ARTICLE IN PRESS

0301-4797/$ - se

doi:10.1016/j.je

�CorrespondE-mail addr

srinivasvalavala

(F. Montes).

Journal of Environmental Management 81 (2006) 42–49

www.elsevier.com/locate/jenvman

Permeability predictions for sand-clogged Portland cement perviousconcrete pavement systems

Liv M. Haselbach�, Srinivas Valavala, Felipe Montes

University of South Carolina, Civil and Environmental Engineering, 300 Main Street, Columbia, SC 29208, USA

Received 11 April 2005; received in revised form 2 August 2005; accepted 30 September 2005

Available online 24 March 2006

Abstract

Pervious concrete is an alternative paving surface that can be used to reduce the nonpoint source pollution effects of stormwater runoff

from paved surfaces such as roadways and parking lots by allowing some of the rainfall to permeate into the ground below. This

infiltration rate may be adversely affected by clogging of the system, particularly clogging or covering by sand in coastal areas. A

theoretical relation was developed between the effective permeability of a sand-clogged pervious concrete block, the permeability of sand,

and the porosity of the unclogged block. Permeabilities were then measured for Portland cement pervious concrete systems fully covered

with extra fine sand in a flume using simulated rainfalls. The experimental results correlated well with the theoretical calculated

permeability of the pervious concrete system for pervious concrete systems fully covered on the surface with sand. Two different slopes

(2% and 10%) were used. Rainfall rates were simulated for the combination of direct rainfall (passive runoff) and for additional

stormwater runoff from adjacent areas (active runoff). A typical pervious concrete block will allow water to pass through at flow rates

greater than 0.2 cm/s and a typical extra fine sand will have a permeability of approximately 0.02 cm/s. The limit of the system with

complete sand coverage resulted in an effective system permeability of approximately 0.004 cm/s which is similar to the rainfall intensity

of a 30min duration, 100-year frequency event in the southeastern United States. The results obtained are important in designing and

evaluating pervious concrete as a paving surface within watershed management systems for controlling the quantity of runoff.

r 2006 Elsevier Ltd. All rights reserved.

Keywords: Runoff; Permeability; Pervious concrete; Stormwater management

1. Introduction

Land development can augment the amount of imper-vious surfaces on a site so that stormwater runoff and itsassociated pollutant loads to receiving waters are increased(Barrett et al., 1998). Pervious concrete is being used as oneof the solutions for decreasing the runoff by allowing rainwater to drain into the land surface. Pervious concrete is aspecially placed concrete mixture with little or no fineaggregate where there are many more larger pores throughthe structure than in conventional concrete (Tennis et al.,2004).

e front matter r 2006 Elsevier Ltd. All rights reserved.

nvman.2005.09.019

ing author. Tel.: +1803 777 8318; fax: +1 803 777 0670.

esses: haselbac@engr.sc.edu (L.M. Haselbach),

@richlandonline.com (S. Valavala), montesf@engr.sc.edu

Runoff for any given rainfall event depends on thepotential permeability of the land surface (Zouaghi et al.,2000). The porosity of pervious concrete usually rangesfrom 15% to 30% and typically water flows through alayer of pervious concrete with rates from 0.2 cm/s togreater than 1 cm/s depending on the materials andplacement (Montes et al., 2005; Tennis et al., 2004). Thisrate is greater than most rainfall rates and the perviousconcrete surface is many times designed to handle bothdirect rainfall (passive runoff) and runoff from upslopeareas (active runoff).As a pavement surface, the pervious concrete layer is

placed over a subbase, which can consist of the naturalsoils or a specially designed subbase such as crushed rockor gravel. In this combination, the infiltration rate ofstormwater through the pervious pavement system is manytimes limited by the permeability of the subbase. Typicalextra fine sand (for example Foster Dixiana sandblast

ARTICLE IN PRESS

Nomenclature

AB the total surface area of the pervious concreteblock (cm2)

AP the surface area of the pervious concrete blockoccupied by pores (cm2)

I intensity of rainfall (cm/h or cm/s)P average porosity of the block is the ratio of the

volume of the voids to the total volume of theblock given in percent.

Ptop average porosity of the top quarter of the blockas determined by an equation developed fromlaboratory analyses of other blocks taken fromthe same slab and given in percent

ksand permeability of sand (cm/s)keff theoretical effective permeability of sand-

clogged or covered pervious concrete blocksystems (cm/s)

kclog experimental permeability of sand-clogged per-vious concrete block system (cm/s)

L.M. Haselbach et al. / Journal of Environmental Management 81 (2006) 42–49 43

abrasive BX-40) may have a permeability of about0.023 cm/s, which is an order of magnitude or more lessthan the pervious concrete layer permeability, and limitsthe flow through the system (Valavala et al., 2006).However, the pervious concrete top layer is full of voidsthat can be used for water storage during a rainstorm andtherefore may prevent stormwater from producing runoffuntil the system is fully saturated.

In many coastal areas, where a pervious concretepavement may be placed over sand, there is concern aboutthe pervious concrete pores becoming clogged or coveredwith blowing sand and the effect that this will have on thesystem. Initially, it may appear that the system will then belimited for its infiltration capacity of both passive andactive runoff by the permeability of the sand. However, theauthors propose that the system permeability will insteadbe reduced by this extreme condition to a fraction of thepermeability of the sand, where this fraction can berepresented by the porosity of the pervious concretesurface as

keff ¼ ðPtop=100Þksand. (1)

2. Background

There is little literature associated with clogging experi-ments and none particular to pervious concrete. There havebeen experiments analyzing the reduced infiltration rates offield-installed concrete block pavers and plastic grid paversfilled with sand and sandy top soil (Hunt et al., 2002).Clogging problems and the pollution retention capacity fordifferent metal pollutants in block paver systems with widejoints were studied (Dierkes et al., 2002). Tan et al. (2003)studied the sand-clogging potential of cylinders filled withgranite aggregates. The reduction in the permeabilityvalues due to clogging in systems of coarse aggregate canalso be analyzed by the Kozeney–Carmen equation, atheoretical empirical formula. However, this equation isusually for systems where the clogging material, if fullydistributed through the aggregate, can fill the voids toeffectively reduce the clogged system permeability to zeroand may not be very representative of the surface clogging

potential for pervious concrete pavements in the field (Tanet al., 2003).It is difficult to simulate a fully sand-clogged pervious

concrete system because it is difficult to get the sand into allthe interior pores. It is also unlikely that a system will befully clogged with sand in an actual application for thesame reason. The connecting pore system is made up ofirregular-sized voids caused by the cement matrix layersaround the aggregate in pervious concrete, and the smallerdiameters can effectively prevent sand from entering intomany of the interior voids. However, it is very common forpavements near sandy soils, such as in coastal areas, to befully covered with a layer of sand on a frequent basis due towind and drifting. It was therefore concluded that apervious concrete pavement system fully covered with alayer of sand would adequately simulate the worstconditions of sand clogging.

3. Theory development

Fig. 1 portrays a simplified view of a block of perviousconcrete. The actual pore structure is more complex withmany different pore sizes and meandering pore paths, butthis figure gives a simple overview of approximately 20% ofthe surface connecting with pores that allow water to flowthrough the system. The pervious concrete matrix aroundthese pores is in itself porous, but the water seepagethrough conventional concrete is negligible as compared tothe flow through the macropores and will therefore beassumed zero since a typical infiltration rate, or perme-ability, for a pervious concrete block is on the order of0.2–1 cm/s (Montes and Haselbach, 2006).Uncompacted sands have a much lower permeability

than pervious concrete, typically ranging from 0.001 to0.10 cm/s depending on the coarseness of the sand (Das,2002). It would therefore seem reasonable to assume thatthe flow of stormwater through a system of perviousconcrete covered with sand would be limited by the flowrate through the sand and therefore have an effectivepermeability of the sand.However, there is an additional limiting bottleneck in the

system. At the sand/pervious interface, the effectivedrainage area of the sand is reduced to that of the area

ARTICLE IN PRESS

Fig. 2. Photo of experimental flume.

Fig. 1. Block section of pervious concrete with exaggerated pore structure

representing approximately 20% porosity.

L.M. Haselbach et al. / Journal of Environmental Management 81 (2006) 42–4944

of the open pores at the surface. At the top, theeffective surface area of the pores is the sum of areas ofthe open circles on the top as depicted in Fig. 1, not theentire square surface of both concrete and pores. There-fore, even though the permeability of the sand throughthese pores may still remain the same, the surfacethat is available for infiltration has been reduced by afactor corresponding to the effective reduction insurface area from a unit block of sand to the open porearea and the total flow would be reduced by this factorassuming that there is no additional head created above thesurface. Assuming that the porosity of a block of perviousconcrete is fairly consistent throughout its height, thepercent of the surface area that is covered by pores can beapproximated by the porosity. Or, if the concrete has avertical porosity distribution, then the lowest verticalporosity can be assumed to represent the limiting percentof the open pore surfaces. If the lowest porosity is near thetop, then the effective permeability of the system would belimited by this interfacial condition and could be approxi-mated as the porosity of the concrete near the top surface(Ptop), given as a volumetric ratio, times the permeability ofthe sand.

These conditions can be represented mathematically inthe following way. The effective area available at theinterface for infiltration is the area of the pores, which is afraction of the surface area of the block or system and canbe approximated as

AP�ðP=100ÞAB. (2)

The effective permeability of the system is the limitingflow at this interface which is the ratio of areas times theunit permeability of sand

keff ¼ ðAP=ABÞksand. (3)

If Eq. (1) is substituted into Eq. (2), then the effectivepermeability of the sand-clogged system at the extreme canbe predicted as the permeability of the sand reduced by afactor representing the porosity of the pervious concrete(Eq. (1)).

3.1. Theoretical effective permeability of sample sand-

clogged pervious concrete block

Theoretically, based on Eq. (1), an effective permeabilityequation for sand-clogged pervious concrete, a perviousconcrete block with a 19% porosity near the top surfaceand which is fully covered with a sand layer that has apermeability of ksand ¼ 0:023 cm=s would have an effectivesystem permeability of

keff ¼ ð19=100Þ0:023 cm=s ¼ 0:0044 cm=s: (4)

4. Materials and methods

4.1. Flume

An adjustable wooden flume was constructed in thelaboratory. The flume was approximately 158 cmlong� 28 cm wide and had an adjustable gap in betweenfor holding the pervious concrete block. The flume can beadjusted for surface slope. Fig. 2 shows a photo of theflume set-up.

4.2. Simulation of rainfall

Rainfall was simulated on to the pervious concretesystem through a perforated hose placed on the flumelocated approximately 30 cm upslope of the pervioussystem. The rainfall rate was calculated as the volumetricflow rate of water pumped to the flume from a reservoir

ARTICLE IN PRESS

Table 1

Rainfall intensity values for Columbia, SC

Storm

frequency

(years)

Intensity of rainfall at respective time of concentration

values (cm/h)

5 10 15 20 25 30

2 13.83 12.24 10.98 9.95 9.10 8.38

5 16.31 14.37 12.84 11.60 10.58 9.72

10 18.07 15.87 14.15 12.76 11.62 10.66

25 20.74 18.13 16.10 14.48 13.16 12.06

50 22.93 19.97 17.68 15.87 14.40 13.17

100 25.10 21.78 19.23 17.22 15.60 14.26

Courtesy of SCDoT (2004).

L.M. Haselbach et al. / Journal of Environmental Management 81 (2006) 42–49 45

divided by the surface area of the pervious concrete blockin the flume. The reservoir rested on a computer-monitoredscale and the change in weight over time as measured bythe scale was used to calculate the volumetric simulatedrainfall flow rate. A portion of this simulated rainwater isinfiltrated into the system, and once the effective perme-ability of the system is surpassed, the remaining portion ofrainwater flows to the lower end of the flume as runoffwhere it can be collected into a basin resting on acomputer-monitored scale. Having the rainwater flow fromjust upstream of the pervious concrete sample helpsalleviate the boundary condition error which can be causeddue to the bouncing of the water from the perviousconcrete system into the runoff area.

Rainfall intensity values for any region are a function ofrainfall frequency. Based on available historical data,intensity/duration/frequency (IDF) curves are availablefor the geographical region from which the rainfallintensity values can be found (Chow et al., 1988; Guo etal., 2002). The applied intensities used initially corre-sponded to those that would simulate rainfall events inSouth Carolina. For illustrative purposes, the interpolatedrainfall intensity values typical in Columbia, SC for majorstorm events are shown in Table 1 (SCDoT, 2004).However, the theoretical effective permeability of thesystem is estimated at 0.0044 cm/s (16 cm/h) which isgreater than many of the largest predicted storm intensitiesfor the region. Therefore, no runoff would be expected.However, pervious concrete systems can also acceptadditional stormwater from contiguous upstream areasand in order to simulate water flows coming both fromdirect rainfall and from neighboring runoff flows, theintensities used were increased beyond the typical rainfallintensity values, but are still referred to as I (rainfallintensity in cm/h).

4.3. Measuring runoff

At the lower end of the flume, runoff was collected in acontainer that rested on a computer-monitored scale. Ascale under the runoff collection basin was used to recordthe change in weight of the water collected over time. This

recorded weight change was then used to calculate thevolumetric runoff rate based on the density of water.

4.4. Subbase

The pervious concrete system used for the simulationsconsisted of a pervious concrete block placed on self-compacted 15 cm (approx) thick sand which in turn wasplaced on a perforated wooden board wrapped with a layerof geo-synthetic filter fabric to contain the sand. This entireset-up was inside a plastic container with a drainage hole atthe bottom. The drained reservoir system beneath thepervious and sand layers had a capacity sufficient to handleall the tested permeability rates. The sand used as thesubbase was Foster Dixiana Sandblast Abrasive BX-40,considered to be an extra fine sand.The permeability and storage capacity of the system

varied with the water content of the sand subbase. Henceall the trials were performed under similar antecedentmoisture conditions. This condition was achieved bywetting the system with 6 l of water approximately 30minprior to the experimental run and allowing the system todrain during the time lapse.

4.5. Surface slope

The runoff volume and the permeability of a system canvary based on the slope of the pervious concrete surface.Experimental trials were run for calculating the experi-mental permeability of sand-clogged pervious concreteblock systems for slopes of 2% (considered fairly flat) and10% (considered steep for roads and parking areas). It isexpected that runoff would be greater in the system withthe higher slope due to horizontal vector components ofthe flow coming out of the pervious concrete.

4.6. Pervious concrete block

The Portland cement pervious concrete block used in theexperiments was sawcut from a pervious concrete slabpoured in Charleston, SC on April 1, 2004. The block wasapproximately 31 cm in length, 31 cm in width and 16 cm indepth with a surface area of 948 cm2. The slab poured inCharleston was made from a mix that consisted of 237 kgof Portland cement, 41 kg of fly ash, 1113 kg of #789granite as the coarse aggregate, 0.6 kg of WRDA17admixture (a retarder) and 72 l of water. Some additionalwater was added just prior to placing the concrete in thefield to develop a sheen on the mixture.The porosity that is assumed appropriate for the

calculations in the experiment is the porosity of the topsection of the block. The overall porosity was measuredbased on a method developed at the University of SouthCarolina (Montes et al., 2005) and represents the totalporosity of the block. This method is based on waterdisplacement in a specific gravity tank. Other experimentson blocks cut from this slab indicate that there is a

ARTICLE IN PRESSL.M. Haselbach et al. / Journal of Environmental Management 81 (2006) 42–4946

significant vertical porosity distribution within the blockand that the porosity in the top quarter can be representedby the following equation:

Ptop ¼ 1:07P� 7. (5)

The total porosity of the block was measured atapproximately 24% and the porosity in the top quarterof the block was therefore estimated to be 19% based onEq. (5) (Haselbach and Freeman, 2006).

4.7. Permeability of extra fine sand

The permeability of the extra fine sand was measuredexperimentally by the Constant Head Permeability TestMethod in accordance with ASTM D2434/AASHTOT215: Standard Test for Permeability of Granular Soilsby Constant Head Method (ASTM, 2000; AASHTO,1993). The sand used for both the subbase and the sandcovering was from the same batch as obtained from FosterDixiana Sandblast Abrasive and specified as BX-40.

4.8. Sand coverage

The permeability of this pervious concrete block systemover extra fine sand without sand clogging was previouslydetermined (Valavala et al., 2006). To simulate a varyingrange of sand clogging, dampened sand was placed over thepervious block at three different depths. The sand was alsodistributed over the flume surface for an additional 30 cmupslope of the block to minimize channeling effects at theflume/pervious interface. The three sand depths used were1.3, 2.5 and 5 cm. Fig. 3 shows a picture of the sandcovering.

Fig. 3. Photograph of run with sand covering.

4.9. Experimental effective permeability of clogged pervious

concrete block

Laboratory experiments were conducted in orderto determine the validity of Eq. (1), an effective perme-ability equation for sand-clogged pervious concrete. Inthese experiments, a pervious concrete block was coveredwith varying layers of sand and rainfall intensitieswere simulated that were great enough to produce runoff.After runoff was initiated, it was assumed that thevolume storage of the pervious concrete was saturated.Therefore, a mass balance on the steady-state systemwould dictate that the volumetric rainfall rate minusthe volumetric runoff rate would equal the infiltrationrate, or

kclog ¼ ðRainfall rate�Runoff rateÞ=Area of the block:

(6)

The mass balance was performed beginning at the first5min time interval after runoff was initiated to ensure thatthe steady-state condition was achieved, and the ratesaveraged over the remaining time. All the trials wereterminated after 30min.

5. Results and discussion

Experimental trials representing two different slopes(2% and 10%), different rainfall rates, and three differentsand coverage depths were performed on the same perviousconcrete block with a sand subbase. The experimentaleffective permeabilities (kclog) were calculated for all theruns and the results are shown in Table 2. The results areonly given when runoff occurred and were usually only forrainfall intensities that far exceeded any expected rainfallevent in South Carolina and therefore represent activerunoff flowing onto the pervious concrete from otherupslope surfaces.As can be seen in Table 2, the average kclog as determined

for surface slopes of 2% (0.0044 cm/s) is nearly identicalto the theoretically predicted value (0.0044 cm/s)for this system. There was very little variation regardlessof sand cover or rainfall intensity. A small increasewith increased sand cover could be expected to representa condition of slightly increased head above the surfaceand an increase in head gives an increase in flow. The datain Table 2 show a slight trend of a higher permeability witha deeper sand depth, but the variation is not significantwith respect to the standard deviation of the measuredvalues.The average value for the simulations representing a

10% slope was 0.0037 cm/s. It was expected that theeffective permeability would be reduced from the theore-tical prediction when the slope is increased. The effectivepermeability as predicted in Eq. (1) is based on net verticalflow into a horizontal surface. With an increase in slope,there will be a horizontal component to the flow which willdischarge horizontally and add to the runoff. Estimating a

ARTICLE IN PRESS

Table 2

Pervious concrete system permeabilities with sand cover

Slope Sand cover

depth (cm)

Approx rainfall

intensity (cm/h)akclog (cm/s)b For similar rainfall, sand

cover and slope

For similar sand cover and

slope

For same slope

Average kclog(cm/s)

Standard

deviation

Average kclog(cm/s)

Standard

deviation

Average kclog(cm/s)

Standard

deviation

2% 1.3 17 0.0041 0.0038 0.0003 0.0039 0.0004 0.0044 0.0011

17 0.0038

17 0.0035

22 0.0032 0.0040 0.0007

22 0.0044

22 0.0043

29 0.0035 0.0039 0.0003

29 0.0041

29 0.0041

2.5 22 0.0054 0.0036 0.0016 0.0042 0.0015

22 0.0029

22 0.0024

30 0.0063 0.0048 0.0013

30 0.0039

30 0.0043

5 22 0.0051 0.0053 0.0006 0.0055 0.0006

22 0.0059

22 0.0048

27 0.0052 0.0057 0.0007

27 0.0053

27 0.0065

10% 1.3 18 0.0049 0.0038 0.0010 0.0040 0.0006 0.0037 0.0008

18 0.0031

18 0.0033

21 0.0043 0.0042 0.0002

21 0.0040

21 0.0044

22 0.0046 0.0040 0.0006

22 0.0035

22 0.0038

2.5 19 0.0023 0.0026 0.0005 0.0032 0.0009

19 0.0023

19 0.0032

22 0.0026 0.0037 0.0009

22 0.0045

22 0.0043

22 0.0033

5 17 0.0034 — — 0.0038 0.0005

21 0.0043 — —

23 0.0038 — —

aA typical rainfall intensity for a 30-min, 100 year storm in Columbia, SC is less than 15 cm/h. Therefore, these rainfall intensities represent a

combination of direct rainfall and additional ‘active’ runoff from adjacent land areas.bkclog is taken as the average for each storm event approximately 5min after runoff starts.

L.M. Haselbach et al. / Journal of Environmental Management 81 (2006) 42–49 47

10% reduction in vertical infiltration would result in anestimated permeability of approximately 0.0040 cm/s andthis value compares very well to the measured value of0.0037 cm/s as the variation is less than the standarddeviation range of 20–25% as calculated for the experi-mental runs. (As can be seen in Table 2, the standarddeviation for the clogged runs was around 0.0011 cm/swhich is nearly an order of magnitude less than thedifference between the expected permeability when clogged

with sand (0.0040 cm/s) and the limiting permeability of thesand (0.023 cm/s).)Throughout the experiments, some channeling through

the sand layer on top of the pervious concrete occurred.Sometimes these channels led to a more rapid runoffbreakthrough. Channeling would be expected to have agreater effect in this laboratory flume experiment than inan actual field situation, due to scaling and the decreasedprobability of a channel being routed to another sand-

ARTICLE IN PRESS

Fig. 4. Comparison of permeabilities for pervious concrete (719% porosity near the top surface), extra fine sand-clogged systems, unclogged systems and

extra fine sand.

L.M. Haselbach et al. / Journal of Environmental Management 81 (2006) 42–4948

covered area. Hence, an increase in the variation in theexperimental kclog permeability values would be expected inthe laboratory.

From the above results, it can be concluded that thecalculated experimental permeability of the system cloggedwith sand (kclog) is consistent with the theoretical effectivepermeability (keff). This is illustrated in Fig. 4, where theexperimental system permeability values in the cloggedsystem are graphically compared to other associatedpermeabilities. The permeability of the pervious concreteblock, without a subbase or sand clogging, ranges from0.20 to 1 cm/s (Montes and Haselbach, 2006). The systempermeability over an extra fine sand subbase, without sandclogging was determined previously at approximately0.02 cm/s, similar to the permeability of extra fine sand(0.023 cm/s) (Valavala et al., 2006). The effective perme-ability of the sand-covered system was consistentlydetermined to be near the theoretically derived perme-ability of the system (0.0044 cm/s) based on Eq. (1), aneffective permeability equation for sand-clogged perviousconcrete. Additionally, as expected, runoff did not occurfor rainfall intensities less than kclog and this region ishighlighted in Fig. 4.

6. Conclusions

Runoff simulations on a Portland cement perviousconcrete block clogged with a surface layer of sand werecarried out in the laboratory. The pervious concrete systemclogged with the same sand as used in the subbase resultedin negligible runoff for both the 2% and 10% slopedsurfaces with simulations of typical rainfall intensities of upto 100 year frequencies for the Columbia, SC region.Runoff was observed when the rainfall intensities

represented a condition that might imitate the additiveeffect of both direct rainfall and additional runoff (activerunoff) from neighboring surfaces reaching the perviousconcrete surface. A relation between the permeability of thesand-clogged pervious concrete block system, the porosityof the block near the surface and the permeability of thesand was established theoretically as an effective perme-ability equation for sand-clogged pervious concrete andverified experimentally in the flume experiments for theconditions tested with high rainfall intensities. Thisequation represents a limiting permeability of the systemwhen fully covered with sand, and actual field conditionswould be expected to have higher permeabilities. The

ARTICLE IN PRESSL.M. Haselbach et al. / Journal of Environmental Management 81 (2006) 42–49 49

actual permeability range would therefore be expected tobe between the lower limit, as theoretically calculated withthe effective permeability equation for sand-cloggedpervious concrete, and the expected system permeabilityof the unclogged system. The expected system permeabilityfor the unclogged system is usually limited by thepermeability of the subbase for pervious concrete or thesoil subgrade below. Information for these extremeconditions can be useful in the design of pervious concretepavements that will also infiltrate runoff from othersurfaces or catchments. Care should also be taken in fielddesign to minimize the effects of increased runoff caused bychanneling through sand on the surface.

Acknowledgements

We would like to thank Chapman Concrete of Spartan-burg, S.C., Gordon Singletary of S & W Ready Mix,Van-Smith Concrete of Charleston, S.C., and members ofthe Carolina Ready Mixed Concrete Association(CRMCA) for their contributions to this project. Dr.Charles Pierce, Dr. Joseph Flora, Dr. Ken Harrison, Mr.Kevin Pulis and Mr. Avery Fox of the Civil andEnvironmental Engineering Department at the Universityof South Carolina generously gave their time in support ofthis work. We are also grateful for the support and fundingfor this research made available through the Center forManufacturing and Technology at the University of SouthCarolina.

References

American Association of State Highway and Transportation Officials

(AASHTO), 1993. AASHTO Guide for the Design of Pavement

Structures. Washington, DC.

American Society for Testing and Materials (ASTM), 2000. ASTM D2434

Standard Test Method for Permeability of Granular Soils (Constant

Head). West Conshohocken, PA.

Barrett, M.E., Irish, L.B., Malina, J.F., Charbeneau, R.J., 1998.

Characterization of Highway Runoff in Austin, Texas, Area. Journal

of Environmental Engineering 124(2), 131–139.

Chow, V.T., Maidment, D.R., Mays, L.W., 1988. Applied Hydrology.

McGraw-Hill Series in Water Resources and Environmental Engineer-

ing, New York.

Das, B.M., 2002. Principles of Geotechnical Engineering, fifth ed. Brooks/

Cole Publishers, Pacific Grove, CA.

Dierkes, C., Kuhlmann, L., Kandasamy, J., Angelis, G., 2002. Pollution

retention capability and maintenance of permeable pavements. In:

Proceedings of the Global Solutions for Urban Drainage, Ninth

International Conference on Urban Drainage, American Society of

Civil Engineers, September 8–13, Portland, OR.

Guo, C.Y., James, Urbonas, B., 2002. Runoff capture and delivery curves

for storm-water quality control designs. Journal of Water Resources

Planning and Management 128, 208–215.

Haselbach, L.M., Freeman, R., 2006. Vertical porosity distributions in

pervious concrete pavement, American Concrete Institute (ACI)

Journal of Materials, submitted.

Hunt, B., Stevens, S., Mayes, D., 2002. Permeable pavement use and

research at two sites in eastern north Carolina. In: Proceedings of the

Ninth International Conference on Urban Drainage, American Society

of Civil Engineers, September 8–13, Portland, OR.

Montes, F., Haselbach, L., 2006. Hydraulic conductivity of pervious

concrete. Environmental Engineering Science, accepted for publica-

tion.

Montes, F., Valavala, S., Haselbach, L., 2005. A new test method for

porosity measurements of portland cement pervious concrete. Journal

of ASTM International 2 (1).

South Carolina Department of Transportation (SCDoT), 2004. Rainfall

Intensity values utilized by South Carolina Department of Trans-

portation. http://www.dot.state.sc.us/doing/pdfs/rainfall_intensity_

chart.pdf (08/04/04), South Carolina Department of Transportation,

USA.

Tan, S., Fwa, T., Han, C., 2003. Clogging evaluation of permeable bases.

Journal of Transportation Engineering 129 (3).

Tennis, P.D., Leming, M.L., Akers, D.J., 2004. Pervious Concrete

Pavements. Portland Cement Association (PCA), Skokie, IL.

Valavala, S., Montes, F., Haselbach, L., 2006. Area rated rational

coefficients values for portland cement pervious concrete pavement.

American Society of Civil Engineers (ASCE) Journal of Hydrologic

Engineering 11(3).

Zouaghi, A., Kumagai, M., Nakazawa, T., 2000. Fundamental study on

some properties of pervious concrete and its applicability to control

stormwater run-off. Transactions of the Japan Concrete Institute 22,

43–50.