Time-Dependent Drainage Capacity and Runoff of

Pervious Block Subjected to Repeated Rainfall Simulation

Dawa Seo 1; Tae Sup Yun 2; Kwang Yeom Kim 3; and Kwang Soo Youm 4
Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

Abstract: To mitigate flood damage in cities, pervious concrete has been developed as a viable and sustainable alternative to traditional
concrete to facilitate drainage. Previous studies have tended to evaluate the drainage capacity of pervious blocks through permeability and
drainage tests in simplified conditions, giving little consideration to multiple environmental factors such as rainfall rate and temporal changes
in the blocks’ drainage capacity. This study presents experimental results of the runoff and drainage capacity of pervious blocks subjected to
time-dependent evaporation and corresponding changes in their degree of saturation. Different levels of repeated water charging at designated
time intervals simulated the urban environment, and both runoff and drainage were continuously monitored. The results highlight that runoff
can take place after certain time intervals despite the same water charge because of evaporation and prewetting-induced changes in water-
retention capacity. The effects of the surface layer, the bedding layer, and clogging (all part of the urban areas) on the drainage were
also observed. The findings underscore the significance of the actual, rather than the simplified laboratory-based, drainage capacity in urban
areas. DOI: 10.1061/(ASCE)MT.1943-5533.0001800. © 2016 American Society of Civil Engineers.
Author keywords: Pervious concrete; Drainage test; Repeated rainfall; Evaporation; Drainage capacity.

Introduction related to the heat island effect and reduced traffic noise; (5) mini-
mized splashing on roads; and (6) less shrinkage than conventional
The cost of flood damage in urban areas caused by localized heavy concrete (Ferguson 2005; Neithalath 2004; Wanielista and Chopra
rain has escalated in recent decades. The annual loss arising from 2007; Yang and Jiang 2003). However, the porous structure of per-
inundation is estimated to be US$3.2 billion in the United States vious concrete is vulnerable to reduced drainage capacity caused
and US$6.8 billion in Europe (Changnon 2003; Jongman et al. by clogging, and it has low mechanical strength and is susceptible
2014). In Korea, the strongest typhoons in 2002 and 2003, Rusa to freeze-thaw effects (Kevern et al. 2008; Tennis et al. 2004;
and Maemi, resulted in damage repair costs of US$5.03 million Wanielista and Chopra 2007). Nevertheless, such concrete has been
and US$4.03 million, respectively (Choi and Moon 2008). More widely used on roads, parking lots, sidewalks, and routes with low
than 75% of the damage reported in urban areas affects public fa- traffic loadings (Tennis et al. 2004; Tong 2011). The common fea-
cilities, and this accounts for 70% of the total damage repair costs tures of pervious concrete include a void ratio of 15–30%, a flow
(European Commission 2013; Choi and Moon 2008). To minimize rate of 0.4–19.0 mm=s, and a strength of 3.5–28 MPa (Table 1). In
the damage resulting from flooding and to establish a systematic situ drainage capacity and void ratio can be measured by using the
drainage system in urban areas, sewerage and reinforcement struc- standard test methods ASTM C1701 (ASTM 2009) and ASTM
tures (e.g., containing gates and detention tanks) have been devel- C1688 (ASTM 2014), respectively, although results may vary.
oped, and the high efficiency of such a system has been widely X-ray computed tomography can be used to assess the quantitative
reported (Cembrano et al. 2004; Hsu et al. 2000; Schmitt et al. characteristics of pore space in pervious concrete, including effec-
2004). Pervious concrete has recently received attention as an al- tive porosity, pore-size distribution, and tortuosity (Manahiloh et al.
ternative and effective material to mitigate flooding in urban areas, 2012; Sansalone et al. 2008). Previous studies have investigated the
where impervious surfaces cover between 65 and 70% of the land infiltration capacity of pervious concrete on the basis of tests of
surface (Lee et al. 2012). Such concrete has the following advan- permeability and drainage by performing rain simulations and
tages: (1) outstanding drainage ability through highly connected slope modeling (Andersen et al. 1999; James and Von Langsdorff
pores; (2) groundwater recharging; (3) better urban environment 2003; Lee et al. 2012). However, in situ pervious concrete is sub-
for planting and vegetation; (4) reduced energy consumption jected to high levels of flow charge that accumulates from imper-
vious regions, natural evaporation to the atmosphere, and repeated
1
Graduate Student, Dept. of Civil and Environmental Engineering, rain events; consequently, the various boundary and initial condi-
Yonsei Univ., Yonsei-ro 50, Seodaemun-gu, Seoul 120-749, Korea. tions of pervious concrete need to be considered for hydraulic
2 characterization.
Associate Professor, Dept. of Civil and Environmental Engineering,
Yonsei Univ., Yonsei-ro 50, Seodaemun-gu, Seoul 120-749, Korea This study aims to assess the drainage capacity of pervious
(corresponding author). E-mail: taesup@yonsei.ac.kr blocks by considering the effects of evaporation, retained water
3
Research Fellow, Korea Institute of Civil Engineering and Building in the block, and repeated rainfall conditions. Three types of speci-
Technology, 283 Goyangdae-ro, Ilsanseo-gu, Goyang 411-712, Korea. men were prepared: a porous concrete block, a block with a less
4
Senior Manager, Infra-Structure Team, Technical Division, GS E&C,
porous surface layer, and a porous concrete block over bedding
33 Jong-ro, Jongno-gu, Seoul 110-130, Korea.
Note. This manuscript was submitted on February 11, 2016; approved
sand. The flow charge conditions were simulated by using desig-
on August 29, 2016; published online on November 16, 2016. Discussion nated time intervals, and runoff and drainage behaviors were simul-
period open until April 16, 2017; separate discussions must be submitted taneously monitored throughout each test. Additionally, the
for individual papers. This paper is part of the Journal of Materials in Civil influence of clogging, a critical drainage-controlling factor, was as-
Engineering, © ASCE, ISSN 0899-1561. sessed by conducting repeated drainage tests at set time intervals.

© ASCE 04016273-1 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

 Table 1. Strength, Permeability, and Porosity in Previous Studies Table 3. Measured Properties for Each Tested Specimen
Strength Permeability Porosity Specimens
References (MPa) (mm=s) (%)
Properties P10 SP10 P10 þ BL
Tennis et al. (2004) 3.5–28 2.0–12.0 15–30
Strength O — —
Kevern et al. (2008) 9.9–20.6 0.1–11.8 15–32.1
Young’s modulus, E O O —
Ferguson (2005) 6.7–17.5 0.3–6 14–31
Permeability O O O
Yang and Jiang (2003) 7.1–13.8 7.8–8.9 —
3D X-ray CT O O —
Huang et al. (2010) 5–6.5 17.0–19.0 26–30
Evaporation O O O
Montes and Haselbach (2006) — 1.32–4.04 16–25
Drainage-runoff O O O
— 4.6–13.2 22–32
Clogging O — —
Note: CT = computed tomography; O = test specimen.
Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

Table 2. Design Mixture of Pervious Concrete Strength

Composition Proportions The compressive and bending strength values were measured
Water-cement ratio (%) 23.3 according to Korean standard tests KS F 2405 (KS 2010) and
Cement (g) 330 KS F 2408 (KS 2000). The cored specimen (diameter: 100 mm;
Water (g) 77 height: 60 mm) was subjected to compression (3500KN, SATEC,
Coarse aggregate (g) 300 Instron, Norwood, Massachusets). A slab specimen (198-mm long,
Fine aggregate (g) 1,200 100-mm wide, and 60-mm high) was used for bending strength
Admixture (g) 4.95 measurements (600KN, SATEC, Instron, Norwood, Massachu-
sets). Strength values were measured for three specimens, and the
average values are reported.
Materials and Methods
Permeability
The cored specimen was placed in an acrylic cylindrical cell
Materials (90-mm internal diameter; 5-mm thick), and a top cell was posi-
Coarse aggregate (i.e., No. 8 size; 9.252-mm diameter) is com- tioned on top of the specimen (i.e., the specimen was sandwiched
monly used in pervious concrete pavement results in enhanced between the top and bottom cells); the components were connected
porosity at the expense of reduced strength. However, a well- by heat-shrink tubing and clamps (Fig. 1). This arrangement al-
graded aggregate should be avoided, as it reduces porosity lowed the vertical flow simulation to be performed without leaking
(Abdel-Jawad and Abdullah 2002). Three block specimens were along the specimen sides. Before testing the permeability, the speci-
prepared for analysis in this study: one composed of aggregate men was fully saturated in a vacuum chamber for 24 h to remove
comprising particles of less than 10 mm in diameter and a block air from pore spaces. A constant pressure head of 136 mm was
height of 60 mm (P10); another with the same aggregate size and a maintained, and the volume of outflow was measured for 30 s. The
height of 50 mm, covered with a 10-mm-thick layer of aggregate permeability was then computed as follows:
(particles with a diameter of 1–2 mm) (SP10); and a third including
QL
a bedding layer (P10 þ BL). Such a surface layer for SP10 is com- k¼ ð1Þ
monly used in commercial pervious blocks to ensure pedestrian Aht
safety and for cosmetic reasons. where Q = volume of outflow measured over t seconds; L =
Type I ordinary portland cement (OPC or CEM I 42.5N) with a specimen height (e.g., 60 mm); A = cross-sectional area of the
Blaine surface area of 3,315 cm2 =g was made by a professional side- specimen; and h = constant pressure head.
walk manufacturer. The mix proportion of pervious blocks is listed in
Table 2. An additive based on naphtalene was added in small X-Ray Computed Tomography
amounts for medium-range water reducing. The manufactured spec- The porosity of the pervious blocks can be measured by calculating
imens satisfied the maximum aggregate particle size of 13 mm stated the differential weight between dry and fully saturated conditions.
in the Korean standard KS F 4419 (KS 2014). The specimens were As water can exist within pore spaces and be absorbed in the ma-
made by using the compression method in the same way as commer- trix, which may not contribute to flow capacity, three-dimensional
cial production and were 198 mm in both width and length. They (3D) X-ray computed tomographic imaging was performed to
were cured at 20°C and 70% humidity for 2 days, and recured in quantitatively obtain the pore volume with a resolution of
an open-air storage yard for 7 days. After curing, the specimens were 0.135 mm per pixel. This also allowed an assessment of both con-
cored (100-mm diameter) for the flow experiments. The Jumunjin nected and isolated pores. The two-dimensional (2D) slice images
sand (grain size: 2.36–4.75 mm; specific gravity: 2.31) was used were sequentially stacked and binarized to reconstruct the 3D
for the bedding layer in this study, as designated by Seoul Metro- volume of pore space, which served as a calibration factor in the
politan Facilities Management Corporation (2013), which requires evaporation and drainage tests.
a maximum grain size of 5 mm and bedding layer height of
30 mm. The United States and United Kingdom regulations require Evaporation
a bedding layer height of 40–50 mm, whereas Japanese regulations The degree of saturation of a pervious concrete block can be used to
require a height of 30 mm (Interpave 2010; Smith 2000). calculate its water-retention capacity and the changed amount of
water in pores after rainfall attributed to gravity-driven drainage
and evaporation to the atmosphere. As the surface of the in-place
Methods
pervious block is mainly exposed to the air, the water held in pores
Table 3 lists the properties measured for each specimen. The test would be able to evaporate. Therefore, evaporation must be con-
procedures are summarized as follows. sidered to analyze the impact of water redistribution on the drainage

© ASCE 04016273-2 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 1. Flow cell for measuring permeability and drainage rate for pervious concrete block specimen

capacity of the block. The sides of cored specimens were sealed in In the intervening intervals, evaporation from the specimen was al-
the present study by using a heat-shrink tube. Each specimen was lowed to evaluate the effect of the degree of saturation on drainage.
saturated in a closed chamber by vacuum pump for 24 h, and then The weight of drained water at the base and the accumulated runoff
natural evaporation was allowed at room temperature (23°C) and height over the specimen were monitored every second. The initial
humidity (54%) for 400 h. During the evaporation period, the speci- degree of saturation for the inflow tests at each elapsed time point
men was placed on a balance and weighed every minute. was calibrated by using results from the evaporation test.

Drainage and Runoff Clogging

Pervious blocks can be subjected to high volumes of water inflow Clogging causes the deposition of granules in pore space, which
from impervious regions over short periods of time. The inflow reduces porosity and permeability (Deo et al. 2010; Tong 2011).
rates determined in this study were based on the average of five Clogging tests were usually repeatedly performed by placing sedi-
1-h-maximum rainfall values in Seoul, Korea (Korea Meteorologi- ment on the specimen before the permeability test or by mixing
cal Administration 2014), which is equivalent to 78.28 mm=h for sediment with water, followed by pouring (Mata 2008). The sedi-
10 min or 0.06 L=min for the surface area of the specimen. Four in- ment used for the clogging test in the present study comprised 60%
flow rates were used: 0.06, 0.13, 0.25, and 0.5 L=min. The specimen of coarse grains that were retained between #8 and #200 sieves, and
was initially oven dried at 40°C for 48 h, and the designated inflow rate 40% of fine grains that passed through a #200 sieve. Ten grams of
was evenly spread over the specimen from a height of 200 mm by clogging material was spread evenly on the specimen, calculated
using the same experimental setup as that used for permeability. from a previous study by comparing the volume of pores (Deo et al.
As described in Fig. 2, one set for a specific inflow rate con- 2010). The inflow rate was 0.13 L=min during the clogging experi-
sisted of four drainage tests and three time intervals (5 min, 1 h, ment. The clogging test was conducted in the same way as the
and 12 h). The dry specimen was subjected to the designated inflow drainage test (Fig. 2) to explore clogging-induced changes in the
drainage capacity of the substratum, which largely comprised
rate for 600 s (10 min), and the supply of water was ceased for 300 s
pervious concrete with large pores.
(5 min) to let the specimen drain and evaporate. This was the first
drainage test and the first time interval. Next, water with the same
inflow rate as the first one was supplied for 600 s, and then the
second time interval (1 h) was applied to allow water in the speci-
Results and Discussion
men to evaporate. Subsequently, after 1 h and again after 12 h,
similar drainage repetitions with the same rate were carried out. Code Adherence of the Strength and Permeability of
the Pervious Concrete Block
The compressive and bending strength values for the P10 speci-
men were 28.8 and 5.8 MPa, respectively, thereby complying with
the Korean standards for pervious block (Seoul Metropolitan
Facilities Management Corporation 2013), as presented in Fig. 3
(compression strength: 16 MPa; bending strength: 4 MPa). The
Young’s modulus was 20.66 GPa. The high porosity and low
binding capacity of the pervious block resulted in low strength
values and failure along the interface between aggregate and the
Fig. 2. One set of drainage test with time intervals
cement matrix.

© ASCE 04016273-3 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 3. Compressive strength, bending strength, and Young’s modulus

for P10 specimen; the lines designate the code-specified requirement
for strength values

Table 4. Measured and Calculated Values of Permeability

Bedding layer P10 þ BL
Permeability P10 (mm=s) SP10 (mm=s) (mm=s) (mm=s)
Measured 0.831 0.796 6.253 1.907
Calculated — — — 1.814

The permeability of pervious blocks should be greater than Fig. 4. Pore-size distribution curve of specimen: (a) P10; (b) SP10
0.1 mm=s, as designated by KS F 4419 (KS 2014); the values mea-
sured in the present study are listed in Table 4. The cover layer on
the SP10 specimen was unlikely to reduce permeability, given its
low thickness; instead, the permeability was controlled mainly by maximum diameter of 4 mm, and the mean pore diameter was
the substratum. The bedding layer of Jumunjin sand had a per- 1.10 mm (at the 50% cumulative percentage of pore size).
meability of 6.253 mm=s, which was 7.5 times higher than that Fig. 4(b) presents the pore-size distribution of SP10; the mean
of the P10 specimen; this high permeability resulted from its high diameter was 0.69 mm (0.46 mm in the surface layer, 0.97 mm
porosity of 0.4. The permeability of the P10 specimen with a bed- in the substratum of SP10). The pore size of the bedding layer var-
ding layer was 1.907 mm=s. As the pervious block and the bedding ied from 0.363 to 0.983 mm because the diameter of Jumunjin sand
layer were vertically layered to the flow direction, the validity of the varied from 2.36 to 4.75 mm. The range in pore size covers D=6.4
measured permeability can be assessed as follows: to D=2.4 of the mean particle size of D (Santamarina et al. 2001).
According to Benavente et al. (2015), the water flow through pores
H with diameters of 2–3 mm is governed by gravity rather than capil-
kvðeqÞ ¼     ð2Þ
H1 H2 lary force, whereas capillary force plays a role in pores smaller than
kv1 þ kv2
2 mm. Therefore, the hydraulic behavior of the specimen would be
where H = total specimen height (160 mm; specimen: 60 mm; bed- influenced by either capillary force or gravity, depending on the
ding layer: 100 mm); and subscripts 1 and 2 ¼ P10 and Jumunjin condition of the pores.
sand, respectively. The equivalent permeability was estimated to be
Evaporation
1.814 mm=s, which was within 10% error of the measured per-
The evaporation rates for the P10, SP10, and P10 þ BL specimens
meability (1.907 mm=s), thereby indicating that the measured per-
are shown in Fig. 5. Bilinear behavior and inflection points clearly
meability values were reasonable. The results indicate that the
appeared in the cases of P10 and SP10. In the first stage of bilinear
permeability of the bedding layer is a critical factor, provided that
evolution, the degree of saturation sharply decreased as connected
the pervious block satisfies the code.
water in the pores evaporated to the atmosphere by capillary-driven
flow. When hydraulically connected water in pores broke, the sec-
Drainage Capacity with Consideration of Evaporation ond stage started, and evaporation rate abruptly decreased by vapor
diffusion of not only the remained water in pore spaces but also the
Spatial Configuration of Pore Space absorbed water in paste whose pore size is much finer (<2 nm) than
The bulk porosity of P10 was estimated to be 0.263, whereas the the bulk pore (Lehmann et al. 2008; Shokri et al. 2010; Korpar and
porosity of the connected pores was 0.239 (90.87%), as determined Trettin 2006). Furthermore, as hydration, which is a chemical re-
by analysis of X-ray computed tomographic images. Effective action of cement paste and needs water, progressed slowly even in
porosity was the sole contributor to fluid flow and was used in this mature cement paste, the nominal amount of water would be con-
study in calculating the degree of saturation and in other relevant tinuously consumed in wet condition of cement paste, and this
analyses. The skeletonization of a binary image set together with could delay the evaporation rate simultaneously (Thomas and
the Euclidean distance (from the solid boundary, to extract skeletal Jennings 2006).
structures) corresponded to the pore-size distribution, as shown in As the pore size of the material on the upper surface of the SP10
Fig. 4(a) (Dong and Blunt 2009). Pore size ranged widely up to a specimen was smaller than that of the P10 specimen, the SP10

© ASCE 04016273-4 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

 consequently lead to drainage. Accordingly, the runoff diminished
[Figs. 6(a and b)].
The drainage rate of water was calculated by dividing the moni-
tored accumulated drainage volume by the elapsed time (Fig. 6).
The drainage rate was linked to the runoff behavior to satisfy the
mass conservation. A sharp increase in drainage rate followed by
a quasi-constant value was observed when no runoff occurred
[e.g., the 0- and 5-min cases in Figs. 6(a and b)]. The termination
of inflow made the drainage rate instantaneously drop to zero. Once
runoff occurred, the drainage rate increased and tended to remain
mainly constant after the runoff height reached a maximum value
Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

[12-h case in Figs. 6(a and b)]. For the specimen with poor drain-
age, both runoff height and drainage rate increased until inflow
ceased [1- and 12-h experiments in Fig. 6(c)]. As explained previ-
ously, the air trapped during the period between experiments gen-
Fig. 5. Evaporation of tested specimens (P10, SP10, P10 þ BL) with erated runoff and interrupted drainage. This decreased the initial
time drainage rate [C in Fig. 6(c)]. As increased runoff led the air to be
highly pressurized, the air overcame the surrounding water pressure
and capillary force. Consequentially, air penetrated into the water
through the pores; thus, the reduced drainage rate in the early stage
specimen had less area exposed to the atmosphere and had a cor-
[C in Fig. 6(c)] of the test showed a recovery to the rate in the no-
respondingly lower evaporation rate; also, water in smaller pores
runoff period [D in Fig. 6(c)]. In other words, the drainage rate
had higher capillary forces and was more difficult to evaporate be-
stabilized.
cause the capillary force is inversely proportional to the pore radius
The evolution of the degree of saturation with time and the
at the first evaporation stage. Both specimens showed a similar
initial saturation for each step were calibrated and plotted by con-
transition time, after which the evaporation rates plateaued and
sidering the evaporation changes in Fig. 6. The changes in the
declined similarly. In the underlying coarse layer, air was able
degree of saturation can be divided into four sections: a sharp in-
to invade the upper part of the coarse layer and reduce the water
crease on the start of inflow; a slow growth and a constant evolu-
content, whereas the fine layer showed a tendency to remain satu-
tion; a rapid decrease after stopping inflow; and a stable period
rated throughout the test (Shokri and Or 2011). Hence, the decrease
(e.g., P10 dramatically increases to 34.12% after inflow starts, in-
in evaporation rate was slowest in the P10 specimen with a bedding
creases continuously, and then drops from 57.18 to 40.49%, main-
layer of sand, which retained more water than the pervious blocks.
taining a final value at 40.49%). The degree of saturation tended to
Runoff Behavior rise continuously in the P10 specimen [Fig. 6(a)], whereas SP10
Fig. 6 shows the entire evolution of runoff height and drainage rate and P10 þ BL maintained a degree of saturation within a specific
at 0.13 L=min for the three time intervals (5 min, 1 h, and 12 h) and range at slightly higher values than for P10 [Figs. 6(b and c)]. By
the calculated degree of saturation for the P10, SP10, and P10 þ using the entire evolution of runoff and drainage behaviors, the
BL specimens. For the P10 specimen, no runoff was observed at the quantitative values of maximum runoff height and drainage capac-
start or at 5 min. Runoff began at 1 h of elapsed time and reached ity can be obtained and assessed.
the maximum runoff height (5 mm) following by a gradual decrease Fig. 7 shows the maximum runoff height for the entire drainage
before the inflow ended. In other words, evaporation for 1 h in- tests for P10, SP10, and P10 þ BL at flow rates of 0.13, 0.25, and
duced a change in the pore conditions, and runoff occurred despite 0.50 L=min. There was no runoff observed at 0.06 L=min, whereas
the use of the same inflow rate as at the start and at 5 min. In the runoff showed variable behavior with elapsed time in the case of
addition, drainage to the specimen base was more effective than a flow rate of 0.13 L=min [Fig. 7(a)]. In other words, the occur-
runoff along the top. The SP10 specimen exhibited a similar runoff rence of runoff was dependent not solely on material properties
evolution (i.e., runoff after 1 h, maximum runoff height 6 mm) as such as porosity and pore structure but also on temporal changes
the P10 specimen, even though it had a slightly greater height. in the degree of saturation and the unobservable distribution of
Unlike the P10 and SP10 specimens, the P10 þ BL specimen dis- water in pore spaces. The experiments showed that the highest run-
played an earlier runoff occurrence, and the runoff height of this off heights were mainly observed in the following order: P10 with a
specimen increased over time because of the bedding layer. After bedding layer, SP10, and P10, and runoff height tended to increase
the imposed inflow was terminated, the runoff showed a linear with increasing inflow rate and over time. On average, P10 with a
decrease [B in Fig. 6(c)]. bedding layer and SP10 presented 1.34 and 2.41 times maximum
Gravity-driven drainage created unsaturated conditions within runoff height of P10, respectively. The SP10 specimen showed
pore spaces. Evaporation during the period between successive a gradual increase in runoff height, which was not seen for the
inflow tests resulted in menisci forming within pores and a rise in P10 and P10 þ BL, and occurred regardless of inflow rate. As
capillary pressure (Nishiyama et al. 2012). In addition, air entered described previously, the SP10 specimen exhibited delayed evapo-
the pores where water left during evaporation. When inflow re- ration, and consistent capillary forces were readily generated within
started after the period between successive inflow tests, the flowing the small pores. That is, the surface layer composed of smaller ag-
water approached that already held in the pores. Then, the air was gregates than the substratum and the bedding layer caused runoff
trapped between the liquids, and a thin layer of air appeared that can only reoccur after a considerable period.
(Mehdi-Nejad et al. 2003). This process acted as a barrier against The drainage capacity was defined as the ratio of the initial
inflow into pores (Weizu and Freer 1995) and restricted the down- drainage rate (C in Fig. 6) to its quasi-constant value in the case
ward movement of water, generating runoff. To overcome the en- of no runoff (D in Fig. 6). Fig. 8 indicates that the drainage capacity
trapped air pressure, an external force such as elevated pressure depended on the inflow rate and the elapsed time, for all specimens
head over the specimen was required (Perkins 1957) and would (P10, SP10, and P10 þ BL). For an inflow rate of 0.06 L=min, the

© ASCE 04016273-5 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. Evolution of runoff and drainage rate during four repeated drainage cycles at t ¼ 0 min, 5 min, 1 h, and 12 h for tested specimens with
0.13 L=min: (a) P10; (b) SP10; (c) P10 þ BL

© ASCE 04016273-6 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Maximum runoff height in each time interval (start, 5 min, 1 h, Fig. 8. Drainage capacity in each flow rate (0.06, 0.13, 0.25, and
and 12 h) of the specimens (P10, SP10, P10 þ BL): (a) 0.13 L=min; 0.50 L=min) with time intervals (start, 5 min, 1 h, and 12 h):
(b) 0.25 L=min; (c) 0.50 L=min (a) P10; (b) SP10; (c) P10 þ BL

inflow water was fully drained without runoff. The drainage capac-
ity tended to decrease as time passed, as the inflow rate increased; of water within pore spaces, which in turn determines the drainage
this was observed for all specimens. The minimum drainage capac- capacity of the pervious blocks.
ity was reached by P10 and P10 þ BL at 5 min and 1 h, respec-
tively. The drainage capacity of SP10 showed a continuous decline Correlations between Measured Quantities during the
and exhibited the lowest drainage capacity throughout the 12-h Drainage Tests
experiment. This trend in drainage capacity was similar to that The measured changes in the degree of saturation were estimated to
in runoff evolution. The maximum runoff height among all the be derived from the difference between the maximum saturation
specimens was exhibited by P10 þ BL, which also showed the larg- before inflow stopped and the stable state value [A in Fig. 6(a)]
est difference between the minimum and maximum drainage rate and were correlated with the maximum saturation for each step
(64.96%). This result indicates that the drainage capacity of the in Fig. 9(a). As the inflow rate increased, both the maximum
bedding layer is strongly affected by evaporation. Evaporation saturation and the difference with the saturation in the stable state
causes a reduction in saturation and changes the spatial distribution increased. The drainage of inflow water is controlled mainly by

© ASCE 04016273-7 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

 P10 þ BL, with smaller pores, showed measured properties that
followed a similar tendency.
Fig. 9(b) illustrates the reduction rate of runoff height alongside
the maximum runoff height. At a given inflow rate, the specimen
had a similar reduction rate of runoff height even where the
maximum runoff height varied (e.g., the average reduction rate
in runoff height was 0.242 mm=s at 0.13 L=min, 0.509 mm=s at
0.25 L=min, and 0.935 mm=s at 0.50 L=min for measurements
over the entire specimen). The results can be approximately divided
into three groups according to flow rate (0.13, 0.25, and
0.50 L=min) and can be described as a series of shadowed regions
(e.g., the box in the 0.13-L=min region is involved in 0.25 L=min).
Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

The average reduction rate of runoff was increased by 1.97 times in

proportion to inflow rates. However, the range of maximum runoff
height became wider as the inflow rate increased. Even though the
overall reduction rate of runoff height was increasingly propor-
tional to maximum runoff height, the P10 specimen (whose pore
size and volume are largest) exhibited the highest rate of runoff
height reduction, followed by SP10 and P10 þ BL. For an inflow
rate of 0.13 L=min, the difference between minimum and maximum
runoff height was 18.55 mm (i.e., the minimum runoff height was
5.45 mm for P10, and the maximum height was 24 mm for P10 with
a bedding layer), whereas an inflow rate of 0.50 L=min yielded a
difference of 77 mm (the maximum runoff height was 118 mm for
P10 with a bedding layer, and the minimum was 41 mm for P10). If
the reduction rate of runoff height was controlled by the hydrostatic
pressure, which was quasi-equivalent to the runoff height in this
study, it would linearly increase with runoff height. Yet, the reduc-
Fig. 9. Comparison between test results of P10, SP10, and P10 þ BL: tion rate was grouped by inflow rate, and it varied widely with the
(a) maximum degree of saturation and change in degree of saturation type of specimen. This result indicates that hydrostatic pressure has
after inflow stopped; (b) maximum runoff height and reduction rate of a relatively minor effect on drainage rate and that the inflow rate
runoff height controls the reduction rate of runoff height. The existence of a sur-
face layer and a bedding layer affects the runoff height but has an
insignificant effect on the reduction rate of runoff volume.
gravity, and the water-retention capacity (e.g., the degree of
saturation of specimens) is determined by pore volume, pore size,
and microstructural networks (Beven and Germann 1982). There- Effect of Clogging
fore, P10, the specimen with the largest mean pore size in this Fig. 10(a) shows the effect of clogging on runoff height. The
study, showed the highest values of saturation change; SP10 and P10 specimen showed time-dependent runoff. The tests in which

Fig. 10. Clogging test results of P10 specimen: (a) maximum runoff height with time intervals for P10 with and without clogging; (b) weight of
flushed soil with elapsed time; (c) drainage rate of tested specimens in clogging tests

© ASCE 04016273-8 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

 clogging occurred showed noticeable runoff from the beginning, Research Foundation of Korea (NRF) grant funded by the Korean
whereas the maximum height of runoff was observed at 1 h Government (MSIP) (Nos. 2012-0008929, 2011-0030040,
(e.g., 42.54 mm compared with 5 mm for the P10 specimen). 20133030000240).
Clogging by soil either reduced the pore size or blocked the pore
pathway such that the drainage capacity decreased with increased
runoff height. References
The weight of soil flushed during drainage showed a sharp de-
crease over time [i.e., at the start time and after 5 min, 1 h, 12 h, Abdel-Jawad, Y. A., and Abdullah, W. S. (2002). “Design of maximum
and 2 weeks; Fig. 10(b)], and the collected soils were very fine in density aggregate grading.” Constr. Build. Mater., 16(8), 495–508.
size. This result indicates that large particles blocked the pore space Andersen, C. T., Foster, I. D. L., and Pratt, C. J. (1999). “The role of urban
in the pervious concrete, thereby reducing the drainage capacity surfaces (permeable pavements) in regulating drainage and evaporation:
Development of a laboratory simulation experiment.” Hydrol. Process.,
and allowing only fine particles to be flushed. The clogging soils
Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

13(4), 597–609.
were precipitated and dried in the pore space, and subsequent drain- ASTM. (2009). “Standard test method for infiltration rate of in place
age flushed the fine particles, whose total weight per drainage cycle pervious concrete.” ASTM C1701, West Conshohocken, PA.
decreased over time [Fig. 10(b)]. ASTM. (2014). “Standard test method for density and void content of
There was no significant difference in the drainage rate caused freshly mixed pervious concrete.” ASTM C1688, West Conshohocken,
by clogging in the first drainage test [Fig. 10(c)], when soil particles PA.
could float in and flow with the water through the pore network, Benavente, D., et al. (2015). “Predicting water permeability in sedimentary
with some becoming clogged in pores. In the second drainage test rocks from capillary imbibition and pore structure.” Eng. Geol.,
(e.g., at 5 min of elapsed time), the drainage rate at the beginning of 195(2015), 301–311.
the drainage test was low and tended to converge toward the rate Beven, K., and Germann, P. (1982). “Macropores and water flow in soils.”
Water Resour. Res., 18(5), 1311–1325.
measured in the nonclogging case. In addition, when the inflow
Cembrano, G., Quevedo, J., Salamero, M., Puig, V., Figueras, J., and Marti,
stopped, the drainage rate decreased gradually because of runoff J. (2004). “Optimal control of urban drainage systems. A case study.”
water [Fig. 10(a)]. Similar results were obtained for the subsequent Control Eng. Pract., 12(1), 1–9.
drainage tests at 1 and 12 h, even though the weight of fine particles Changnon, S. A. (2003). “Shifting economic impacts from weather ex-
being flushed was decreasing. This result indicates that large tremes in the United States: A result of societal changes, not global
particles blocked the pore spaces and inhibited the drainage rate, warming.” Nat. Hazards, 29(2), 273–290.
whereas fine particles made a smaller contribution to the clogging- Choi, E. S., and Moon, I. J. (2008). “The variation of extreme values in the
induced reduction in drainage rate. In other words, the clogging precipitation and wind speed during 56 years in Korea.” Atmos. Korean.
reduced the drainage rate at both the beginning and end of the test. Meteor. Soc., 18(4), 397–416.
This occurred because large soil particles would block the pore Deo, O., Sumannasooriya, M., and Neithalath, N. (2010). “Permeability
reduction in pervious concretes due to clogging: Experiments and mod-
space and effectively control the drainage rate unless the clogged
eling.” J. Mater. Civ. Eng., 10.1061/(ASCE)MT.1943-5533.0000079,
particles were cleared. This observation leads to the inference that 741–751.
the flushed fine particles may work as a barrier and reduce the en- Dong, H., and Blunt, M. J. (2009). “Pore-network extraction from micro-
tire drainage capacity and runoff by blocking pores when pervious computerized-tomography images.” Phys. Rev. E., 80(3), 036307.
concrete is set on a roadbed. European Commission. (2013). “Community research and development in-
formation service, FLOODPROBE (technologies for the cost-effective
flood protection of the built environment.” 〈http://cordis.europa.eu
Conclusions /result/rcn/153483_en.html〉 (Jan. 18, 2013).
Ferguson, B. K. (2005). Porous pavements, Taylor & Francis, New York.
This study presents a series of experimental results focusing on Hsu, M. H., Chen, S. H., and Chang, T. J. (2000). “Inundation simulation
time-dependent drainage and runoff behaviors of pervious blocks for urban drainage basin with storm sewer system.” J. Hydrol., 234(1),
dominated by multiple factors, and the following salient observa- 21–37.
tions are made: Huang, B., Wu, H., Shu, X., and Burdette, E. G. (2010). “Laboratory evalu-
• The prewetting condition caused by the preceding water inflow ation of permeability and strength of polymer-modified pervious
concrete.” Constr. Build. Mater., 24(5), 818–823.
makes the block unsaturated and prompts the runoff at the sub-
Interpave. (2010). “Permeable pavement: Guide to the design, construction
sequent inflow cycles, exhibiting the unique balance between and maintenance of concrete permeable block pavements.” Precast
runoff and drainage; Concrete Paving and Kerb Association, Leicester, U.K.
• The existence of surface layer with fine aggregates and bedding James, W., and Von Langsdorff, H. (2003). “The use of permeable concrete
layer underneath the previous block induces the higher runoff block pavement in controlling environmental stressors in urban areas.”
height, owing to the interplay between capillary force in small 7th Int. Conf. on Concrete Block Paving, National Association of City
pores and the spatial distribution of water phase during evapora- Transportation Officials, New York.
tion; and Jongman, B., et al. (2014). “Increasing stress on disaster-risk finance due to
• The clogging by fine grains deteriorates the drainage capacity large floods.” Nat. Clim. Chang., 4(4), 264–268.
despite the consistently flushed sediments. Kevern, J. T., Schaefer, V. R., Wang, K., and Suleiman, M. T. (2008).
“Pervious concrete mixture proportions for improved freeze-thaw
The experimental results shown in this study can help in design-
durability.” J. ASTM Int., 1–12.
ing the drainage systems in urban areas under heavy rainfall to mit- Korea Meteorological Administration. (2014). “National climate
igate the runoff after repeated rainfall, highlighting the significance data.” 〈http://www.kma.go.kr/weather/climate/extreme_daily.jsp?type=
of prewetting condition. hour_pre&mm=7&x=10&y=6〉 (Dec. 16, 2014).
Korpa, A., and Trettin, R. (2006). “The influence of different drying meth-
ods on cement paste microstructures as reflected by gas adsorption:
Acknowledgments Comparison between freeze-drying (F-drying), D-drying, P-drying and
oven-drying methods.” Cem. Concr. Res., 36(4), 634–649.
This work was supported by the Korea Carbon Capture and KS (Korean Industrial Standards). (2000). “Standard test method for
Sequestration R&D Center (KCRC) grant and the National flexural strength of concrete.” KS F 2408, Seoul.

© ASCE 04016273-9 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273

 KS (Korean Industrial Standards). (2010). “Standard test method for Santamarina, J. C., Klein, K. A., and Fam, M. A. (2001). Soils and waves:
compressive strength of concrete.” KS F 2405, Seoul. Particulate materials behavior, characterization and process monitor-
KS (Korean Industrial Standards). (2014). “Concrete interlocking block for ing, Wiley, New York.
side walk and road.” KS F 4419, Seoul. Schmitt, T. G., Thomas, M., and Ettrich, N. (2004). “Analysis and modeling
Lee, S. H., Kim, J. W., Yoo, I. K., and Kim, N. S. (2012). “Analysis on of flooding in urban drainage systems.” J. Hydrol., 299(3), 300–311.
runoff reduction effects of detachable permeable block.” J. Korean. Seoul Metropolitan Facilities Management Corporation. (2013). “Design-
Soc. Hazard. Mitigation, 12(6), 157–162 (in Korean). ing and building manual of sidewalk.” Seoul.
Lehmann, P., Assouline, S., and Or, D. (2008). “Characteristic lengths affect- Shokri, N., Lehmann, P., and Or, D. (2010). “Evaporation from layered
ing evaporative drying of porous media.” Phys. Rev. E., 77(5), 056309. porous media.” J. Geophys. Res., 115(B6), B06204.
Manahiloh, K. N., Muhunthan, B., Kayhanian, M., and Gebremariam, S. Y. Shokri, N., and Or, D. (2011). “What determines drying rates at the onset of
(2012). “X-ray computed tomography and nondestructive evaluation of diffusion controlled stage-2 evaporation from porous media?” Water
clogging in porous concrete field samples.” J. Mater. Civ. Eng., 10.1061 Resour. Res., 47(9), W09513.
/(ASCE)MT.1943-5533.0000484, 1103–1109. Smith, D. R. (2000). “Permeable interlocking concrete pavements.”
Downloaded from ascelibrary.org by UFC - Universidade Federal do Ceara on 05/07/23. Copyright ASCE. For personal use only; all rights reserved.

Mata, A. (2008). “Sedimentation of pervious concrete pavement system.” Interlocking Concrete Pavement Institute, Washington, DC.
PCA R&D Serial No. SN3104, Portland Cement Association, Skokie, IL. Tennis, P. D., Leming, M. L., and Akers, D. J. (2004). “Pervious concrete
Mehdi-Nejad, V., Mostaghimi, J., and Chandra, S. (2003). “Air bubble pavements.” Portland Cement Association, Skokie, IL.
entrapment under an impacting droplet.” Phys. Fluids, 15(1), 173–183. Thomas, J. J., and Jennings, H. M. (2006). “A colloidal interpretation of
Montes, F., and Haselbach, L. (2006). “Measuring hydraulic conductivity chemical aging of the CSH gel and its effects on the properties of
in pervious concrete.” Environ. Eng. Sci., 23(6), 960–969. cement paste.” Cem. Concr. Res., 36(1), 30–38.
Neithalath, N. (2004). “Development and characterization of acoustically Tong, B. (2011). “Clogging effects of portland cement pervious concrete.”
efficient cementitious materials.” Ph.D. thesis, Purdue Univ., West Ph.D. thesis, Iowa State Univ., Ames, IA.
Lafayette, IN. Wanielista, M., and Chopra, M. (2007). “Performance assessment of port-
Nishiyama, N., Yokoyama, T., and Takeuchi, S. (2012). “Size distributions land cement pervious pavement.” Stormwater Management Academy,
of pore water and entrapped air during drying-infiltration processes of Univ. of Central Florida, Orlando, FL.
sandstone characterized by water-expulsion porosimetry.” Water Weizu, G., and Freer, J. (1995). “Patterns of surface and subsurface
Resour. Res., 48(9), W09556. runoff generation.” Proc., IAHS Publications—Series of Rep.—Int.
Perkins, F. M., Jr. (1957). “An investigation of the role of capillary forces in Association of Hydrological Sciences, International Association of
laboratory water floods.” J. Petrol. Technol., 9(11), 49–51. Scientific Hydrology, Wallingford, U.K., 265–274.
Sansalone, J., Kuang, X., and Ranieri, V. (2008). “Permeable pavement as a Yang, J., and Jiang, G. (2003). “Experimental study on properties of
hydraulic and filtration interface for urban drainage.” J. Irrig. Drain pervious concrete pavement materials.” Cem. Concr. Res., 33(3),
Eng., 10.1061/(ASCE)0733-9437(2008)134:5(666), 666–674. 381–386.

© ASCE 04016273-10 J. Mater. Civ. Eng.

J. Mater. Civ. Eng., 2017, 29(5): 04016273