Chapter 6 Headwater Streets · The greatest share of total impervious cover in most communities is from the roads, sidewalks, parking lots and driveways used to get us to where we

_____________________________________________________________

Street systems are organized along a hierarchial network, with the smallest access streets feeding into collectors,which in turn connect to larger arterial streets and freeways.

Chapter 6

Headwater Streets

Introduction

The greatest share of total impervious cover inmost communities is from the roads, sidewalks,parking lots and driveways used to get us towhere we work, live or shop. This reflects thestrong influence that the car has in shaping thedesign of our communities. In this chapter, weexamine techniques to reduce the imperviouscover created by residential streets. The term“headwater streets” is usedhere to distinguish residential streets from thewider and more heavily travelled roads and

highways that are also a part of the urbanlandscape.

Some Street Geometry and Terminology

Road networks resemble stream systems inmany respects. For example, they areconnected in a hierarchial network that is quitesimilar to stream order. Small access streetsgenerate the traffic that is routed to collectorstreets that in turn connect with arterial roads,that ultimately feed freeways (Fig. 33). Likestreams, the capacity and width

FIGURE 33: THE HIERARCHY OF STREETS

Site Planning for Urban Stream Protection

_____________________________________________________________ 130

of roads tend to increase in a downstreamdirection. And just like headwater streams,local streets comprise the majority of the roadlength of the entire road network in acommunity. Recent studies indicate that theyrepresent between 50 and 65% of the length ofthe entire road network (Carrol County 1992).

The analogy with streams is not altogetherperfect, however. The most notable differenceis in the direction of flow. Runoff only travelsin a downstream direction, whereas roads aredesigned for two way traffic of vehicles. Thisof course, introduces a safety problem—howto keep vehicles travelling in oppositedirections from colliding with each other.

The traditional street classification system

Streets are classified according to the trafficvolume they are expected to carry. Trafficvolume is computed in a fairly simple manner.Each single family home generates a numberof vehicle trips each day. As it happens, atypical single family home generates about tentrips every day. Thus, the expected trafficvolume is simply the product of the averagenumber of trips per residence and the numberof residential units located along the street.This statistic, known as the average dailytraffic or ADT, can be calculated for any streetor road. For example, a residential access streetthat serves 15 homes would have an ADT ofabout 150.

Road designers use ADT to classify streets,and set road design standards. The hierarchialclassification system assigns a street to one offour general categories, based on its ADT.

Thus, in ascending order of traffic volume, wehave access streets, collector streets, arterialstreets and freeways.

Access streets occupy the lowest rung in thestreet hierarchy. They conduct traffic betweenindividual dwelling units and higher orderstreets (such as collectors, arterials andfreeways) Also known as local roads, theygenerally handle no more than 500 to 1,000ADT.

Collector streets are used to funnel trafficbetween smaller access streets and largerarterial roads. They act as the primary trafficroute within a residential or commercial area,and can handle from 1,000 to 3,000 ADT.

Arterial streets provide a direct route for longdistance travel to different parts of acommunity, and are fed by collectors streets atcontrolled intersections. Arterial streets aredesigned for greater speed and volume andusually handle from 3,000 to 10,000 ADTs.Arterial streets may eventually feed into evenlarger freeways that allow for high speed travelfrom one region to the other.

Freeways can handle 30,000 or more tripseach day, and are designed for limited andcontrolled access.

This chapter focuses exclusively on the designof smaller “headwater” streets (i.e., access andcollectors) for two reasons. To begin with,most local communities only have authority todevelop or modify design standards forheadwater streets of a subdivision. Second,even if they had authority to modify the design

Chapter 6: Headwater Streets

_____________________________________________________________ 131

of larger arterials and freeways, there would befew opportunities to make these roadsnarrower (indeed, many larger road systemsare continually widened to keep up with everexpanding traffic volumes).

Why are Residential Street So Wide?

The design standards that govern the geometryof roads are derived from two basic sources:the American Association. of State Highwayand Transportation Officials (AASHTO 1990)and the Institute of Transportation Engineers(ITE 1987, 1991). One of these sets ofstandards must be followed for any street orroad project built with state or federal funds,

them without change for local roads. Asummary of existing design standards ispresented in a very condensed form in Table33.

When it comes to residential access roads, thecurrent AASHTO and ITE recommendationsadopt a one–size–fits–all approach (Stabenfeldt1995). For example, AASHTO onlyrecognizes one basic design for residentialaccess streets, that has a minimum pavementwidth of 26 feet, and 24 extra feet for the rightof way. The ITE standards, which are lessfrequently used as the basis for residentialstreet design, do allow for a greater range ofwidths, depending on terrain, housing density,and whether on–street parking will be

TABLE 33: CONDENSED SUMMARY OF NATIONAL DESIGN STANDARDS FOR RESIDENTIAL STREETS

DESIGN CRITERIA AASHTO ITE HEADWATER STREETS

Residential StreetCategories

1 3, depending onland use density

4, depending on ADT

Minimum StreetWidth

26 ft min. 22–27 ft >2 du28–34 ft 2–6 du36 ft < 6 du

16 ft (>100 ADT)20 ft (100–500 ADT)26 ft (500–3,000 ADT)32 ft ( >6 du/ac)

Additional Right of Way 24 ft 24 ft 8 to 16 ft

Design Speed, LevelTerrain

30 mph 30 mph 15 to 25 mph

Curb and Gutter generallyrequired

generally required not required oncollectors

Cul–de–sac Radii 30 ft 40 ft 30 ft

Turning Radii in Cul-de-sac

20 ft 25 ft 17 ft


_____________________________________________________________ 132

The width of residential streets is determined by the number and width of moving and parking lanes provided. Inheadwater streets, a parking and moving lane are shared, thereby reducing street width.

provided on one or both sides of the street.Thus, under the ITE standards, pavementwidth can range from 22 to 36 feet, with moststreets between 28 and 34 feet. In practicalterms, street widths are determined by thenumber of parking and moving lanes provided(Figure 34):

9 three eight foot lanes, devoted to movingor parking = 24 feet

9 one ten foot moving lane, plus two eightfoot parking lanes = 26 feet

9 two ten foot moving lanes, plus one eightfoot parking lane = 28 feet

9 two ten feet moving lanes, plus two eightfoot parking lanes = 36 feet

mind. First, wide residential streets aredesigned to promote relatively rapid trafficflow, at an average speed limit of 30 mph onlevel or rolling terrain. The wider streets andbetter sight distances, however, encouragemany drivers to exceed even this relativelyhigh speed. High speeds, however, are notdesirable in residential neighborhoods. Oneway to force drivers to slow down is to reducestreet width.

Second, street width is generally determined bythe planned function of the street, rather thanthe actual traffic capacity it experiences. As aconsequence, street width is fixed, regardlessof whether it serves 10 homes or

FIGURE 34: PARKING AND MOVING LANES IN RESIDENTIAL STREETS


_____________________________________________________________ 133

100 homes. To put this into perspective,consider the relationship between the numberof dwelling units and the traffic they generate(Table 34). A residential street serving tenhomes can be expected to handle about 100cars each day, which equates to an average ofroughly 15 minutes between each car trip (6minutes during the peak hour). A second streetserving 100 homes will typically handle athousand car trips each day, with one car tripgenerated every 90 seconds (about 30 secondsduring the peak hour). In the first case, vehiclescan share a common moving lane since onevehicle can pull into a parking lane to allowanother vehicle to pass. This rather minorinconvenience does not occur very often whenthe number of homes served by the street issmall (see Table 34). Shared moving lanesbecome a major inconvenience,

and a possible safety hazard, however, once astreet serves more than 50 dwelling units.

Third, streets are often utilized as a spilloverparking area in residential neighborhoods, withone or more on–street parking lanes beingprovided. Residential parking demand hasgrown sharply over the last decade in responseto ever increasing trends in car ownership. Forexample, over a third of all families in the USnow own two or more cars (ULI 1990).Consequently, two to three parking spacesmust usually be supplied per dwelling unit toaccommodate the future parking needs ofresidents and their visitors. In large lotdevelopments, however, on–street parkinglanes creates a surplus of unused parking areas.

TABLE 34: RELATIONSHIP BETWEEN NUMBER OF DWELLING UNITS, TRAFFIC GENERATION, AND

RESIDENTIAL CONGESTION

No. of SF Homes

AverageDaily Trips

Peak TripsPer Hour

Minutes between cars (average)

Minutes betweencars (peak)

5 50 5 30 12

10 100 10 15 6

25 250 25 6 4

50 500 50 3 1.5

75 750 75 2 45 secs

100 1,000 100 1.5 35 secs

150 1,500 150 1 20 secs

300 3,000 300 30 secs 10 secs

Many residential streets carry relatively few vehicles each day. For example, streets serving less than 25 homes areso lightly travelled each day (and during peak hours) that shared parking and moving lanes make sense.


_____________________________________________________________ 134

The Headwater Street Alternative

The design of headwater streets is directlylinked to the traffic and parking demandgenerated by the homes that are served. Ingeneral, streets are designed to the narrowestwidth capable of fully meeting the traffic andparking demand. A revised classificationsystem for headwater streets is presented inTable 35, where street width declines withdecreasing ADT. The classification systemrepresents a composite of innovative residentialstreet standards drawn from severalcommunities around the country. Fiveresidential street categories are defined basedon traffic and parking demand. They are:

Lane: serves less than 15 homes with a densityof 2 dwelling units per acre or more. This lowspeed street is only 16 feet wide, with anadditional 8 to 16 ft. right–of–way. Parkingdemand is met by driveways or grassshoulders, and drainage is typically providedby grassed channels.

Access: these streets serve 15 to 50 homes, andare only 20 to 22 feet in width (either twomoving lanes, or a parking lane and a sharedmoving lane). The right–of way may extend 8to 24 feet, depending on whether a grasschannel or curb/gutters are used for drainage.A sidewalk is located on one side of the street.

Standard street: this street category handlestraffic from 50 to 100 homes, and is 26 feetwide, thereby allowing for one moving laneand two parking lanes (one of which doublesas a moving lane most of the time). Drainage isusually provided by curb and gutters.

Dense street: a wider street (32 to 34 feet) isoften needed when housing densities exceedfour dwelling units/acre in order to meetresidential parking demand. Parking lanes areprovided on both sides of the street, althoughone or both lanes can be eliminated if a streetsection is more than 200 feet away from thenearest residence (i.e., hourglass streets). Thedense street often will have curb and gutter,and can have sidewalks on one or both sides ofthe street.

Collector: the primary function of this streetcategory is to funnel traffic fromneighborhoods to arterial streets, and becauseof the high traffic level (1,000 to 3,000 ADT),no frontage lots are allowed. The typical widthis 22 feet which consists of two moving lanes(with grass shoulders for emergency parking)or 28 feet if a spillover parking lane is needed.

The headwater street classification systemrepresents the minimum width that canprobably be achieved without compromisingsafety or traffic flow (Fig. 35). Communitieswill need to modify it to reflect their uniquestreet terminology, trip generation rates, snowstorage, utility and pedestrian accessrequirements.

Residential Street Features

Residential streets do more than just carry cars.They also serve as a major corridor for utilitiesand pedestrian movement in a community. Asa result, most communities require a minimum12 to 15 foot wide right of way on each side ofthe street. Underground lines provide water,sewer, telephone, gas or


_____________________________________________________________ 135

TABLE 35: CONDENSED SUMMARY OF DESIGN STANDARDS FOR FIVE CATEGORIES OF HEADWATER

STREETS

Design Factor

Lane Access StandardStreet

DenseStreet

Collector

ADT less than 100

100 to 500

500-1,000 100-1,000at 4 du/ac

1,000 to3,000

Width 16 20 26 32 22 to 28

Extra ROW 8 to 16 ft.* 8 to 24 ft.* 20 ft. 20 ft. 22 to 28 ft.

Off–StreetParking

privatedriveways

privatedriveways

privatedriveways

multi-familyparking lots

none

On–StreetParking

None one-lane one-lane two-lane emergencyshoulders

Drainage Swale Swale orcurb/gutter

curb/gutter curb/gutter Swale orShoulder

Design Speed 15 mph (max) 20 mph 25 mph 25 mph 25 mph

Sidewalks none one-side one or twosides

two sides one side

Frontage Lots

yes yes yes yes no

The terms used to classify the five kinds of headwater streets are illustrative only, and many communities maywant to use other terms. The dense street refers to a street section that serves 4/du/ac or more that may requireon–street parking or multi–family parking lots to meet parking demand.

*higher right of way length is needed for grass channels and grass roadway shoulder.

electricity service to each home within thisprotected easement. In addition, sidewalks andstreet trees are located inside the right of way,usually on both sides of the street. In northernclimates, the right of way is also used to storeexcess snow during the winter months.

The right–of–way is also the primary corridorfor moving stormwater runoff away from adevelopment. Runoff is conveyed along thestreet network in one of two ways, either (a) in

an open grass channel located in the right ofway or (b) in an enclosed storm drain locatedunder the street or right of way. Storm drainsare fed by a system of curb and gutters thatchannels street runoff into a pipe inlet. Openchannels and storm drains are both sized tohandle large runoff volumes, and typicallyhave a capacity to carry the runoff from a tenyear design rainfall event.


_____________________________________________________________ 136

FIGURE 35: COMPARISON OF HEADWATER STREET WIDTHS

Several options for sharing parking and moving lanes are at the heart of the headwater street concept.

The use of an open channel or storm drain in aparticular street is determined by a number offactors, such as drainage area, slope, length,housing density, and street type. Openchannels can be used on smaller streets, but atsome point, runoff velocities become tooerosive to be safely handled in an earthenchannel, and they must be enclosed in a stormdrain. This limit, known as the critical erosivevelocity, is typically around 4 to 5 feet persecond. A channels' maximum velocity isgenerally defined and computed using the peakdischarge rate under the two year design stormevent. Open channels can have many streamprotection benefits. For example, stormwaterpollutants are filtered through grass or soil asthey pass through the channel. The lack of acurb eliminates a major trap of airbornepollutants. In addition, runoff can infiltrate into

the soil during small and moderate stormevents. Performance monitoring, however, hasshown that drainage channels only realizethese benefits under ideal conditions (e.g., lowslope, sandy soils, dense grass cover, longchannel lengths, etc.—Dorman et al. 1989,Harper 1988, Yousef et al. 1985). When theseconditions are not met, drainage channels canhave a low or even negative removal capabilityfor many pollutants (MWCOG 1983, Dormanet al. 1989).

Only recently have engineers recognized thevalue of designing open channels explicitly forpollutant removal during small andmoderate–sized runoff events (Seattle METRO1992, Claytor and Schueler 1995). Channeldimensions are intentionally set to promotelonger residence times and/or to promotegreater runoff infiltration. Depending on thedepth to the water table, they are known as


_____________________________________________________________ 137

either grass channels, dry swales or wetswales. (Fig. 36). Checkdams, underdrains,stone inlets, prepared soil mixes andlandscaping are also used to enhance thepollutant removal capability of swales. The useof grass channels or swales along headwaterstreets is an economical and effective elementof a BMP system, as long as the criticalvelocity is not exceeded. In addition, openchannels in residential areas must be designedto prevent standing water, to ensure thatmowing and snow removal operations areconvenient, and to avoid odors, mosquitos orother nuisances associated with stagnant water.

Residential parking demand can be met byon–street parking lanes, private driveways orlots, or a combination of the two. Privatedriveways can usually meet the entire parkingdemand in larger–lot developments.Eventually, however, a threshold is crossedwhere parking demand can no longer be metsolely by private driveways. Typically, thistransition to on–street parking lanes occurs athousing densities of about 3 to 4 dwelling unitsper acre (Arendt 1994). Another residential street feature involvesturnarounds. If an access street has a dead end,provisions must be made to allow for vehiclesto conveniently turnaround. The most commonapproach utilizes a circular turnaround knownas a cul–de–sac. Traditional cul–de–sac designwere based on the turning radii needed forlarge vehicles–fire trucks, garbage trucks,moving vans, school buses, and resulted indiameters of 80 to 100 feet or more (Reed1991). As time as gone by, many communitiesrecognize that the diameter of cul–de–sac isexcessive (ULI 1990). A better alternative on

headwater streets are T–shaped orhammerhead turnarounds, which create lessimpervious cover. Pollutant Generation From Streets

Streets are a key source area in the urbanlandscape where stormwater pollutantsaccumulate. Pollutants can take a wide numberof pathways before they are trapped on thestreet surface (Fig. 36). While the deposition ofpollutants from the atmosphere by dryfall andwetfall remains the primary pathway, a streethas many other pollutants pathways. Forexample, trace metals such as cadmium, copperand zinc are often the product of gradual wearof tires or brakepads of cars as they pass overthe street. The contribution of metals from thispathway can be regionally significant; theSanta Clara NPS Program (1994) reported thatover 50% of the copper, cadmium and zinccould be attributed to this source. Cars are alsothought to be an important source ofhydrocarbons, through car emissions, leaks, orspills.

In northern climates, sand, salts or otherdeicing agents are applied for road traction andcan be a major seasonal pollutant source onstreet surfaces. Snow plowed to the curb or theright–of–way can also become a pollutantstorage area. Chemicals, grit and litteraccumulate in the road–side snowpack over anentire winter, and when snow melts in thespring, pollutant concentrations often exceed


_____________________________________________________________ 138

Open channels can be designed in one of four ways—as either (a) a drainage channel, (b) a grassed channel, (c) adry swale, or (d) a wet swale. All open channels are typically designed to convey the ten year design storm, andprevent critical erosive velocities during the two year design storm. The grass channel is designed to achieve acritical velocity during a water quality design storm. The dry swale is designed to capture and treat the entire waterquality volume in the swale. The same is true for the wet swale, except that the storage is provided by a pool of water,due to the presence of a high water table.

FIGURE 36: OPEN CHANNEL OPTIONS FOR RESIDENTIAL STREETS


_____________________________________________________________ 139

those recorded in warmer months (Oberts1994).

One poorly understood pathway is thebreakdown of the pavement surface itself.Little research has been performed todetermine whether asphaltic compounds arereleased shortly after resurfacing, gradually over time, or not at all.

Pollutant loads generally increase as averagedaily traffic volume increase. Runoffmonitoring by the Federal HighwayAdministration indicates that pollutantconcentrations are greatest along urbaninterstate highways with a traffic volume

greater than 30,000 ADT (Table 36). Ruralhighways are reported to have pollutant levelssimilar to those measured in residential andcommercial runoff (which are dominated bystreets and parking lots).

Lastly, the very nature of the street itself trapspollutants that blow in from outside areas. Eventhe modest vertical break of a curb sheltersairborne pollutants that may have blown in by thewind. Thus, dust, pollen, leaves, grassclippings, and organic matter can be trapped bythe curb, where they remain until they arewashed into the stormdrain system. Some ideaof the trapping potential of curbs and gutters isf o u n d i n t h e d a t e o f

TABLE 36: COMPARISON OF HIGHWAY AND URBAN RUNOFF POLLUTANT CONCENTRATION DATA

(ADAPTED FROM FHWA 1990 AND US EPA 1983)

Median PollutantConcentration

Highway (a)> 30,000 ADT

Highway (a)< 30,000 ADT

NURP RunoffData (b)

TSS (mg/l) 142 41 100

COD 114 49 65

Nitrate–N 0.76 0.46 0.82

TKN 1.83 0.87 0.68

Total Phosphorus 0.40 0.16 0.33

Copper (ug/l) 54 22 34

Lead 400 80 140

Zinc 329 80 160

Sources (a) FHWA runoff data N=993, (b) NURP Runoff data N=2300, includes primarily residential,commercial and mixed use sites that include roads and parking areas. ADT= Average Daily Traffic volume


_____________________________________________________________ 140

Curbs provide an effective trap for airborne pollutants, snow, vehicle emissions; as well as a very efficient means forwashing pollutants into the storm drain systems.

Bannerman (1994). Figure 37 shows thecomparative concentration of coliform bacteriafrom various urban sources. Bacterialconcentrations were one to two orders ofmagnitude higher in the street curbs comparedto parking lots or roof runoff. Residentialstreets, in particular, were discovered to havethe highest concentrations of harmful bacteria.

Once pollutants accumulate on a street surface,their delivery to the stream system is almostassured (Fig. 38). The crown of the streetdirects runoff over to the curb, where it is soonrouted to a storm drain inlets. Since the stormdrain network is designed to be self– cleansing,pollutants have a very high probability ofreaching the stream without modification.

FIGURE 37: BACTERIA LEVELS MEASURED FROM VARIOUS URBAN SOURCE AREAS ( BANNERMAN

1994)


_____________________________________________________________ 141

General Model of Residential StreetImpervious Cover

The amount of impervious cover created by aresidential street system can be analyzed usinga simple model. Total impervious cover (Ic) isthe sum of the impervious cover produced byfive residential street features:

Ic = R + Rw + S + T + D

where:

R = road length Rw = road widthS = sidewalksT = turnarounds, and D = driveways

The specific amount of impervious covercreated by each individual residential streetfeature is easily computed, based on theiraverage length and width, which is governedby the prevailing subdivision codes of acommunity. For example:

Road length (R). The minimum distance of theroad for a given zoning category is computedas:

R = {(Number of Dwelling Units) (AverageFeet of Frontage Required)}/2 As the equation shows, the only significantway to reduce road length is to lower theminimum frontage requirement, which istypically set by local subdivision codes.Communities that allow designers flexibility inthis key lot dimension to promote streamprotection cluster can greatly reduce theamount of impervious cover created, regardlessof the road width. As was discussed in Chapter

4, stream protection cluster can reduce thelength of the road network by as much as 50%.

Road width (Rw). Once the minimum lengthof the road network is established, the next keyvariable to address is road width. If thetraditional residential street width of 36 feet isused as a baseline, we can analyze the impactof narrower headwater streets on the creationof impervious cover. This relationship isgraphically displayed in Figure 39, whichshows the number of impervious acres createdas a function of the number of dwelling unitsand street width, for single family homessituated on one–acre lots. As an example,consider a subdivision containing 50 homes. A36–foot–wide street system creates about 2.5acres of impervious cover, while a 26–footstreet creates only 1.8 acres of imperviouscover, or a savings of 28%. The use of evennarrower headwater streets (16 or 20 feet)produce even greater savings. Sidewalks (S). The next residential streetvariable are sidewalks used for pedestrianmovement. Most communities require that theybe installed on one or both sides of a street.The minimum width of sidewalks is four feet(which allows a wheelchair adequate passage,under the requirements of the Americans withDisabilities Act), but some communities oftenrequire that they be five or even six feet wide.Based on these parameters, the calculation ofimpervious cover created by sidewalks can becomputed in a straight forward manner:


_____________________________________________________________ 142

FIGURE 39: IMPERVIOUS COVER CREATED AS A FUNCTION OF ROAD WIDTH AND NUMBER OF

DWELLING UNITS SERVED

The model of residential street impervious cover indicates that significant reductions in impervious cover can beachieved through narrower streets.

S = (Sn) (R) (Sw)

where:

Sn = local sidewalk requirements (0, 1 or2 sides).

R = road length (linear ft.) Sw = local sidewalk width (4, 5 or 6 ft.)

Using the same example of a 50 acre singlefamily home subdivision, we can thendetermine the amount of impervious covercreated by sidewalks, using differentassumptions about local requirements (Fig. 40).The acreage of impervious cover can be seento range from none to slightly less than oneacre (0.83 acres).

Turnarounds (T). Dead end streets inresidential subdivisions are usually required tohave an acceptable option for vehicles toturnaround, with the circular cul–de–sac beingthe most common. A range of five differentturnaround options are depicted in Figure 41.In each case, they provide a minimum internalturning radius of 17 to 20 feet to accommodatethe larger vehicles. The sharp differences in theamount of impervious cover produced by eachturnaround option is shown


_____________________________________________________________ 143

Five options for turnarounds include (a) 40 foot radius circle (b) a 30 foot radius circle,(c) a 40 foot circle withpervious donut, (d) a 30 foot radius circle with pervious donut and (e) a 60 by 20 foot “T”—shaped or hammerhead.

FIGURE 40: IMPERVIOUS COVER CREATED BY TYPICAL RESIDENTIAL SIDEWALK STANDARDS

Subdivision requirements on both the width and number of sidewalks can create almost an acre of impervious coverin a 50 acre residential subdivision.

FIGURE 41: FIVE TURNAROUND OPTIONS AT THE END OF A RESIDENTIAL STREET


_____________________________________________________________ 144

The greatest impervious reduction is achieved using either a hammerhead, or by reducing the radius of a circularturnaround. Donuts have a minor effect on the amount of impervious cover created.

in Figure 42. It is clearly evident that largecul–de–sac radii create needless imperviouscover, even if a landscaped “donut” is installedin the center (see Figure 41). Approximately50% less impervious cover is created simply bydropping the radius from 40 feet to 30 feet. Aneven greater reduction occurs when T–shapedor “hammerhead” turnarounds are used. Inhammerheads, a vehicle must make a requiredthree–point turn to completely reversedirection, compared to a two point turn (in 30foot radius cul–de–sacs).

Driveways (D). The last residential streetfeature that creates impervious cover areprivate driveways. Most communities requirea standard width of 20 feet for the driveway(with a slightly greater apron width, where the

driveway meets the street). A driveway thiswide allows two cars to be easily parked sideby side. Deriving the average length of thedriveway is a more complicated affair. Mostcommunities specify that homes must be setback a fixed distance from the streetright–of–way, depending on the residentialzoning category. This requirement issometimes expressed as a fixed length (40 or60 feet), or a percentage of the lot depth (e.g.,the home must be set back at least 40% of thedistance between the front yard and backyardboundary). Since the driveway needs to extendfrom the street to the home, its length isfundamentally determined by the setbackrequirement. Thus, if the setback is 60 feet, thelength of the driveway will be at least 60 feetplus the 12 or 15 feet of right–of–way from thestreet (for a total of 72–75 feet).

FIGURE 42: IMPERVIOUS COVER CREATED BY EACH TURNAROUND OPTION


_____________________________________________________________ 145

Driveways contribute a surprising amount of impervious cover in the landscape. The amount increases as drivewaylengths become greater (due to front yard setbacks) or housing density increases.

The last variable that we must define before wecan compute driveway impervious area is thetotal number of driveways needed per linearfeet of street (N). This is computed simply as:

N = R / (Fd/2)

where:

R = road length (linear feet)Fd = average frontage distance in feet. which allows us to compute the imperviouscover associated as driveways (D) as: D = (N) (Dl)(Dw)

where:

N = number of drivewaysDl = average driveway length = front yard

setback

Dw = driveway width = 20 feet, but can be12 feet in rural areas.

Driveways contribute a major share ofimpervious cover in residential streets. Usingthe 50–acre single family home subdivisionexample presented earlier, anywhere from 1.0to 1.5 acres of impervious cover can be createdsolely by driveways. Driveways tend to createmore impervious cover when lot size decreases(and a greater number of individual drivewaysare needed (Fig. 43). The strong influence ofdriveway length on the creation of imperviouscover is also evident in the figure. It should benoted that after about 40 feet, a driveway canfully meet the parking demand for most homes,and additional length only functions to connectthe street to the home.

FIGURE 43: DRIVEWAY IMPERVIOUS COVER AS A FUNCTION OF LENGTH AND LOT SIZE


_____________________________________________________________ 146

FIGURE 44: IMPERVIOUS COVER CREATED UNDER THREE RESIDENTIAL STREET DESIGN

SCENARIOS

The residential street model is useful in analyzing the effect of local street standards on the creation of imperviouscover. This example shows the difference between the most generous street standards, and the proposed headwaterstreet standards.

The simple model presented above has greatvalue in identifying the best opportunities toreduce impervious cover created by residentialstreets. For example, consider the case of asimple 1,000 foot dead–end street that servessingle family homes situated on one–acre lots(Fig. 44). Using the model, we can see therelative share of impervious cover created byeach of the five residential street features undertraditional subdivision design standards usingthe following assumptions:

R = 1,000 feetRw = 36 feetS = 5 feet wide, both sides of the streetT = one 40 foot radius cul–de–sac.D = 15 driveways that average 75 feet in

length.

The total amount of impervious cover createdunder this scenario is about 1.7 acres (Fig. 45).The value of alternative headwater streetdesign scenarios can then be assessed bymodifying design assumptions, such as shownbelow.

Scenario A: Scenario B:Headwater Streets Headwater Streets and ½ acre cluster

R 1,000 500Rw 20 feet 26 feetS 4 foot, one–side 4 foot, one sideT hammerhead 30 ft radius

cul–de–sacD 15 @ 60 feet 15 @ 45 feet

The cumulative impact of these measuresresults in a 40% (scenario A) or 60% (scenarioB) reduction in impervious area.


_____________________________________________________________ 147

FIGURE 45: RELATIVE FRACTION OF TOTAL IMPERVIOUS COVER BY THE FIVE RESIDENTIAL

STREET DESIGN COMPONENTS

.

While street width is important in the creation of impervious cover, a large fraction is produced by other features, suchas driveways, sidewalks, and turnarounds. Communities may want to examine subdivision code requirements thatinfluence those features when designing headwater streets.

Benefits of Headwater Streets

Beyond their obvious benefit in reducing siteand watershed imperviousness, headwaterstreets can provide many other environmentaland economic benefits. For example,headwater streets:

Q reduce the amount of clearing and grading needed at the development site

Q allow for treatment of runoff adjacent to the street

Q reduce speeds in neighborhoodsQ make a neighborhood more pedestrian

friendly Q reduce the capital construction costQ preserve more area for lotsQ reduce road maintenance costsLocal Experience with Headwater Streets

Most communities in the US have yet to takeadvantage of the benefits of headwater streets.The gradual evolution in residential streetdesign standards has been well documented byStabenfeldt (1995) and ULI (1989) and issummarized below:

1. Most local governments model theirresidential street design standards after stateand/or federal highway criteria, although thetraffic capacity and function of their streetsystem is considerably different fromhighways. The key reasons for the lack oflocally developed residential street designs,according to Stabenfeldt (1995), areperceptions about congestion, emergencyaccess, liability and parking demands. Manylocal traffic engineers have simply accepted the


_____________________________________________________________ 148

notion that wider streets adequately addressthese concerns, and that wide streets are saferstreets.

In most regions of the country, localgovernments have authority to create narrowerdesign standards for residential streets.Exceptions include a number of states that stillretain control over local street design. Forexample, state oversight or review is stillrequired for local road construction inConnecticut, Montana, North Carolina, Texasand Virginia.

2. Consequently, very few communitiesrecognize any local road categories that aredifferent from established state and federalstreet categories. A recent national survey ofcounties conducted by the Urban LandInstitute (1990) indicated that fewer than halfrecognize street standards that are substantiallydifferent from those of ITE and AASHTO.

3. Road dimensions have increased sharplyover time, indeed, the total width of residentialstreets has increased by over 50% since theSecond World War, in response to concernsabout safety, traffic flow, emergency access,spillover parking, more utilities, pedestriansafety, snow removal and liability (ULI 1989).Traffic engineers have adopted a uniformstandard for roads, regardless of the trafficgenerated by the road. Thus, street functiondominates over actual land use as the primarydeterminant of road type. Experience andcommon sense, have shown that theseconcerns can be fully met by narrower streetswhen traffic volume is light (ULI 1989).

4. A number of communities haveexperimented with narrower residential streets,and found them to be an attractive and safealternative. These include Bucks County PA,Portland OR, Boulder CO, Dade County FL,Olympia WA and King County WA (ULI1989; Bray and Rabiner 1991; Fernandez1994; Wells 1994; and Bucks County 1980).A summary of some of the residential streetterminology and geometry that have been usedin these communities is provided in Tables 37and 38, respectively.

Performance Criteria for Headwater Streets

The overall objective for the design ofheadwater streets is to reduce needlessimpervious cover while still meetingcommunity needs for safety, traffic flow andparking. Where feasible, headwater streets arealso designed to utilize grass channels orswales to maximize pollutant removal andstormwater infiltration. Seven recommendedperformance criteria for headwater streetdesign are to: 1. Reduce total road length2. Design narrower headwater streets3. Limit right–of–way4. Reduce number and size of cul–de–sacs5. Limit driveway lengths6. Design safe pedestrian movement7. Use open channel stormwater treatment The performance criteria are intended to be


_____________________________________________________________ 149

TABLE 37: DESIGN STANDARDS FOR HEADWATER STREETS IN THREE LOCAL COMMUNITIES

(SOURCE: STABENFELDT 1995)

Source/Street Type Volume

(ADT)DesignSpeed

Right-of-Way

PavementWidth

Parking Curb andGutter

BUCKS COUNTY, PAAccess Streets 200 25 mph 16 ft

18 ft26 ft

nonenoneone side

not requirednot requiredrequired

Residential Subcollector 200-1,000

20 ft22 ft28 ft36 ft

nonenoneone sideboth sides

not requiredrequiredrequiredrequired

Residential Collectors to3,000

20 ft22 ft

nonenone

not requirednot required

Special Purpose Streets/Alleys 12 ft

BOULDER, COAccess Lane

150 15 mph 28 ft 20 ft allowed

Access Street 350 20 mph 48 ft 20 ft22 ft26 ft

noneone sideboth sides

required

Residential Street 500-1,000

25 mph 48 ft 20 ft26 ft32 ft


required

Residential Collector 1,000-3,000

25 mph 50 ft 22 ft28 ft34 ft


required

Alley 20 ft 18 ft none

PORTLAND, ORThrough & Cul–de–sac

n.r. n.r. 35 ft 20 ft one side n.r.

Queuing n.r. n.r. 40 ft 26 ft both sides n.r.


_____________________________________________________________ 150

TABLE 38: HEADWATER STREET CLASSIFICATIONS USED IN THREE LOCAL COMMUNITIES

(SOURCE: STABENFELDT 1995)

Locality Street Type/Definition

Bucks County ResidentialAccess

ResidentialSubcollectors

ResidentialCollectors

Have the sole purpose of providing frontage for service andaccess to private lots.

Are access streets which provide frontage for residential lotsand may carry a small amount of through traffic “collected”only from through tributary access streets.

Conduct and distribute traffic between other residentialstreets of lower order in the streets hierarchy and higherorder streets or major activity centers.

Boulder, CO Access Lane

Access Street

ResidentialStreet

ResidentialCollector

Designed exclusively for access to a limited number ofproperties, serving no more than 15 properties.

Provides access to a limited number of properties, servingno more than 25 units.

Designed to provide access to individual properties and toalso provide access to the subcollector and collector.

Design to provide access to individual properties and tostreets of lower and higher function. They are alsodesigned to accommodate higher traffic volumes with someof the tries using these streets to access the collectors andarterial street network.

Portland, OR QueuingStreet

Local TrafficStreet

Intended for two–way traffic; are comprised of a singletraffic lane and a parking lane on one or both sides.Queuing streets are possible for both cul–de–sac andthrough streets.

Permits two travel lanes plus on–street parking on bothsides of the street.


_____________________________________________________________ 151

general in nature; communities may choose todevelop more specific criteria after analyzingthe impervious cover created by their currentstreet classifications and design standards.

Criteria 1. Reduce the total road length neededto serve residential development.

The road network of a residential developmentshould be the shortest possible length neededto serve the total number of dwelling units.Four simple techniques that can be used tocreate a shorter road network are:

Q stream protection cluster (cf Chapter 4) Q grid or curvilinear road patterns that

serve more lots per unit road length thanlooping or branching patterns

Q shorter frontage requirements for individual lots (to a minimum frontage

distance of 60 ft) also allow more lots toserved per unit road length. A smallnumber of flag lots may be permitted ifthey act to reduce road length as well

Q shorter centerline radii requirements forroad turns. For many headwater streets,a centerline radii of 150 feet is sufficientfor safe turning

Criteria 2. Reduce the pavement width ofheadwater streets that carry less than athousand ADT and serve less than fourdwelling units per acre.

The pavement width of smaller residentialstreets should be as narrow as possible. Somerecommended pavement widths for a range ofheadwater street sizes are shown in Tables 35and 38.

Narrower headwater streets typically have oneless parking lane and/or moving lane,depending on their traffic volume and on–streetparking needs. These narrower streets mayoccasionally require that one car must pull intoa parking lane to let an oncoming car pass. Forheadwater streets that carry less than 500ADT, this queuing behavior presents little orno inconvenience to motorists.

Current local street classifications and designstandards should be analyzed to identifyopportunities for creating narrower headwaterstreets. The new design standards willundoubtedly vary somewhat from one localityto another, based on such factors as residentialparking demand, lot size, curb and gutterrequirements, design speeds, emergency accessand snow storage.

Criteria 3. Make right of way requirementsonly as wide as needed to accommodatestructures that are actually built on each sideof the street.

Since most headwater streets will never bewidened in the future, a narrower right of wayof 8 to 15 feet on each side of the street can beeasily justified. A narrower right of way canreduce the need for tree clearing or grading. Inmost cases, the right of way should include(from the pavement edge outward): a 3 to 5foot wide grass strip, a four foot widesidewalk, and an extra foot of grass. Utilitiesand storm drains are located underneath thestreet or within the right of way. Some otherconsiderations to keep in mind when settingright of way include:


_____________________________________________________________ 152

G If utility corridors or sidewalks are installedon only one side of the street, the right ofway should be reduced proportionately.

G A wider right of way may be needed forgrass channels or swales

G Clearing of trees within five feet of thepavement edge should be avoided in forestareas, wherever possible

G In snow regions, the grass strip area in theright of way may need to be increased to 6to 8 feet for storage of plowed snow

Criteria 4. Reduce the use and effective radiusof cul–de sac turnarounds

Large cul–de–sacs are expensive, unattractive,and create needless impervious cover. Their useshould be discouraged. Since headwater streetsare located at the end of the road network, theprimary design objective is to get directemergency access to homes. To turnaround andgo back, however, may occasionally require atwo point turn. Thus, for most headwater streetsserving less than 25 homes, a minimumcul–de–sac open turnaround radius of 30 feet isrecommended. A landscaped donut can beplaced in the center of the cul–de–sacturnaround as long as it maintains an internalturning radius of 17 to 20 feet.

Alternative turnarounds, such as the T–shaped“hammerhead,” create less impervious coverthan any circular option, and should beencouraged in shorter cul–de–sacs, particularlyin rural areas.

Criteria 5. The length of driveways should belimited to satisfy residential parking demandand access requirements.

As noted earlier, driveways generate asurprisingly large fraction of the imperviouscover created by a residential street. Drivewaylength can be limited to 30 to 40 feet in mostlarge lot residential lots that have two cargarages, and still fully meet residential parkingdemand. The key site design parameter thatinfluences the length of driveways; however, isthe mandatory front yard setback to the home.Local codes should be reviewed to determine ifexcessive setback requirements of 60 to 75 feetcan be modified to accommodate shorterdriveways. Some other techniques for reducingthe impervious cover created by drivewaysinclude:

G shortening the minimum driveway widthfrom 20 to 18 feet.

G limiting impervious surfaces to two tracks,with the remainder of the driveway in grassor pervious surface

G utilizing a shared driveway to connect 3 or4 units together (rather than a wider road)

Criteria 6. Design for safe pedestrianmovement through the community

Safe pedestrian movement does not alwaysentail wide sidewalks on both sides of thestreet. Indeed, many adults and children stillmove through a community along the street,even when sidewalks are available. The keysafety consideration is that headwater streetshave both low traffic speeds and goodvisibility. If these considerations are met, it ispossible to relax the double–wide sidewalk


_____________________________________________________________ 153

requirement. For example:

G sidewalks can be located on only one sideof the street

G sidewalks can be replaced by walkwayslocated within community open space andaway from streets

G the width of sidewalks can be as narrow asfour feet

G sidewalks should be graded so thatstormwater runoff travels to the front yardand not into the street

Criteria 7. Wherever possible, the right of wayof headwater streets should be utilized for thetreatment of stormwater quality using opengrass channels.

Headwater streets provide engineers with anideal location for effectively treating the qualityof stormwater runoff near its source. The basictechnique is to design an open, vegetatedchannel within the right of way (and eliminatecurb and gutters).

In general, road designers should have todemonstrate that open channels are not feasibleon each headwater street before any curb andgutter are accepted. The following conditionsare evidence that a headwater street cannotsupport an open channel:

G longitudinal slopes greater than 5%G computed runoff velocities for the two

year design storm event that exceed thecritical erosive velocity of 4 to 5 feet persecond

G local climate or soils make it impossible toestablish dense turf throughout the year

G presence of the water table within a footbelow the proposed channel bottom,

G or a housing density exceeding 3 dwellingunits per acre

Headwater Street BMPs

If an open channel remains feasible for the site,designers can use one of four open channeldes igns fo r runof f t r ea tmen t—drainage channels, grassed channels, dryswales and wet swales (see Figure 36). Somedesign guidance on each design option areoutlined below:

a. Drainage ChannelsThis form of open channel is solely designed tohave enough capacity to safely convey runofffrom large storm events without erosion. Thechannel cross–section has a hydraulic capacityto handle the peak discharge rate for the tenyear storm event, and channel dimensions (i.e.,slope and bottom width) are carefully selectedso that the critical erosive velocity is notexceeded during the peak discharge rate for thetwo year storm event. Since drainage channelsare not explicitly designed to treat runoff frommore frequent storm events, they provide onlylimited water quality benefits, unless soils areextremely sandy (MWCOG 1983).Consequently, the use of drainage channels isprimarily restricted to runoff pretreatment (i.e.,trapping coarse sediments in the channel beforethey are delivered to a downstream pond,wetland, filter or infiltration facility).

b. Grass ChannelsGrass channels are different from drainagechannels in that they are designed to meetrunoff velocity targets under three stormconditions—a water quality design storm, thetwo year design storm and the ten year designstorm (see Figure 46). In addition, the total


_____________________________________________________________ 154

length of the channel must provide at least tenminutes residence time for the water qualitystorm. In some regions of the country, grasschannels are termed “biofilters” (SeattleMETRO 1992). To meet these criteria, grasschannels have broader bottoms, lower slopesand dense vegetation. Performance monitoringhas shown that grass channels are more areliable technique for removing pollutants fromstormwater than drainage channels.Reasonably high removal rates have been

reported for sediment and hydrocarbons.Grassed channels, however, have proven to beless effective in removing soluble nutrients,soluble metals or bacteria. In addition, fieldassessments indicate that many grassedchannels are not constructed to specification,lack dense vegetation or have standing water(Horner 1988), and may not be suitable for allresidential sites. Thus, while grass channelsmay satisfy local requirements for stormwaterquality treatment, they need to be carefullyconstructed, inspected and maintained. Somegeneral guidance on grass channel design canbe found in Table 39, and detailed designmethods are provided in Seattle METRO1992, and Claytor and Schueler 1995.

FIGURE 46: SCHEMATIC OF A GRASSED CHANNEL


_____________________________________________________________ 155

TABLE 39: DESIGN GUIDANCE FOR BIOFILTER SWALES (ADAPTED FROM HORNER 1988,REEVES 1995)

Geometry:Preferred geometry minimizes sharp corners and has gentle slopes, parabolic or trapezoidal shapes, withsideslopes no greater than 3:1 (h:v).

Longitudinal Slope: Should be in the range of 2 to 4%. Checkdams should be installed if slopes exceed 4% and underdrainsinstalled if slopes are less than 2%.

Swale Width: Should be no wider than 8 feet, unless structural measures are used to spread flow.

Maximum Residence Time:Try to achieve a hydraulic residence time for the 6 month 24 hour storm of about 9 or 10 minutes.

Maximum Runoff Velocity:no more than 0.9 fps for 6 month, 24 hour storm, and 1.5 fps for 2 year storm event.

Manning’s n Value: Recommend the use of a 0.20 value in design.

Grass Height: Normal grass height should be at least two inches above design flow depth.

Mowing: Routine mowing is used to keep grass in active growth phase, and to maintain dense cover.

Biofilter Soils: A sandy loam topsoil layer, with an organic matter content of 10 to 20%, and no more than 20% clay. If soiltest indicates that the current soil does not meet these criteria, a surface layer topsoil amendment may be used.

Water Table: Designer should check to determine the level of the seasonally high water table. If it is within a foot of thebottom of the biofilter, it may be advisable to select wetland species.

Plant Selection: Select grass species that produces a uniform cover of fine–hardy vegetation that can withstand the prevailingmoisture condition. Wetland adapted species such as Juncus and Scirpus may be utilized if drainage is poor.

Landscaping: Other plant material can be integrated into a biofilter; but care should be taken to prevent shading or leaf fallinto swale.

Construction: Use of manure mulching or high fertilizer hydroseeding to establish ground cover should be avoided during

construction, as these can result in nutrient export.


_____________________________________________________________ 156

c. Dry SwaleDry swales are designed to completely storethe runoff volume from the water qualitystorm event and filter it through 30 inches ofswale soil before it is collected by anunderdrain (Fig. 47). Consequently, dryswales are expected to have the highestremoval rates of any open channel system(Yousef et al. 1985, and Harper 1988). Toachieve such rapid rate of infiltration, it isoften necessary to modify the parent soils toimprove their infiltration rate and/or to allowup to 18 inches of temporary ponding abovethe swale. Some other key design criteria forthe dry swale include:

G Pretreatment is required to protect theswale. For pipe inlets, 0.1 inch percontributing acre should be temporarily

stored behind a checkdam. For lateralinflows, gentle slopes or a pea graveldiaphragm can be used.

G Swales include a prepared soil filter bedthat is 30 inches deep and composed of50% sand and 50% silt loam.

G Swale filter beds are drained by alongitudinal perforated pipe to keep theswale dry after storm events.

G Swales are parabolic or trapezoidalshapes, with gentle side–slopes (nogreater than 3:1 h:v), and bottom widthsranging from 2 to 8 feet.

G Geotechnical tests are required todetermine the location of the water table.

FIGURE 47: SCHEMATIC OF A DRY SWALE


_____________________________________________________________ 157

The dry swale is the preferred open channeloption for most residential settings since it isdesigned to prevent standing water problemsthat generate homeowner complaints. Theswale is designed to rapidly dewater, therebyallowing front yards to be easily mowed.Design methods for dry swales can be foundin Claytor and Schueler (1995).

d. Wet SwalesIn some regions of the country, the water tableis located very close to surface. When swalesare excavated in these regions, it is likely thatsoils will be fully saturated, or standing waterwill be present in the swale during all or parto f t h e y e a r . T h e b e s t

design in these situations is the wet swale(Figure 48), which essentially acts as a verylong and linear pocket wetland. Monitoringstudies in Florida indicate that wet swales canhave reasonably high pollutant removal rates(Yousef et al. 1985 and Harper 1988). Itshould be clearly noted that wet swales areoften not appropriate in many residentialsettings, as the public seldom accepts standingor stagnant water in their front yards.Application of wet swales is thus limited tolarge lot rural developments, and headwaterstreet sections that do not front individual lots.Design methods for wet swales, includingwetland plant selection, can be found inClaytor and Schueler (1995).

FIGURE 48: SCHEMATIC OF A WET SWALE


_____________________________________________________________ 158

Runoff Treatment Options for EnclosedStorm Drain

If open channels are not a feasible drainageoption for a headwater street, then stormwaterquality treatment must be provided within thestorm drain network (in–line treatment) or atthe outfall of the storm drain (end–of–pipetreatment).

The in–line treatment strategy involvestrapping sediments within special structureswithin the storm drain network. Theseunderground structures include inlets,catchbasins, sump pits or oil/grit separators.As a general rule, these structures often havelimited storage capacity compared to the largeflow rates they must handle during intensestorm events. Consequently, pollutantstrapped within the structures during smallerstorm events often are resuspended duringlarger ones. Pollutant storage is temporary,and good removal rates can only be achievedwhen pollutants are physically removed on afrequent basis. Recent research hasdemonstrated that trapped pollutants must becleaned out almost monthly if meaningfulpollutant reduction is expected (Mineart andSingh 1995, Schueler and Shepp 1993). Localpublic works departments should carefullyevaluate whether they have adequate budgetsand staff to perform frequent cleanouts beforeaccepting in–line treatment options.

The end–of pipe strategy has been the mostcommon approach for treating the quality ofstreet runoff. Runoff is rapidly delivered to adownstream point where it is treated in a largestormwater pond, wetland or filtering system.

Resources Needed for Implementation

A great deal of local inertia must be overcometo modify existing residential street designstandards. The greatest barrier to realizingheadwater streets are often local highway,public works and fire department personnelwho may be reluctant to change the way thingshave always been done. Consequently, it isoften necessary to create an interagency taskforce that includes each agency to oversee thedevelopment of revised residential streetdesigns. Through this outreach process, manycommunities have found more commonground than they had expected (Wells 1994).Once agreement has been reached, headwaterstreet standards will still need to be formallyadopted in subdivision codes or ordinances,following a local public involvement plan. References AASHTO. 1990. A Policy on GeometricDesign of Highways and Streets. Washington,DC. 46 pp.

Arendt, R. 1994. Rural by Design–maintaining small town character. AmericanPlanning Assoc. Washington, DC. 460 pp. Bannerman, R. 1994. Sources of UrbanStormwater Pollutants Defined in Wisconsin.Wat. Prot. Techniques. 1(1): 30–32. Bucks County Planning Commission. 1980.Performance Streets–a concept and modelstandards for residential streets. Bucks County, PA. 46pp.

Bray, T and K. Rabiner. 1991. Report on New


_____________________________________________________________ 159

Standards for Residential Streets in Portland,Oregon. Bureau of Transportation Engineering.Portland, Oregon.

Carrol County, 1992. Design Manual-Roads andStorm Drains. Dept. Of Public Works. Book No.0190. Bureau of Engineering. Carrol County, MD.

Claytor, R. and T. Schueler. 1995. Design ofStormwater Filtering Systems. Chesapeake BayResearch Consortium. Center for WatershedProtection, Silver Spring, MD 202 pp.

Dorman et al. 1989. Retention/detention andoverland flow for pollutant removal in highwayrunoff. FHWA/RD – 89/202 pp

Federal Highway Administration. 1990. PollutantLoadings and Impacts from Highway StormwaterRunoff. FHWA– RD88–006. Washington, DC. 440pp.

Fernandez, J. 1994. Boulder Brings BackNeighborhood Streets. Planning. 60(6): 21–26.

Harper, H. 1988. Effects of StormwaterManagement Systems on Groundwater Quality.Final Report. Environmental Research and Design,Inc. Florida Dept. of Environmental Regulation. 460pp.

Horner, R. 1988. Biofiltration systems for stormrunoff water quality control. Washington Dept. ofEcology. 84 pp.

Institute of Transportation Engineers. 1987b. TripGeneration Rates Washington, DC.

Institute of Traffic Engineers. 1991. ResidentialStreet Design Guidelines– recommended practice.Washington, DC. 68 pp.

Metropolitan Washington Council of Governments.1983. Final Report. Nationwide Urban Runoff Project.Department of Environmental Programs. US EPA.Washington, D.C. 222 pp.

Mineart, P and S. Singh. 1995. The value of morefrequent cleanouts of storm drain inlets. Wat. Prot.Techniques. 1(3): 129–130.

Oberts, G. 1994. Influence of Snowmelt Dynamics onStormwater Runoff Quality. Wat. Prot. Techniques.1(2): 55–61.

Reed, C. 1991. Subdivision Development andConstruction Standards. Part 1:Street ROW andCross Sections. The Zoning Report. 9(14):1–8.

Reeves, E. 1995. Performance and Longevity ofBiofilters in Washington, Wat. Prot. Techniques 1(3):117–120.

Santa Clara Valley NPS Program. 1994. Cars areLeading Source of Metal Loads in California. Wat.Prot. Techniques. 1(1): 28.

Schueler, T. and D. Shepp, 1993. The quality oftrapped sediments in oil grit separators in suburbanMaryland. MWCOG. MDE. Washington, DC. 88pp.

Seattle METRO. 1992. Biofiltration swaleperformance: recommendations and designconsiderations. Publication No. 657. Washington Dept.Of Ecology. 220 pp.

Stabenfeldt, L. 1995. Residential street strategies forurban watersheds. Environmental Land PlanningSeries. Metropolitan Washington Council of Gov’ts.US EPA. Washington, DC. 66 pp.

US Environmental Protection Agency. 1983. Results


_____________________________________________________________ 160

of the National Urban Runoff Monitoring Project.Washington, DC.

Urban Land Institute, 1989. Survey of ResidentialStreet Design. ULI. Washington, DC. 78. pp.

Urban Land Institute. 1990. Residential Streets.(second edition). American Society of Civil

Engineers and National Association ofHomebuilders.106 pp.

Wells, C. 1994. Impervious Surface ReductionStudy. Draft Report. City of Olympia Public WorksDepartment. Washington Department of Ecology.

182 pp..

Williams, K, T. McCauley and M. Wyckoff. 1990.Land division and access controls. Planning andZoning Center. Michigan Society of PlanningOfficials, Lansing, MI.76 pp.

Yousef, Y., M. Wanielista, H. Harper, D.Pearce and R. Tolbert. 1985. BestManagement Practices – removal of highwaycontaminants by roadside swales. FinalReport. University of Central Florida. FloridaDept. Of Transportation. 166 pp.

Documents

Chapter 6 Headwater Streets · The greatest share of total impervious cover in most communities is from the roads, sidewalks, parking lots and driveways used to get us to where we