands for aesthetic quality and amenityimprovement. The inter-event dry period betweenstorms is not often a factor considered when streetsweeping programs are formulated. However, ifmunicipalities are willing to incorporate stormwatermanagement objectives into street sweepingprograms, the occurrence of rainfall events shouldbecome a significant design factor.

The minimisation of pollutant washoff, particularlyfine particulates and associated contaminants, fromstreet surfaces requires compatibility of streetsweeping frequency and timing with rainfall

characteristics and the daily activities in thecatchment. Fine particulates and associatedcontaminants are often mobilised with even thesmallest amount of runoff while gross pollutants oftenrequire a minimum runoff rate to be reached beforethey are mobilised. In areas which are not sweptdaily, the selected street sweeping frequency shouldideally reflect the relationship with the inter-event dryperiod (time between storm events) typical of thecatchment. For those catchments currently on a dailystreet sweeping regime, the time of day when streetsweeping is conducted should be selected to limit theperiod in which the pollutants deposited on streetsurfaces are exposed to the risk or likelihood of wash-off associated with a storm event.

7.2 Inter-Event Dry Period

It can be assumed that the majority of pollutantstransported into the stormwater system occur duringrainfall event periods. Therefore if the street cleaningfrequency is longer than the average inter-event dryperiod it can be expected that the accumulatedpollutants, on road surfaces, will have a higherlikelihood of being washed into the stormwatersystem before being collected by the street sweeper.

Melbourne rainfall was characterised from analysis ofrainfall over a 105 year period by Wong (1996). Theanalysis identified storm events as having a thirtyminute minimum storm duration. A six hourminimum period of no rainfall to define theconclusion of a rainfall event. Using this definitionfor a storm event, the analysis found the mean periodbetween storms in Melbourne to be 62.4 hours (2.6days) with a standard deviation of 76.8 hours (3.2days). There is an apparent trend in Melbourne oflonger periods between storms in summer months,with a maximum mean period of 108 hours (4.5 days)in February and a minimum mean period of 45 hours(1.9 days) in August as shown in Figure 7.1. Wong(1996) also carried out an analysis of the rainfall datafor a number of major cities in Australia, and thestatistics according to their respective months arepresented in Table 7.1. The influence of seasonalityon the period between storms for the cities is shownin Figure 7.2.


22

Figure 7.1 Melbourne Mean Monthly Inter-Event Dry Period

Table 7.1 Mean Inter-Event dry Periods (Hours).

Mean

CITIES JANUARY FEBRUARY MARCH APRIL MAY JUNE

Adelaide 165.93 189.42 156.52 94.81 61.07 51.16

Brisbane 65.39 57.28 58.08 74.48 93.68 111.03

Darwin 33.02 32.10 41.40 116.14 130.32 561.14

Hobart 72.33 83.26 74.79 60.86 56.24 50.69

Melbourne 97.38 107.55 89.56 66.68 55.21 49.46

Perth 250.70 238.29 200.54 89.21 58.02 39.91

Sydney 70.30 64.68 66.58 69.27 70.19 73.36

CITIES JULY AUGUST SEPTEMBER OCTOBER NOVEMBER DECEMBER

Adelaide 44.02 44.45 54.94 69.63 93.96 128.95

Brisbane 133.87 141.20 126.21 90.91 81.91 72.38

Darwin 416.95 240.36 217.41 120.79 62.21 58.72

Hobart 47.94 46.93 50.47 47.26 49.03 59.92

Melbourne 49.57 45.01 50.63 53.39 65.32 75.32

Perth 39.96 53.79 62.20 88.16 141.96 193.17

Sydney 91.48 98.50 97.78 77.87 68.92 76.31

Of the cities analysed, Darwin shows the most inter-event dry period variability between seasons, rangingbetween 32 hours (1.3 days) and 561 hours (23.4days), with the longer periods, unlike Melbourne,occurring during the winter months. The variablenature of inter-event dry periods, both betweenseasons and capital cities highlights the importance ofstreet sweeping program design being specific tolocation and flexible to accommodate for seasonvariability.

Based on consideration of typical inter-event dryperiods, one would question the effectiveness ofcurrent Australian street sweeping practices ineffectively preventing pollutants entering thestormwater system if the street sweeping frequency,

designed for aesthetic objectives, is significantly lower than the frequency of storm events. If streetsare only swept every six weeks then it is likely thatstorm events occurring within this period will flush alarge proportion of the accumulated pollutants intostormwater drains before sweeping has theopportunity to collect it. In the case of grosspollutants, Allison et al. (1998) suggested a minimumrainfall amount before there is sufficient runoff to re-mobilise these larger size pollutants. As a grosspollutant export control, sweeping frequencyequivalent to approximately three times the meaninter-event period appears to be appropriate (seeSection 8.1).


23

Sweeping Frequency and Rainfall Patterns:

� The variable nature of inter-event dry periods, both in terms of seasonal variation and dependence onclimatic locations, highlights the importance of street sweeping program designs which are specific tolocation and flexible to accommodate the local meteorological conditions and seasonal variability.

� It is anticipated that if street sweeping occurs at a longer interval than the inter-event dry period of thecatchment, street surface pollutants will have a much higher likelihood of being flushed into thestormwater system before being collected by the street sweeper.

Figure 7.2 Mean Inter-Event Periods for Australian Cities


24

Figure 7.3 Daily Litter Generation (Hall and Phillips., 1997)

Street Sweeping Timing:

� Recorded gross pollutant load generation over a typical day indicates that the accumulation of litter in ashopping strip begins at 8:00 and effectively ends around 17:00 hours.

� Early morning street sweeping allows the exposure of deposited street surface litter items to a higherlikelihood of being transported into the stormwater drainage system.

processes. Street sweeping is most commonlyconducted in the early morning leaving theaccumulated pollutants, especially litter from the daybefore, to a longer exposure period and the likelihoodof over night rainfall events capable of flushing theminto the stormwater system.

The study by Hall and Phillips (1997) also involvedcomparing accumulated litter items from streetsurfaces and side entry pit traps (SEPTs) in drainsfollowing rainfall events. The Carnegie urbancatchment was monitored over a seven day period,and litter material was measured from bins, footpaths,street surfaces and SEPTs located in stormwater draininlets. Footpath litter items were not consideredwhen determining the effect of rainfall due to theirsurfaces being sheltered from rainfall and associatedwashoff mechanisms. When only street material isconsidered, up to 77% of the calculated street itemsentered the stormwater system during rainfall events.These data suggest that street washoff is the principalmechanism for transport of gross pollutants into thestormwater system.

7.3 Street Sweeping Timing

Analysis of street and footpath litter accumulationalong a 280 m section of strip shopping centre in theMelbourne suburb of Carnegie during a typicalbusiness day was conducted by Hall and Phillips(1997). This commercial land-use area is subject totypical street sweeping operations carried out daily by the Glen Eira municipality. Detailed recording of thegross pollutant load generated over a day from 5:15 to18:30 commenced immediately after street sweepingand footpath cleaning and concluded when trade hadeffectively ended. The data indicates that the rate ofaccumulation of litter is highest between the times of8:00 and 17:00 with litter accumulation effectivelyending around 17:00 hours in the evening (see Figure7.3).

The data plotted in Figure 7.3 suggest that the time ofday a rainfall event occurs can alter the amount oflitter available for re-mobilisation to the stormwatersystem. The time of day at which street sweeping ispracticed is expected to have an effect on the amountof litter entering the stormwater system due to theexposure time of deposited pollutants to wash-off

Footway

gutter

Total


25

8 Gross Pollutant Wash-Off Characteristics

8.1 Gross Pollutant Load Generation

The study by Allison et al. (1998) showed thatstormwater runoff is the principal means by whichgross pollutants are transported to the stormwatersystem. Ten storm events (larger than 3 mm ofrainfall) and their transported gross pollutant loads inthe Melbourne suburb of Coburg were monitoredusing the CDS unit from May to August 1996(Allison et al., 1998). Monitoring was carried out in a50 hectare catchment and the amount of grosspollutants transported during each of the 10 eventswas found to be correlated with the event rainfalldepth as shown in Figure 8.1. A similarly highcorrelation between the gross pollutant load retainedin the CDS unit and event runoff was also obtained asshown in Figure 8.2.

According to the fitted relationship between the wetgross pollutant load generated and the depth ofrainfall (see Figure 8.1), events of less than 3.7 mmmay be considered to be insufficient for re-mobilisation and transport of deposited street surfaceloads. The corresponding threshold for runoff (seeFigure 8.2) is 0.70 mm. The fitted relationships

between gross pollutant wet load and event rainfalldepth or runoff show a trend of increasing grosspollutant load with increasing rainfall or runoff.Although the curves are monotonically increasing, therate of increase in gross pollutant loads decreaseswith rainfall and runoff indicating a possible upperlimit of gross pollutant load transported into thestormwater system during large rainfall or runoffevents. The fitted curves in Figure 8.1 and 8.2 maybe interpreted as indicating that the limitingmechanism for stormwater gross pollutant transport,in the majority of cases, is not the supply of grosspollutants but rather the processes (ie. the stormwaterrunoff rates and velocities) influencing themobilisation and transport of these pollutants.

If the mobilisation and transportation of grosspollutants from the street surface depends on arainfall depth greater than 3.7 mm, it is likely that theinter-event dry period for gross pollutant transportingstorm events, in Melbourne will be longer than thecalculated 2.6 days for all recorded storm events.Analysis of the cumulative frequency distribution ofevent rainfall depth for Melbourne over a 105 yearrecord is presented in Figure 8.3. The analysis showsthat approximately 35% of all recorded rainfall eventsare greater than 3.7 mm giving an average inter-eventdry period of 178 hours (7.4 days) for gross pollutanttransporting storm events.

Figure 8.1 Gross Pollutant Wet Loads v’s Rainfall (after Allison et al., 1998)


26

Figure 8.2 Gross Pollutant Wet Loads v’s Runoff (after Allison et al., 1998)

Figure 8.3 Cumulative Probability Distribution of Event Rainfall Depth for Melbourne


27

dry days numerous gross pollutant items aretransported into the stormwater system by factorsother than stormwater runoff (eg. wind or directdumping). That study focused on measuring thenumber of litter items as well as material compositioncollected daily over seven days, from identifiedcatchment pollutant sources. SEPTs were placed indrain entry pits located in the study area to determinethe number of litter items reaching the stormwatersystem from the identified catchment pollutantsources (including bins, footpaths and street surfaces).The results showed that up to 78 items of litter in total(per day) were collected in SEPTs during periodswithout rainfall. A substantial amount of the materialtrapped during recorded dry days were lighter items(polystyrene) although numerous heavier items werealso found, indicating possible direct littering ratherthan wind blown transportation of street surfacepollutants.

The Coburg gross pollutant wet load data haveincorporated the effect of Moreland City Council’sstreet sweeping practices which range from daily tofortnightly, depending on land-use. How exactly anyalterations made to the street sweeping frequencywould affect the gross pollutant load in stormwater(see Figure 8.2) is not known and cannot beascertained from the data collected. However, it ispossible for some inference of the effectiveness ofstreet sweeping in limiting the export of grosspollutants from street surfaces to the stormwatersystem to be made, and this will be discussed inSection 9.2.

Despite rainfall wash-off being the dominant factortransporting gross pollutants from street surfaces,litter can also reach the stormwater system during dryweather periods. The litter monitoring study,conducted by Hall and Phill ips (1997), in theCarnegie commercial catchment indicated that during

Gross Pollutant Load Generation:

� Data collected in the Coburg catchment indicated washoff of gross pollutants becomes significant forstorm events greater than 3.7 mm of rainfall depth or 0.70 mm of runoff.

� The limiting mechanism affecting the transport of gross pollutants in the majority of cases appears to bere-mobilisation and transport processes (ie. stormwater runoff rates and velocities) and not the supply ofgross pollutants.

� Approximately 35% of all recorded rainfall events in Melbourne are greater than 3.7 mm giving anaverage inter-event dry period of 178 hours (7.4 days) for gross pollutant transporting storm events.


28

8.2 Influence of Catchment Land-use

As part of the same project, Allison et al. (1997b)investigated the effectiveness of side entry pit traps(SEPT’s) by monitoring 192 SEPTs installed in allpublicly owned side entry pits of the 50 hectareCoburg catchment as shown in Figure 8.4. The studyaimed to assess the effectiveness of SEPTs by using aCDS unit located at the outlet of the catchment tocollect any gross pollutants which may pass theSEPTs. The SEPTs were monitored from 2 August 96to 15 November 96. During these four months, thetraps were cleaned out on four separate occasions.For each of these clean-outs the total SEPT load (wet& dry) for each trap was calculated. Gross pollutantload data from that study are used for further analysisin this study.

SEPT gross pollutant wet load data were groupedaccording to the practiced street sweeping regimedefined by catchment land-use. Figure 8.5 displaysthe three identified land-use sub-catchments in theCoburg catchment (50ha) as the daily swept South

East (SE) commercial sub-catchment (13ha), thefortnightly swept North West (NW) & South West(SW) residential sub-catchments (24.5ha) and thedaily / fortnightly swept North East (NE) mixed land-use sub-catchment (12.5ha).

The total SEPT gross pollutant wet loads werecalculated and categorised according to streetsweeping regime, defined by the three sub-catchmentland-use types and are presented in Table 8.1. Thedays between clean outs, total rainfall between cleanouts and the number of storm events, are alsopresented in Table 8.1. For the purpose of thisanalysis a storm event was identified as a storm thathad the potential to re-mobilise deposited solids fromthe road surfaces and is described as a gross pollutanttransporting event (ie. greater than 3.7 mm afterAllison et al., 1997b). The SEPT wet loads have beennormalised into a load (g) per unit catchment area(ha) to enable gross pollutant loads from the sub-catchments to be compared.

Figure 8.4 SEPT installations in the experimental 50ha Coburg Catchment (source Allison, 1998)


29

Figure 8.5 Land-use Sub-catchments in the 50ha Coburg Catchment (source Moreland City Council and Merri Creek Management Committee, 1997)


30

As indicated in Table 8.1, calculated total SEPT wetloads ranged from 1.8 kg/ha for the mixed land-usesub-catchment to as much as 20.2 kg/ha for thecommercial sub-catchment. Figure 8.6 displays thecomparison between land-use and total SEPT wetload, indicating commercial land-use contributeslarger loads of gross pollutants per hectare comparedto residential and mixed land-use catchments. This isin spite of daily street sweeping in the commercialsub-catchment compared to once every two weeks inresidential and mixed land-use areas. Three of thefour clean outs showed the ratio of gross pollutantload generation between the commercial andresidential areas to be approximately 2.0. There washowever, one clean out, that of the 15 November 96which gave a significantly lower ratio of 1.1. It isinteresting to note that the gross pollutant loadgenerated from the mixed land-use was the lowest inall the four clean outs.

Many factors other than land-use contribute to thedifferences observed in the amount of gross pollutantsexported from the different areas, including wind,traffic volume, topography, population density,community awareness and importantly the hydrologicconveyance system. Hydrologic conveyance factorswhich can influence gross pollutant export include thenumber of side entry pits in the stormwater system(ie. the average distance to entry pits from within thecatchment), the degree of catchment areaimperviousness and the extent of “supplementaryareas” (defined as pervious areas over which runofffrom impervious areas needs to traverse whendischarging towards the stormwater drainage system)in these sub-catchments. Catchment topography,

average distance along roadside kerbs and the extentof supplementary areas influence the required energyto re-mobilise and convey deposited gross pollutantsto the stormwater system. The fractionimperviousness of the catchment influences themagnitude of the runoff from the catchment which inturn determines the energy available for re-mobilisation and transport of deposited grosspollutants in the catchment.

The results presented in Figure 8.6 are consistent withresults from a separate study by Allison undertakenduring 1995 to investigate the transport of grosspollutants from different land-uses within a 150hectare catchment in Coburg. Gross pollutant loadsfrom two storm events (27 January 95 and 31 May95) were monitored at three locations representingmixed commercial/residential, residential and lightindustrial land-uses as shown in Figure 8.7 (Allison etal., 1998). On commencement of storm runoff,specifically designed gross pollutant samplers(Essery, 1994) were lowered, at varying timeintervals, into the flow and used for gross pollutantsampling as illustrated in Figure 8.8.

Gross pollutant loads from the two storm eventsmonitored for each land-use area are presented as drymass per hectare of catchment area in Table 8.2. Thecomputed unit area dry loads for the different land-uses were compared against the weighted average dryload for the three combined sub-catchments. Thesedata indicate that commercial land-use catchmentsgenerate approximately twice the amount of grosspollutants compared to residential land-use and asmuch as three times the amount generated from lightindustrial land-use catchments.

Table 8.1 Event Rainfall and Related SEPT Total Gross Pollutant Wet Loads

Daysbetween

clean-outs

TotalRainfall

(mm)

StormEvents

(>3.7mm)

Single EventRainfall

(mm)

EventRainfall

(mm)

CommercialWet Load

(g/ha)

ResidentialWet Load

(g/ha)

MixedWet Load

(g/ha)

Clean-outDate

29-Aug-96 27 55 5 (6.4) (10) (5) (15) (12) 48 5000 2408 1760

30-Sep-96 32 74 6 (11) (12.3) (16.4) (4.4) (9.4) (11.5) 65 20154 10041 6880

15-Oct-96 15 25 2 (8.2) (14) 22 6462 3143 1840

15-Nov-96 31 47 2 (7) (35.4) 42 6538 5878 1920


31

Figure 8.6 SEPT Wet Loads for Different Land use Catchment in Coburg.

Figure 8.7 Coburg Land-use Monitoring Areas in the 150 ha Coburg Catchment (source Allison et al. 1998)


32

Allison (1997b) noted that material often blinded theSEPT basket pores, leading to overflows from thebaskets and thus a reduction in trapping efficiencies.The field study into the efficiency of SEPTs, foundthe trapping efficiency of SEPTs to be between 60%and 70% (Allison et al. 1998). The SEPT total wetloads given in Table 8.1 can thus be assumed to be anunder estimation of gross pollutant loads generatedfrom the respective sub-catchments.

The gross pollutant loads for three of the four SEPTclean-outs (see Figure 8.6) show similar relativecontributions from the different land-use catchmentsas that derived from the study by Allison (1998) andsummarised in Table 8.2. The commercial catchmentwas found to have generated the most load of grosspollutants on each of the clean out dates in spite ofdaily street sweeping. As noted earlier, the ratio ofcommercial to residential land-use gross pollutantload from three of the four clean out dates isapproximately 2.0 except for the data from the cleanout of 15 November 96. The gross pollutant loadtransported from the commercial area preceding theclean out of the 15 November 96 was found to besignificantly lower than expected when compared tocorresponding data from the residential area.

The gross pollutant load from the clean-out of 15November 1996 was transported by two grosspollutant transporting storm events (ie.<3.7 mm), onewith an event rainfall of 6.8 mm and the other 35.4 mm.

The lower than expected gross pollutant load from thecommercial area in this clean-out may be related to apossible “supply limiting condition” during the large35.4 mm storm event (a trend not apparent in thefortnightly swept, residential catchment). This notionis explored in Section 9.3.

Total dry load per unit areaLand Use Area

(ha) 27-Jan-95 31-May-95 Value / Weighted(g/ha) (g/ha) Average

Commercial 9.5 423 747 1.6Residential 26.5 292 308 0.8

Light Industrial 2.5 242 63 0.5Total 38.5

Weighted Average 321 400

Table 8.2 Gross Pollutant Dry Mass Loads and Weighted Averages (after Allison et al., 1998)

Figure 8.8 Sampling Gross Pollutants from DifferentLand-use Sub-catchments in Coburg

(source Allison et al, 1998)


33

Influence of Catchment Land-use:

� The fraction imperviousness of a catchment influences the runoff during storm events which influence theavailable energy for mobilisation of deposited gross pollutants.

� Commercial land-uses contribute larger loads of gross pollutants despite more intensive street sweepingfrequencies.

� Relative gross pollutant loads generated from different land-uses show that commercial areas produceapproximately twice the amount of gross pollutants than residential and three times as much as lightindustrial, despite a daily street sweeping regime in the commercial area compared to fortnightly in theresidential and industrial areas.

� A number of transport factors are thought to also influence gross pollutant loads from different land-uses.Some of these factors include:-

� Number of entrances to the stormwater system,

� Fraction of catchment imperviousness,

� Extent of pervious area over which runoff needs to traverse towards the stormwater drainage system.


34


35

9 Discussion

9.1 Gross Pollutant Load and Rainfall DepthRelationship

The relationships between the gross pollutant loadand rainfall depth (Figure 8.1) and runoff (Figure 8.2)derived from the Coburg data incorporate the effect ofa typical Melbourne municipal street sweepingprogram, ranging in frequency from daily tofortnightly sweeping depending on catchment land-use. The relationships clearly show a trend ofincreasing gross pollutant load to the stormwatersystem with increasing rainfall or runoff, indicatingthat the limiting mechanism for stormwater grosspollutant transport in the majority of cases isstormwater runoff rates and velocities. While thecurves are monotonically increasing, the rate ofincrease in gross pollutant loads entering thestormwater system decreases with rainfall and runoffindicating a possible upper limit of gross pollutantload transported into the stormwater system atrelatively high rainfall depths or runoff. This possibleupper limit of gross pollutant load may reflect thegross pollutant load deposited on street surfaceswhich is available for re-mobilisation into the

stormwater system. A modification of the streetsweeping frequency could potentially adjust thisupper limit value, thereby altering the shape of thegross pollutant export curve as conceptualised inFigure 9.1.

9.2 Impact of Street Sweeping on GrossPollutant Loads

It is not known how exactly any further alterationsmade to the street sweeping frequency will affect thegross pollutant export curve. Nevertheless theillustration in Figure 9.1 postulates that if streetsweeping effort were reduced it can be expected thatthe gross pollutant load will increase, initially forthose events with large rainfall depths. Furtherreduction in street sweeping frequency will ultimatelylead to the increase of gross pollutants in stormwatersystems becoming evident for even smaller stormevents. Similarly, by increasing street sweepingeffort, the reduction in gross pollutant load wouldessentially be confined to events of large rainfalldepths. Figure 9.1 postulates that in most grosspollutant export events, the export load is defined bythe size of the storm event rather than the availablepollutant surface load.

Figure 9.1 Hypothetical Gross Pollutant Load and Street Sweeping Effort


36

9.3 Supply Limiting Condition

The lower than expected gross pollutant load from thecommercial area for the clean-out of the 15November 96 noted in Table 8.1 of the previoussection of this report may be explained by a possible“supply limiting condition” occurring during the large35.4 mm storm event. It is possible that during thislarge event the available gross pollutants in thecatchment have been substantially removed from thestreet surface and mobilised into the stormwatersystem, a trend not apparent in the fortnightly swept,residential catchment.

Based on the results of this investigation, it ispostulated that a source limiting storm condition mayhave occurred during the 35 mm storm event. Stormevents greater than 35 mm occur less than 3% of thetime in Melbourne (see Figure 8.3) indicating that theoccurrence of such a gross pollutant supply limitingcondition would be very rare. This may haveimportant implications for assessing the effectivenessof street sweeping. The incremental benefits ofincreasing the present street sweeping effort in theCoburg catchment (from the daily frequency of thecommercial areas and fortnightly frequency in theresidential areas) are expected to be low. The limitingfactor affecting the transport of gross pollutants in themajority of cases appears not to be the supply of grosspollutants but instead the pollutant mobilisation andtransport processes (ie. rainfall patterns and depths,runoff rates and velocities).

9.4 Street Sweeping Efficiency Issues

The use of new street sweeping technologies maycontribute to reducing pollutant loads in thestormwater system as advocated by Sutherland andJelen (1997). Taking into account influencing factorssuch as the inter-event dry period and catchmentcharacteristics may enable the frequency and timingof street sweeping operations to be redesigned to meetspecified stormwater improvement objectives forspecific conditions. Street sweeping frequencies thatare equivalent to three times the mean inter-eventperiod (approximately 8 days for Melbourne) isconsidered to be appropriate as approximately 35% ofstorm events are considered to be gross pollutant

transporting events. Also, conducting street sweepingat a time of day which enables the collection ofpollutants when the rate of load acummulation ofstreet surface has reached its highest would reduce thetime pollutants are potentially exposed to thelikelihood of rainfall events.

Factors contributing to inefficiencies in streetsweeping are not confined to rainfall patterns(affecting the build-up and wash-off processes),frequency and timing of sweeping, size of pollutantsand the sweeper mechanism. Street sweepinginefficiencies are further exacerbated by everydaypractice limitations. Significant practice limitationsassociated with street sweeping include the inabilityof sweepers to access the street surface load due toparked vehicles (see Figure 9.2), inappropriate streetdesign, poor road surface conditions and operatorspeed. Street sweeping program specifications mustaddress these influencing factors as well as improvingsweeper mechanisms before stormwater qualityimprovements may be realised from street sweepingpractices.

The principle objective of street sweeping in meetingcommunity demand for a standard of streetcleanliness, and the perceived success of sweeping tofulfi l this objective makes street sweeping animportant municipal operation. However, there islittle evidence to suggest significant incrementalbenefits in stormwater quality, particularly theremoval of contaminants associated with the fineparticulates, can be gained with increased streetsweeping frequency.

The use of new street sweeping equipment may leadto increased effectiveness particularly for grosspollutants and coarse to medium sized sediment.There are however other operational limitationswhich will reduce the actual effectiveness of streetsweeping from that determined under controlled testconditions. Furthermore, the use of new equipmentwill need to be associated with a street sweepingfrequency that matches the catchment meteorologicalcharacteristics. Their cost effectiveness will need tobe evaluated against the cost of installing andmaintaining end-of-pipe or in-transit gross pollutanttraps.


37

Figure 9.2 Street sweeping pollutant removal effectiveness is limited by parked cars


38


39

10 Conclusions

This study has investigated the effectiveness of streetsweeping for stormwater quality improvement. Anumber of factors are identified as influencing theeffectiveness of street sweeping for the collection ofstreet surface pollutants for stormwater pollutioncontrol rather than just aesthetic requirements. Thesefactors include street sweeping mechanism, pollutanttype, sweeping frequency and timing and alsopollutant wash-off characteristics.

The most important conclusion from this study is thatcurrent Australian street sweeping practices aregenerally ineffective as an at source stormwaterpollution control measure. Current street sweepingpractices are found to be not only ineffective for thereduction of fine sediment and sediment-boundcontaminants but also for larger gross pollutantscapable of entering the stormwater system. CurrentAustralian street sweeping mechanisms and practicesare therefore regarded as providing very little benefitfor stormwater quality improvements, due toinefficiencies at reducing a variety of pollutants fromentering the stormwater system over a range ofconditions. Street sweeping should be thereforeaccompanied by structural pollutant treatmentmeasures to effectively reduce the discharge of grossand sediment associated pollutants in stormwater.

Increasing the frequency of current street sweepingpractices beyond what is required to meet aestheticobjectives is not expected to yield substantialincremental benefits in relation to receiving waterquality improvements. There seems little benefit inconducting detailed field monitoring investigationsinto quantifying the effectiveness of street sweepingas a stormwater pollution control measure for currentAustralian street sweeping mechanisms or operations.Other specific observations from this study are listedbelow.

Sweeping Mechanisms

� Mechanical and regenerative air street sweepingequipment requires a minium threshold load ofsediment on the street surface before they becomeeffective.

� The threshold load can be three times higher forthe mechanical sweeper compared to theregenerative air system.

� Overall the regenerative air sweeper exhibits asubstantially better performance than the regularmechanical sweeper.

� Street sweeping technology is developing andimproving to remove finer street surface particlesfor a variety of street surface loads.

Gross Pollutants

� Significant amounts of gross pollutants aremobilised into the stormwater system duringbursts of rain, wind or both.

� There is little correlation between the frequencyof sweeping and the transport of gross pollutantsinto the stormwater system.

� Street sweeping efficiency increases with particlesize.

� Sweeper efficiency can be up to nearly 80% forparticles greater than 2 millimetres under ‘test’conditions (ie. sweeping more frequently than theoccurrence of rainfall events and effective use ofparking restrictions).

Sediment and Other Suspended Solids

� The removal efficiency of sediment and other fineorganic particles by conventional street sweeperswas found to be dependent upon a threshold levelof load on the surface and the particle size rangeof the surface loads.

� Material smaller than 300 µm was less affected bystreet sweeping.

� No effective removal (>50% removal efficiency)was evident for particle sizes smaller than 125 µmfor conventional street sweepers (excluding thenew small-micron surface cleaning technology).

Contaminants Associated with Sediment

� Significant amounts of metals and nutrients aretransported as sediment-bound contaminants.

� Most of the total mass of contaminants isassociated with the fine particles.

� Conventional street sweeping is generallyineffective at removing particles smaller that 300 µm and therefore will not effectively reducethe export of sediment-bound contaminants suchas nutrients, metals and PAHs.


40

Removal of Sediment and Associated Contaminant

� Limited sampling of sediment in street runoff inAustralia indicates that 70% of particles are lessthan 125µm compared to 20% for overseas data.

� The fine sediments found on Australian streetswould suggest that conventional street sweepingwill have a minimal effect on sediments andassociated contaminants reaching stormwatersystems.

Street Sweeping Frequency

� The variable nature of inter-event dry periods,both in terms of seasonal variation anddependence on climatic locations, highlights theimportance of street sweeping program designwhich are specific to location and flexible toaccommodate the local meteorological conditionsand seasonal variability.

� It is anticipated that if street sweeping occurs at alonger interval than the inter-event dry period ofthe catchment, street surface pollutants will havea much higher likelihood of being flushed into thestormwater system before being collected by thestreet sweeper.

Street Sweeping Timing

� Recorded gross pollutant load generation over atypical day indicates that the accumulation oflitter in a shopping strip begins at 8:00 am andeffectively ends around 5:00 pm.

� Early morning street sweeping allows theexposure of deposited street surface litter items toa higher likelihood of being transported into thestormwater drainage system.

Gross Pollutant Load Generation

� Data collected in the Coburg catchment indicatedwashoff of gross pollutants becomes significantfor storm events greater than 3.7 mm of rainfalldepth and 0.70 mm of runoff.

� The limiting mechanism affecting the transport ofgross pollutants in the majority of cases appearsto be re-mobilisation and transport processes (ie.stormwater runoff rates and velocities) and not thesupply of gross pollutants.

� Approximately 35% of all recorded rainfall eventsin Melbourne are greater than 3.7 mm giving anaverage inter-event dry period of 178 hours (7.4days) for gross pollutant transporting stormevents.

Influence of Catchment Land-use

� The fraction imperviousness of a catchmentinfluences the runoff during storm events whichinfluence the available energy for mobilisation ofdeposited gross pollutants.

� Commercial land-uses contribute larger loads ofgross pollutants despite more intensive streetsweeping frequencies.

� Relative gross pollutant loads generated fromdifferent land-uses show that commercial areasproduce approximately twice the amount of grosspollutants than residential and three times asmuch as light industrial, despite a daily streetsweeping regime in the commercial areacompared to fortnightly in the residential andindustrial areas.

� A number of transport factors are thought to alsoinfluence gross pollutant loads from differentland-uses. Some of these factors include:-

� number of entrances to the stormwater system,

� fraction of catchment imperviousness,

� extent of pervious area over which runoff needs to traverse towards the stormwater drainage system.


41

CH2M-HILL, Pitt, R., Cooper and Associates, Inc.,and Consulting Engineering Services, Inc, (1982).Street Particulate Data Collection and Analyses,Prepared for WASHOE Council of Governments,August, 1982, Vol. 2.

Colwill, G.M., Peters, C.J. and Perry, R. (1984).Water Quality of Motorway Runoff, TRRL Supp, Rpt823, Crowthorne, Berkshire, England.

Dempsey, B.A., Tai, Y.L., Harrison, S.G. (1993).Mobilisation and Removal of ContaminantsAssociated with Urban Dust and Dirt, Wat. Sci.Tech. Vol. 28, No.3-5, pp. 225-230.

Ellis, J.B., Revitt, D.M., Harrop, H.O. and BeckwithP.R. (1986). The Contribution of Highway Surfacesto Urban Stormwater Sediments and Metal Loadings,Sci. Total Environ, Vol. 59. Pp 339-349.

Essery, C.I. (1994). Gross Pollutant Water Quality -Its measurement and the performance of remediationtechnologies/ management practices, StormwaterIndustry Association, Best Stormwater ManagementPractice, NSW.

Fergusson J.E and Ryan D.E. (1984). The ElementalComposition of Street Dust from Large and SmallUrban Areas Related to City Type, Source andParticle Size, Sci. Total Environment, Vol.34 pp.101-116, Elsevier Science Publishers B.V., Amsterdam -Printed in the Netherlands.

Hall, M.D. and Phill ips, D.J. (1997). LitterGeneration and Distribution in Commercial StripShopping Catchments, in Future directions forAustralian soil and water management, Stormwaterand Soil Erosion conference, Brisbane, pp 501-508.

Hvitved - Jacobsen, T. and Yousef Y.A. (1991). Roadrunoff quality, Environmental impacts and control inroad pollution, eds R.S. Hamilton and R.N. Harrison,Elseview, London, pp 165-209.

Hvitved - Jacobsen, T., Johansen N.B. and YousefY.A. (1994). Treatment Systems for urban andhighway runoff in Denmark, Sci. Total Environ,Vol146/147. Pp 499-506.

References

Allison, R.A. and Chiew, F.H.S. (1995). Monitoringof Stormwater Pollution for Various Land-uses in anUrban Catchment, Proc. 2nd Int. Sym. On UrbanStormwater Management, Melbourne Australia, 11-13July, IE Aust., NCP 95/03, Vol.2, pp 511-516.

Allison, R.A., Chiew, F.H.S. and McMahon, T.A.(1997a). Stormwater Gross Pollutants, IndustryReport 97/11, Cooperative Research Centre forCatchment Hydrology, December 1997.

Allison, R.A., Rooney, G., Chiew, F.H.S. andMcMahon, T.A. (1997b). Field Trials of Side EntryPit Traps for Urban Stormwater Pollution Control,Proceedings of the 9th National Local GovernmentEngineering Conference, I.E. Aust., Melbourne,Australia.

Allison, R.A., Walker, T.A., Chiew, F.H.S., O’Neill,I.C. and McMahon, T.A. (1998). From Roads toRivers, - Gross Pollutant Removal from UrbanWaterways, Cooperative Research Centre forCatchment Hydrology.

Alter, W. (1995). The Changing Emphasis ofMunicipal Sweeping.....May be Tandem., AmericanSweeper, Volume 4, Number 1, pp.6.

Baker R.A. (1980). Contaminants and Sediment, AnnArbor Science, Publishers inc/the Butterworth Group,Michigan.

Ball, J.E. and Abustan, I. (1995). An Investigation ofParticle Size Distribution during Storm Events froman Urban Catchment, Proceedings of the SecondInternational Symposium on Urban StormwaterManagement, Vol. 2, NCP No .95/03, pp 531-535.

Bannerrnan, R., Baun, K. and Bohn, vf. (1983).Evaluation of Urban Non-Point Source PollutionManagement in Milwaukee County Wisconsin. Vol 1.Urban Stormwater Characteristics, Sources andPollutant Management by Street Sweeping. Preparedfor Environmental Protection agency, Chicago, IL.

Bender, G.M. and Terstriep, M.L. (1984).Effectiveness of Street Sweeping in Urban RunoffPollution Control, Sci. Total Environ, 33: 185-192.


42

Lloyd, S.D. and Wong, T.H.F. (1999). Particulates,Associated Pollutants and Urban StormwaterTreatment, Paper accepted for The 8th InternationalConference for Urban Storm Drainage. Sydney, 30August -3 September.

Mann, R. and Hammerschmid, K. (1989). Physicaland Chemical Characteristics of Urban Runoff fromTwo Catchments in Hawkesbury / Nepean Basin,Australia Water and Wastewater Association 13thFederal Convention, Canberra 6-10 March, 1989.

Nilson, B., Silby, N. and Argue, J.R. (1997). AnInvestigation into Source Control of Gross Pollution,Submitted at the 8th National Local GovernmentEngineering Conference.

North Caroline Department of Natural Resources andCommunity Development (NCDNRCD) (1993). AnEvaluation of Street Sweeping as a Runoff PollutionControl. EPA Publication, PB85-102507.

O’Brien, E.J. (1994). Water Quality ChangesAssociated with the Installation of the Bondi GrossPollutant Traps, Proceedings of the Australian Waterand Wastewater Associations 16th FederalConvention, I.E. Aust., ACT, Australia.

Pitt, R.E. (1979). Demonstration of NonpointPollution Abatement Through Improved streetCleaning Practices, EPA 600/2-79-161, 270pp.

Pitt R.E. and Amy G. (1973). Toxic MaterialsAnalysis of Street Contaminants, USEPA WashingtonDC Report No. EPA-R2-73-283.

Pitt, R.E. and Bissonnette, P. (1984). Bellevue UrbanRunoff Program: Summary Report, Prepared for theNationwide Urban Runoff Program (NURP), WaterPlanning Division, U.S. Environmental ProtectionAgency.

Pitt, R.E. and Shawley, G. (1982). A Demonstrationof Non-Point Source Pollution Management onCastro Valley Creek, Alameda County Flood Controland Water Conservation District, Hayward,California.

Pitt, R.E. and Sutherland, R.C. (1982) WashoeCounty Urban Stormwater Management Program, Vol2 Street Particulate Data Collection and Analysis,Prepared by CH2M Hill for Washoe Council ofGovernments, Reno, Nevada, November.

Powell, M. and Lloyd, S.D. (1999). SuspendedSolids, Metals and Phosphorous in Road Runoff froma Major Transport Route in Melbounre, ACooperative Research Centre for CatchmentHydrology Working Document.

Sansalone J.J., Koran, J.M., Smithson J.A. andBuchberger S.G. (1997), Characterisation of SolidsTransported from an Urban Roadway Surface,Proceedings of the Water Environment Federation,70th Annual Conference, Chicago, Illinois, U.S.A.October 18-22, Vol 4.

Sartor, J.D. and Boyd, G.B. (1972). Water PollutionAspects of Street Surface Contaminants, EPA-R2-72-081, U.S Environmental Protection Agency, 1972.

Sartor, J.D., Boyd, G.B. and Agardy, J.F. (1974).Water Pollution Aspects of Street SurfaceContaminants, Journal of Water Pollution ControlFederation, Vol.46. pp. 458-667, Washington D.C.

Sartor, J.D. and Gaboury, D.R. (1984). StreetSweeping as a Water Pollution Control Measure:Lessons Learnt Over the Last Ten Years, Sci. TotalEnviron., 33: 171-183.

Schorer, M. (1997). Pollutant and Organic MatterContent in Sediment Particle Size Fractions,Freshwater Contamination, Proceedings of RabatSymposium S4, April-May, IAHS Publ. No243.

Shaheen, D.G. (1975). Contribution of UrbanRoadway Usage to Water Pollution, U.S.Environmental Protection Agency.

Shaver, E. (1990). Sand Filter Design for StormwaterTreatment, Waterfall, 4. Pp. 31-36.

Sutherland, R.C & Jelen, S.L. (1993). SimplifiedParticle Transport Model-Users Manual, Version 3.1,66pp.

Sutherland, R.C. and Jelen, S.L. (1996a).Sophisticated Stormwater Quality Modeling is Worththe Effort. Published in Advances in Modeling theManagement of Stormwater Impacts, Volume 4,Edited by Dr William James, CHI Publications.

Sutherland, R.C. and Jelen, S.L. (1996b). StudiesShow Sweeping has Beneficial Impact on StormwaterQuality, Published APWA Reporter, Volume 63,No.10, pp.8.


43

Sutherland, R.C. and Jelen, S.L. (1997). Contrary toConventional Wisdom: Street sweeping can be anEffective BMP. Published in Advances in Modelingthe Management of Stormwater Impact, Volume.,Edited by William James, CHI Publications.

Sutherland, R.C., Jelen, S.L, and Minton, G. (1998).High Efficiency sweeping as an Alternative to the useof Wet Vaults for Stormwater Treatment, Published inAdvances in Modelling the Management ofStormwater Impacts, Volume 6,. Edited by WilliamJames, CH1, Publications.

Terstriep, M.L., Bender, G.M. and Noel, D.C. (1982)Evaluation of the Effectiveness of Municipal StreetSweeping in the Control of Urban Stormwater RunoffPollution, Final Report, Nationwide Urban RunoffProject, Champaign, Illinois, US EnvironmentalProtection Agency.

Wilber W.G. and Hunter J.V, (1979). Distribution ofMetals in Street Sweepings, Stormwater Solids, andUrban Aquatic Sediments, Journal of WPCF, Vol 51,No 12, December.

Wong, T.H.F. (1996). Synthetic Generation of StormSequence using a Multi-module Data GenerationTechnique, Proceedings of the Institute of Engineers,Australia, 23 Hydrology and Water ResourcesSymposium, Hobart, Australia, pp. 439-444, 21-24May 1996.

Woodward-Clyde (1992). Annual Monitoring Report.Prepared for the Santa Clara Valley Urban RunoffPollution Prevention Program, September.

Woodward-Clyde (1993). Weir Study. UnpublishedReport prepared for Alameda Countywide UrbanRunoff Clean Water Program.

Woodward-Clyde (1994). San Jose Street SweepingEquipment Evaluation. Prepared for City of San Jose,California, October.

Woodward-Clyde (1996). Catchbasin Assessment forPollution Removal. Draft Report Prepared forCaltrans. September.

Yousef, A.Y., Lin, L., Sloat, J.V. and Kaye, K.Y.(1991). Maintenance Guidelines for AccumulatedSediments in Retention/Detention Ponds ReceivingHighway Runoff. Prepared for Florida Department ofTransportation. Publication # FL-ER-47-91.

- clearwatervic.com.au · i preface this report investigates the effectiveness of street sweeping...

Documents