Temperature and Testing Impacts on Surface Infiltration Rates of Pervious Concrete

Pervious concrete pavement is widely regarded as a stormwater best management practice that has many stormwater and environmental benefits (Haselbach et al. 2014). It may also be used as a structured pavement surface to handle many types of vehicles (Alam et al. 2012). Furthermore pervious concrete systems are also seen as sustainable building practices by reducing heat island effects (Kevern et al. 2012; Haselbach et al. 2011).

Pervious concrete is ideal for parking lots and other uses where there may be a large amount of runoff or excessive pooling may occur. It is a mixture of aggregate, portland cement, and other cementitious material, water, and various admixtures that vary by manufacturer. The porous structure of pervious concrete is obtained by little or no fines being put into the concrete mix, using narrowly graded aggregates, and using less water than traditional concrete pavement, thus leaving voids in the absence of graded aggregates. Because of compaction of the top of the pervious concrete layer during installation, there is a higher strength, lower permeability, and lower void ratio at the surface than the bottom of the placement (Haselbach and Freeman 2006; Delatte et al. 2009).

Pervious concrete pavement systems have the ability to store runoff in a storage basin and also may allow the water to be released into the soil to provide groundwater recharge. However, the rate of infiltration of water into the pervious concrete system is greatly affected by the presence of fine materials and other debris, which may clog the upper surface. These materials can come either from direct application onto the installment from sources such as vehicles, construction or atmospheric deposition, or from runoff from other pavements and landscape surfaces (runon) carrying these materials onto the installment (Hunt and Bean 2006; Mata 2008). Visual inspection can provide a lot of information about pervious concrete, because when there is clogging, it normally first occurs near the surface of the structure (Delatte et al. 2009). However, there are methods available to quantify the impact of the clogging on the surface infiltration rate of pervious concrete.

Double-ring infiltration tests and falling head permeability tests have been used in the past to measure pervious concrete surface infiltration (Delatte et al. 2009; Bean 2005). However, it is difficult to maintain the head in the outer ring required in the double-ring tests owing to the typically very fast infiltration rates, and the falling head test typically covers a much smaller area than the single ring infiltrometer setup used in ASTM C1701 (ASTM 2009). ASTM C1701 is considered to be an accurate and consistent way to measure infiltration rates for a low cost (Brown and Sparkman 2012).

The infiltration rates of pervious concrete placements vary owing to differences in concrete mixes, aggregate size, and the level of compaction after placement. ASTM C1701 was developed because of the need to have a reliable and repeatable test that would be able to measure the infiltration rate of water at any point on a pervious concrete installment. The criteria of the test were for it to be low cost, uncomplicated, nondestructive to the pervious concrete layer, and to be performed with readily obtainable materials. Using these criteria, a test was developed by using a ring of known diameter and a known volume of water and timing how long it took for the water to exfiltrate from the ring. This test can be used both for early quantifications of hydraulic performance of in place concrete systems and later used to identify and further evaluate troubled areas (Brown and Sparkman 2012).

There has also been some preliminary research on the hydraulic performance of pervious concrete in varying temperatures resulting from the change in density and viscosity of water. Water stays liquid for a large temperature range and does not expand or contract much from changes in temperature in the liquid state. However, there are still variations with its hydraulic conductivity because of the change in dynamic viscosity of water as it varies with temperature (Young et al. 2010). The dynamic viscosity of water was found to vary 163% from 0 to 38 °C (Emerson and Traver 2008). This variation is evident in Fig. 1. Because of the hydraulic conductivity being directly correlated to the dynamic viscosity as the dynamic viscosity increases, the hydraulic conductivity decreases and therefore in theory might decrease the hydraulic performance of the pervious concrete installment.

One study showed that the hydraulic performance of pervious concrete installments at Villanova University in Pennsylvania, differed twofold over the course of 1 year. This study only performed 15 performance observations on installments, but the trends showed that the highest infiltration occurred during the end of the summer when temperatures were greatest and the poorest performance during the end of the winter, when temperatures were lowest (Emerson and Traver 2008).

Another research endeavor used 24 rainfall events over the course of 2 years and showed that there is only a minimal decrease in hydraulic efficiency as temperatures become colder. However, there is a significant reduction of performance in terms of infiltration: infiltration decreased as temperatures became colder, and lag time, the amount of time it took for water to begin infiltration, increased. The decrease in infiltration rates may result from the decrease in hydraulic conductivity and dynamic viscosity as previously stated. However the increase in lag time was probably a result of the frost found on the surface of the area. This frost temporarily stopped the infiltration of water and created longer lag times from precipitation to infiltration (Roseen et al. 2009). While comparing pervious concrete and porous asphalt, Houle (2008) found that both pavements had slower infiltration rates during the colder months in winter and the fastest rates during the summer when temperatures were warmer.

The aforementioned studies were based on trends during precipitation events. To further understand this phenomenon and its impact on the surface infiltration performance of pervious concrete systems during various seasons, it was proposed to study the infiltration during more controlled field research. The objective of this research was to study the magnitude of the impact that varying ambient temperatures might have on pervious concrete with respect to surface infiltration into the system. This work was accomplished through field testing over many seasons on a large pervious concrete sidewalk section located in Pullman, Washington, using ASTM C1701 (ASTM 2009).

ASTM C1701 was used to provide quantitative results for surface infiltration (ASTM 2009). A long sidewalk was selected in front of Sloan Hall on the Washington State University campus in Pullman. This sidewalk was installed in the summer of 2012 and has experienced four winters, which are the most impactful to these surfaces. Previous testing with ASTM C1701 on this location and other pervious concrete installations on the Pullman campus indicate adequate hydraulic performance after several years of service without clogging rehabilitation at most locations on the installations. Exceptions were mainly for locations that received large amounts of runon or other soil depositions, such as from landscape beds or construction (Haselbach and Werner 2015). Fig. 2 is a plan view of the Sloan Hall pervious concrete sidewalk. Locations H and G had very reduced infiltration rates because they received runon from several hundred feet of impermeable concrete sidewalk upslope from the pervious concrete placement.

To decrease the influence of clogging from runon in these tests, the four locations as shown in Fig. 2 that were most downslope were selected to be tested (A, B, C, and D). Mata (2008) showed that if there is excessive runon, then clogging from sedimentation can occur, which may affect lower infiltration rates, so the chosen locations were those most remote from additional sidewalk runon. During the first winter of testing, Location A and then Location B were also eventually excluded from further testing because they had received significant clogging from the neighboring landscape bed and snowpiling (Location A) or from a pile of dirt from landscaping work (Location B).

Before performing each test following ASTM C1701, many variables were recorded: the temperatures of the water used to run the test (this water was set out the night before the tests to bring the water temperature close to the ambient air temperature), the air at the site, and the surface of the pervious concrete installment with an infrared thermal gun. The relative humidity, weather, and time the test took place were also recorded on site. After initial recordings were made, a modified version of ASTM C1701 was used. The method that was changed from ASTM C1701 was to use only a 3.79 L (1 gal.) test no matter how long the prewetting took (ASTM 2009). The use of the test with the smaller amount of water was selected because of the slower infiltration rates when testing the four locations for infiltration.

To perform ASTM C1701, a single-ring infiltrometer was sealed using plumbers putty over the center of the test location. Then a 3.79 L (1 gal.) prewetting was performed, and the time for infiltration was recorded. If infiltration took longer than 30 min, then the test was concluded and insufficient infiltration was recorded. After the prewet test, there were two consecutive separate tests performed with 3.79 L (1 gal.) of water and the time recorded. These will be referred to as tests alpha ($\alpha$) and beta ($\beta$), respectively. Two consecutive tests were originally performed for additional averaging based on timing and variability. However, during the testing periods, other research found that when pervious concrete is in use, a second test may be slower, especially after rejuvenation, most likely because of fines being carried into pores with the testing waters, some of which may have been dislodged during cleaning (Haselbach et al. 2016).

Testing was performed for Locations A–D approximately once per week when the weather was cold, and approximately every 2 weeks once the weather became warmer through parts of two consecutive winter seasons. This was done to provide a variety of weather conditions and temperatures in which the infiltration testing was performed. Following the guidelines that were adapted from ASTM C1701, the time to infiltrate the water was recorded. After the time was recorded, the infiltration rate was calculated based on the surface area in the single ring infiltrometer.

For each location of the four locations on each date, the test was performed twice in succession ($\alpha$ and $\beta$). Table 1 shows the results at each location. As previously noted, Locations A and B became significantly clogged during the testing in the first winter; because of this, some tests at those locations were not continued that year. For the one case at Location B, which became extremely clogged when landscaping took place next to the installment, an effort was made to try and remove the surface dirt. A broom was used to sweep off some of the dirt on March 5, 2015, but it did not make a very noticeable difference in infiltration rates. Fig. 3 is a photograph of the surface clogging that was experienced at Location B. The dot to the right of the dirt in Fig. 3 is a locator point created to center the single ring infiltrometer used to perform the modified version of ASTM C1701 while testing. Subsequent cleaning of those two locations with a garden hose and nozzle in the fall of 2015 resulted in improved infiltration rates, and the locations were added again to the testing protocol. Included in Table 1 are the test dates and times, the air and water temperatures at which the tests were performed; the top row is the average infiltration rate found at these locations in the spring of 2014 (Haselbach and Werner 2015). Table 1 shows that there is a significant loss of performance from the observed areas in just over a year’s time from spring 2014 to spring 2015, this is believed to be primarily attributable to clogging and sedimentation in the area.

Fig. 4 provides a graphical representation of the effects that time (i.e., more clogging, rejuvenation and/or temperature) might have on the average tested ($\alpha$ and $\beta$) surface infiltration rates. Surface infiltration rates for pervious concrete installations can easily change by orders of magnitude over time as soils and other materials fill some of the pores. The overall decrease in rate with time is very apparent, as is the improvement with the hosing rejuvenation. (In addition, there is variability in the measurements owing to timing errors, small changes in surface area as the putty is placed around the infiltrometer, perhaps from temperature.)

Data in Table 1 shows that the second ($\beta$) of the two consecutive tests on each date of testing tends to be less than the first ($\alpha$). Fig. 5 further supports this. The slope of the trend line is 1.05 with an $R^2$ of 0.97. This not only confirms the predictions from the earlier research that the $\beta$ test is slower than the $\alpha$ after rejuvenation by washing, but it also occurs for regular testing, perhaps from new fines being deposited on the surface and transported into the pores in the first ($\alpha$) test.

Fig. 6 provides information about the water temperature and surface infiltration rates over the two winter testing period for all four locations. In Fig. 6, only the rate from the second test ($\beta$) is used because this was previously shown to usually be the slowest, and the water temperature is used because this is more representative of the viscosity variations that might occur. It is again apparent that clogging impacts infiltration rates much more than temperature because no trend can be found with temperature variations from season to season, and Fig. 4 showed the clogging impacts. However, if one inspects the temperature and infiltration variations for Locations A and B closely over the 2015 winter months in Fig. 6, there appears to be a slight correlation of increased infiltration rates with increased temperature, but it is not readily apparent because of the confounding impacts of clogging.

To evaluate the possible impacts of temperature further from the testing periods, as differentiated from the clogging impacts, the water temperature versus infiltration rate data were discretely segregated by infiltration rates as a surrogate measure of relative clogging. Fig. 7 presents this information for segregated information over every $100\mathrm{cm}/\mathrm{h}$ infiltration range. No confirmed trends were seen with increased viscosity at lower temperatures, even in the lowest range (under $100\mathrm{cm}/\mathrm{h}$) of infiltration, where viscosity might have a larger impact.

No statistical correlation is seen between infiltration rates and water temperature in the field experiments performed as hypothesized. The possible impacts of water temperature and viscosity are masked by the impacts of clogging over time, although there is a slight trend of slower rates with lower temperatures when the infiltration rates are very low. Therefore, infiltration testing carried on in warmer months should be used conservatively for estimating wintertime hydraulic performance in areas that experience colder temperatures. This study did not evaluate the impact that icing, snow cover, and wintertime conditions other than temperature might have on pervious concrete surface infiltration rates. As noted in Roseen et al. (2009), surface icing can decrease infiltration rates dramatically. Additional research under these conditions would be beneficial in understanding how a pervious concrete system might recover after significant icing or snow cover.

This study did find that the use of ASTM C1701 for evaluating surface infiltration rates of pervious concrete field placements can vary after an initial test ($\alpha$) as was seen in laboratory tests in previous research (ASTM 2009). This phenomenon occurs not just after rejuvenation by washing, possibly from dislodged debris re-entering the pores, but was found to occur in the majority of field tests performed over time, most likely because of surface deposited fines moving into the pore structure in the first ($\alpha$) test. Therefore it is recommended to always perform the second test as allowed by the standard to obtain more information about the placement’s subsequent performance. However, the volume of the test itself may cause this clogging more than several precipitation events might because of the relatively large amount of water used compared to a typical rain storm.

The authors thank other students at Washington State University for assistance in performing these tests, namely, Daniel Noh, Daniel Hendrickson, and Trace Sendele. Funding for the work was obtained from the Washington State Department of Ecology, the Husseman Fund, and from the USDOT University Transportation Center, Center for Environmentally Sustainable Transportation in Cold Climates (CESTiCC).