Rc

Aa

cb

a

ARRAA

KHFCC

1

a(pCoe

rr

tfriOt(ht

0d

Ecological Engineering 37 (2011) 754–757

Contents lists available at ScienceDirect

Ecological Engineering

journa l homepage: www.e lsev ier .com/ locate /eco leng

eliability, repeatability and accuracy of the falling head method for hydrauliconductivity measurements under laboratory conditions

nna Pedescoll a, Roger Samsóa, Enrique Romerob, Jaume Puigaguta, Joan Garcíaa,∗

Environmental Engineering Division, Department of Hydraulic, Maritime and Environmental Engineering, Technical University of Catalonia,/Jordi Girona 1-3, Building D1, 08034 Barcelona, SpainDepartment of Geotechnical Engineering and Geo-Sciences, Technical University of Catalonia, c/Jordi Girona 1-3, Building D2, 08034 Barcelona, Spain

r t i c l e i n f o

rticle history:eceived 26 January 2010eceived in revised form 6 June 2010ccepted 8 June 2010vailable online 8 July 2010

a b s t r a c t

The aim of this study was to verify under lab conditions the reliability, repeatability and accuracy of thefalling head method (FHM) for hydraulic conductivity measurements. The FHM is a reliable procedure thathas slight variations (less than 10%) in repeated measurements and turns out to be a reliable techniqueto record the hydraulic conductivities typically described for clogged and unclogged subsurface-flowconstructed wetlands (from 4 to ca. 360 m/day). The accuracy of the method is acceptable considering

eywords:ydraulic conductivityalling head methodonstant head methodlogging

difficulties in the measurement of hydraulic conductivity in highly conductive media. Accordingly, resultsshow measurement deviations of 20% when compared with a laboratory constant head method for highlyconductive media (higher than 250 m/day), and 80% for media with low hydraulic conductivity (lowerthan 50 m/day). The main conclusion of the present paper is that of the FHM is a reliable and repeat-able technique for hydraulic conductivity measurements and it is accurate enough for on-site clogging

nstru

ofiwlsrtbePia

tmccs

assessment in full-scale co

. Introduction

Subsurface-flow constructed wetlands (SSF CWs) representsuitable technology for the sanitation of small settlements

Puigagut et al., 2007). However, SSF CWs filter media suffer fromorosity reduction over time (clogging) (Wallace and Knight, 2006).logging leads to hydraulic malfunction that prevents normalperation and, in the worst-case scenario, can affect treatmentfficiency (Rousseau et al., 2005).

When clogging is detected, several techniques can be applied toeverse it. However, such techniques and especially filter mediumeplacement can be very costly (Pedescoll et al., 2009).

For treatment and financial reasons, it is essential to assesshe degree of clogging in SSF CWs. Therefore, indirect measuresor clogging assessment have been given special attention in cur-ent literature. Some of the most widespread measures of cloggingnclude the analysis of accumulated solids in filter media (Caselles-sorio et al., 2007), system hydrodynamics by means of tracer

ests (Bowmer, 1987) and hydraulic conductivity measurements

Sanford et al., 1995; Rodgers and Mulqueen, 2006). Specifically,ydraulic conductivity measurements have proven to be a suitableechnique (Knowles et al., 2010; Pedescoll et al., 2009).

∗ Corresponding author. Tel.: +34 93 401 6464; fax: +34 93 401 7357.E-mail address: joan.garcia@upc.edu (J. García).

2

2

f

925-8574/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.ecoleng.2010.06.032

cted wetlands.© 2010 Elsevier B.V. All rights reserved.

The measurement of hydraulic conductivity to assess the degreef clogging in SSF CWs is not a straightforward procedure. Dif-culties arise because the clogged filter medium of constructedetlands is of non-cohesive nature (generally gravel with accumu-

ated solids), which makes it virtually impossible to take unalteredamples of material to perform standardized and controlled labo-atory tests. For this reason most of the techniques that are usedo study the hydraulic conductivity of constructed wetlands areased on in situ procedures (Reynolds et al., 2000; Langergrabert al., 2003; Caselles-Osorio et al., 2007; Knowles et al., 2010;edescoll et al., 2009). The disturbance caused by these methodss small compared with that caused during the extraction, storingnd transportation of a gravel core to be tested in lab facilities.

The main objective of this study was to assess under lab condi-ions the reliability, repeatability and accuracy of the falling head

ethod for the measurement of hydraulic conductivity. The mainonclusions of the present work will determine the potential appli-ation of the method for on-site clogging assessment of full-scaleubsurface-flow constructed wetlands.

. Methods

.1. The falling head method

The falling head method (FHM) is based on Lefranc’s test withalling heads (NAVFAC, 1986), which has been used to measure

A. Pedescoll et al. / Ecological Engineering 37 (2011) 754–757 755

ory pe

tmatilaesL

K

wmoti

oh

2

scMchmo

FFctpeewvddwTAmofipwa

tpt

K

Fig. 1. Diagram of the laborat

he saturated hydraulic conductivity of constructed wetlands’ filteredia (Pedescoll et al., 2009). FHM consists of measuring the timecolumn of water inside a tube takes to drop a certain height. The

ube (made of steel for example, and perforated at its bottom) isnserted within the granular media by hammering until the waterevel of the wetland is reached. The tube is filled with water inpulse mode. As a result of water flowing out the tube a negativexponential curve of the water level inside the tube is obtained, thelope of which is related to the hydraulic conductivity according toefranc’s formula:

= d2 ln(2L/d)8Lt

lnh1

h2

here K is the hydraulic conductivity of the studied material, in/s; d is the diameter of the tube, in m; L is the submerged length

f the tube, in m; h1 is the height of the water table level insidehe tube at time zero, in m; h2 is the height of the water table levelnside the tube at time t, in m; and t is time, in s.

The quadratic difference between the theoretical curve and thatbtained in the field is minimized to estimate the value of theydraulic conductivity.

.2. The constant head permeameter

The accuracy of the FHM was assessed by comparison with atandardized method for laboratory measurements of hydrauliconductivity (a constant head method, CHM) (Rodgers andulqueen, 2006). A permeameter was built to measure hydraulic

onductivity by the CHM. The aim of the permeameter was to allowydraulic conductivity to be measured in a wide range of sampleaterials. Thus, it was designed to be suitable for measurements

f both low and high conductivity materials (Fig. 1).

wmd

rmeameter (constant head).

To run the CHM test, a constant water flow rate was required.or this purpose, water was stored in an upper reservoir (750 L).rom the reservoir, the water flowed through the permeameterell (made from metacrylate, 70 cm long and 11.4 cm diame-er), in which the granular material was placed. Attached to theermeameter cell were 3 piezometric tubes (10 cm apart fromach other). Once water had passed through the medium, itnded up in a lower reservoir, in which a constant water levelas maintained by a spillway. The height of the lower reser-

oir was changed to vary the overall test pressure gradient onemand. When water flowed through the medium, a pressurerop was obtained by friction with particles. This pressure dropas monitored with a GE Druck LPM 5480 Differential Pressure

ransductor (DPT) that allows measurements from 0 to 15 mbar.calibration procedure was used to correctly convert the infor-ation provided by the DPT (in Volts) to pressure units (in cm

f water column). The DPT calibration was carried out by arti-cially generating pressure differences between the apparatusiezometers (by applying different flow rates) before the testsere run. The calibration curve, with an r2 = 0.999, was Volt-

ge = 0.174 Pressure + 0.823.If we know the vertical distance between the piezometric tubes,

he cross-sectional area of the permeameter cell and the hydraulicressure differences measured in the tests, the hydraulic conduc-ivity of the sample can be determined with Darcy’s Law as follows:

= Q

A

L

�H

here K is the hydraulic conductivity of the studied material, in/s; Q is the flow rate through the sample, in m3/s; L is the vertical

istance between piezometric tubes, in m; A is the cross-sectional

756 A. Pedescoll et al. / Ecological Engineering 37 (2011) 754–757

Table 1Relevant physical characteristics of the laboratory tanks used to evaluate the reliability and repeatability of the falling head test method.

Surface area (m2) Length to width ratioa Water depth (m) Granular medium

D50 Porosity (%) Filter media hydraulic conductivity(m/day)b

Sand tank 0.74 1:0.7 0.25 0.9 31 32Gravel tank 0.74 1:0.7 0.25 7.1 39 265

ap

Fttat

2

ei

tAiammamrerlwnamlwtcTa

gnsmaimestam

cdtoc

2

aAtap

3

3

TAe

n

a Length to width ratio of each individual cell.b Measured in this study according to the constant head method.

rea of the permeameter cell, in m2; �H is the head loss betweeniezometric tubes, in m.

Two conditions must be met in order to apply Darcy’s Law.irstly, the saturation index must be over 85% (Arnold, 1995). Tohis end, the permeameter cell was filled slowly with water fromhe bottom to the top prior to any test, which forced the entrappedir bubbles to leave the medium pores. Secondly, the flow duringhe test must be laminar (Reynolds Number < 12).

.3. Lab experiments

In the present investigation a lab test series was performed tovaluate the reliability, repeatability and accuracy of the FHM forn situ SF CWs clogging evaluation.

In order to assess the reliability and repeatability of the FHMhree experimental conditions were considered during the lab test.ccordingly, the reliability of the method was assessed by address-

ng (1) the effect of the application of the method for highly cloggednd non-clogged wetlands using clean gravel and sand as a modeledia for non-clogged and highly clogged scenarios (first experi-ental condition). The repeatability of the method was assessed by

ddressing (2) the potential differences on hydraulic conductivityeasurements recorded due to the location of the test performed,

egardless the water infiltration properties of the media and (3) theffect of tube extraction/no extraction while the test is performed,egardless the water infiltration properties of the media. For theab tests two 300 L tanks (0.95 m long × 0.70 width × 0.45 m depth)

ere set up. One of the tanks was filled with gravel (as model foron-clogged scenario) and the other one was filled with sand (asmodel for a clogged scenario). Each tank was filled with 30 cmedia layer for each tested substrate. To keep a constant water

evel (5 cm below the material surface), a drain tube in the tanksas connected to a spillway. This was set up at the base of each

ank in each bed (tanks were filled with tap water for the tests). Theharacteristics of the sand and gravel tanks can be found in Table 1.he hydraulic conductivity of the filter media was determined bypplying the FHM in 3 evenly spaced points along the central lon-

rcws

able 2verage values ± SD of hydraulic conductivity obtained with the falling head method apvaluation of reliability and repeatability of the method. Each average based on n = 3 for F

FHM

Measurement without tube extraction Mea

Gravel

Inleta 349.7 ± 41.1 359.Middlea 365.3 ± 55.7 338.Outleta 334.3 ± 17.9 332.

nsd

Sand

sda

Inleta 4.3 ± 0.6 5.0 ±Middlea 4.0 ± 1.0 4.3 ±Outleta 6.0 ± 1.7 4.0 ±

s: no significant differences according to three-way ANOVA; sd: significant differences aa Statistical analyses between tests carried out with FHM and CHM with gravel and sanb Statistical analyses between near inlet, middle and near outlet location of measuremec Statistical analyses between gravel and sand as filter media.d Statistical analyses between tests carried out with or without tube extraction in the t

itudinal section of each tank: near the inlet, in the middle andear the outlet (beside the drain tube). Each measuring point waseparated from the others by about 25 cm. Some of these measure-ents were carried out by extracting the metal tube and inserting it

gain, whereas others were carried out repeatedly without extract-ng the tube. In these experiments, the hydraulic conductivity was

easured three times in each test and the following cases werevaluated: (1) the effect of the filter media (by comparing the mea-urements between filter media), (2) the effect of extracting theube (by comparing measurements with or without extraction),nd (3) the effect of the proximity of the drain tube (by comparingeasurements at different locations).In order to assess the accuracy of the FHM under lab-controlled

onditions, gravel and sand were also analysed following the CHMescribed in Section 2.2 and compared to those results obtained byhe application of the FHM. A total of 14 and 9 tests were conductedn sand and gravel respectively in order to estimate the hydrauliconductivity.

.4. Statistical analyses

Differences among experimental conditions were assessed bypplying the ANOVA test of variance (one-way or three-wayNOVA depending on the experimental design). All data subjected

o the ANOVA test met the required conditions for the test to bepplied. All statistical analyses were carried out using the softwareackage SPSS 17.

. Results and discussion

.1. Reliability and repeatability of the falling head method

Lab tanks filled with sand and gravel were used to assess the

eliability and repeatability of the FHM, and three experimentalonditions that may affect hydraulic conductivity measurementsere considered. The first experimental condition addressed the

uitability of using the FHM for the analysis of low conductive and

plied to the two lab tanks with different filter media (sand and gravel) used forHM and n = 5 and 9 for CHM with gravel and sand, respectively. Data are in m/day.

CHM

surements with tube extraction

7 ± 50.1251.9 ± 78.4 nsb

sdc

3 ± 56.80 ± 29.3

0.037.0 ± 4.7 nsb0.6

0.0

ccording to three-way ANOVA.d as filter media.nts.

anks for gravel and sand as filter media.

l Engi

h(ersfgvrra

wt(aarHubamelfi

3

mr2m

t(t(vTr3a

bwoowhmsdmpc2

4

s

sea

ttcmtici

ocs

A

M(fAvgPCa

R

A

B

B

C

K

L

N

P

P

R

R

R

S

W

A. Pedescoll et al. / Ecologica

igh conductive substrates in SSF CW. According to the resultsTable 2), the FHM appears to be a suitable technique for thevaluation of clogging in SSF CW, since it is sensitive enough toecord high significant differences in hydraulic conductivity mea-urements between the gravel and the sand media (values rangedrom ca. 330 to 365 m/day and from ca. 4.0 to 6.0 m/day for theravel and sand media, respectively). These hydraulic conductivityalues are in the range of those described for different granulomet-ic materials (Wilson et al., 2000), and also match those previouslyecorded in the literature for severely and not severely cloggedreas of SSF CW (Caselles-Osorio et al., 2007).

When both the location factor and the effect of tube extractionere analysed (second and third experimental condition, respec-

ively) differences in hydraulic conductivity within the same mediagravel or sand) are not significant (p < 0.05) (Table 2). The variationmong treatments was less than 10% (around 2 and 7% for gravelnd sand media, respectively). Therefore, the FHM is reliable andepeatable for a wide range of hydraulic conductivity conditions.owever, it is necessary to point out that in full-scale facilitiesnder highly clogged conditions the variation of the method coulde higher than that recorded here (variation for sand media iss low as 7%). Accordingly, higher variation for on-site assess-ent of hydraulic conductivity in highly clogged wetlands could be

ncountered due to the different properties of the solids accumu-ated. The use of sand media to assess the reliability of the methodor highly clogged wetlands has limitations and, therefore, cautions required.

.2. Accuracy of the falling head method

It is widely accepted that the CHM represents a standardizedethodology for the assessment of hydraulic conductivity in a wide

ange of materials under lab conditions (Rodgers and Mulqueen,006). Therefore, by comparing the results obtained with bothethods, the accuracy of the FHM will be assessed.Among CHM tests conducted either on sand and gravel, only

he tests with Reynolds Number less than 12 were considered14 and 5 tests for sand and gravel, respectively). Results onhe hydraulic conductivity measured ranged from 30 to 45 m/dayaverage value of 35 m/day) and from 177 to 364 m/day (averagealue of 265 m/day) for the sand and gravel media, respectively.herefore a fairly variation was observed for the gravel, which iselated to its infiltration velocity (which ranged from 1.3 × 10−4 to.7 × 10−4 m/s and from 2.8 × 10−3 to 7.0 × 10−3 × m/s in the sandnd gravel media, respectively).

Differences in the average values of hydraulic obtained withoth methods were found, regardless of the media considered (one-ay ANOVA test of variance; p < 0.05) (Table 2). Specifically, values

btained from the FHM method were ca. 80% lower than valuesbtained with the CHM for the sand media and ca. 20% higherhen gravel media was considered (Table 2). These deviations,owever, are acceptable in the context of hydraulic conductivityeasurements (Bagarello et al., 2004). To this regard it is neces-

ary to point out that exact values of hydraulic conductivity areifficult to obtain because variations can exist depending on theeasuring technique used, flow rate and system dimensions, sam-

le collection technique and size, and physical and hydrologicalharacteristics of the media (Reynolds et al., 2000; Wilson et al.,000).

. Conclusions

The falling head method (FHM) for hydraulic conductivity mea-urements is a reliable technique for on-site clogging assessment

W

neering 37 (2011) 754–757 757

ince it is able to discriminate among hydraulic conductivities gen-rally recorded for non-clogged and clogged facilities (between 200nd 300 m/day and less than 50 m/day, respectively).

The FHM appears to be a repeatable technique (at least underhe conditions here considered). Accordingly, a hydraulic conduc-ivity variation lower than 10% was recorded among conditionsonsidered (tube extraction/no extraction and the location of theeasurement), regardless the infiltration properties of the media

ested. However, it is necessary to point out that in full-scale facil-ties under highly clogged conditions the variation of the methodould be higher than that recorded here (variation for sand medias as low as 7%). Therefore, caution is required.

Even though the FHM is not as accurate as normalized meth-ds such as the CHM under lab conditions (especially for lowonductive media), it is still accurate enough to assess in situ mea-urements of hydraulic conductivity for clogging assessment.

cknowledgements

This work was made possible by funding from the Spanishinistry of Innovation and Science for the NEWWET2008 Project

CTM2008-06676-C05-01). Roger Samsó also acknowledges theund provided by the Universitat Politècnica de Catalunya (UPC).nna Pedescoll kindly acknowledges the Spanish Ministry of Inno-ation and Science for her scholarship. The authors are particularlyrateful to Elisenda Alba, Cristina Ávila, Javier Carretero and Miriamlanas for their help with lab work. In addition, we acknowledge theatalan Water Agency, Francesc Llenas and Maria Llorens (Sorea)nd Paula Aguirre and Maribel Carrasco (Rubatec).

eferences

rnold, P.W., 1995. Manual of Soil Laboratory Testing, vol. 2. Permeability, Shearstrength and Compressibility tests, 2nd ed. Pentech Press, London.

agarello, V., Iovino, M., Elrick, D.E., 2004. A simplified falling-head technique forrapid determination of field-saturated hydraulic conductivity. Soil Sci. Soc. Am.J. 68 (1), 66–73.

owmer, K.H., 1987. Nutrient removal from effluents by artificial wetland: influenceof rhizosphere aeration and preferential flow studied using bromide and dyetracers. Water Res. 21 (5), 591–599.

aselles-Osorio, A., Puigagut, J., Segú, E., Vaello, N., Granés, F., García, D., García,J., 2007. Solids accumulation in six full-scale subsurface flow constructed wet-lands. Water Res. 41 (6), 1388–1398.

nowles, P.R., Griffin, P., Davies, P.A., 2010. Complementary methods to investi-gate the development of clogging within a horizontal sub-surface flow tertiarytreatment wetland. Water Res. 44 (1), 320–330.

angergraber, G., Haberl, R., Laber, J., Pressl, A., 2003. Evaluation of substrate clog-ging processes in vertical flow constructed wetlands. Water Sci. Technol. 48 (5),25–34.

AVFAC, 1986. Soil Mechanics. Design Manual 7.01. Naval Facilities EngineeringCommand, Alexandria, USA.

edescoll, A., Uggetti, E., Llorens, E., Granés, F., Garcia, D., García, J., 2009. Practicalmethod based on saturated hydraulic conductivity used to asses clogging insubsurface flow constructed wetlands. Ecol. Eng. 35 (8), 1216–1224.

uigagut, J., Villasenor, J., Salas, J.J., Bécares, E., García, J., 2007. Subsurface-flowconstructed wetlands in Spain for the sanitation of small communities: a com-parative study. Ecol. Eng. 30 (4), 312–319.

eynolds, W.D., Bowman, B.T., Brunke, R.R., Drury, C.F., Tan, C.S., 2000. Comparison oftension infiltrometer, pressure infiltrometer, and soil core estimates of saturatedhydraulic conductivity. Soil Sci. Soc. Am. J. 64, 478–484.

odgers, M., Mulqueen, J., 2006. Field-saturated hydraulic conductivity of unsatu-rated soils from falling-head well tests. Agric. Water Manage. 79 (2), 160–176.

ousseau, D.P.L., Horton, D., Vanrolleghem, P.A., De Pauw, N., 2005. Impact of oper-ational maintenance on the asset life of storm reed beds. Water Sci. Technol. 51(9), 243–250.

anford, W.E., Steenhuis, T.S., Parlange, J.Y., Surface, J.M., Peverly, J.H., 1995.Hydraulic conductivity of gravel sand as substrates in rock-reed filters. Ecol.Eng. 4 (4), 321–336.

allace, S.D., Knight, R.L., 2006. Small-scale Constructed Wetland TreatmentSystems: Feasibility, Design Criteria and O&M Requirements. Final Report,

Project 01-CTS-5, Water Environment Research Foundation (WERF), Alexandria,USA.

ilson, M.A., Hoff, W.D., Brown, R.J.E., Carter, M.A., 2000. A falling head permeame-ter for the measurement of the hydraulic conductivity of granular solids. Rev.Sci. Instrum. 71 (10), 3942–3946.