Marine Pollution Bulletin 115 (2017) 154–163

Contents lists available at ScienceDirect

Marine Pollution Bulletin

journal homepage: www.elsevier.com/locate/marpolbul

Oil permeability variations on lagoon sand beaches in the Patos-Guaíba

system in Rio Grande do Sul, Brazil
Elaine Baroni Oliveira ⁎, João Luiz Nicolodi
Institute of Oceanography, Federal University of Rio Grande/FURG, Av. Itália, Km 8, CEP 96203-900 Rio Grande, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Permeability is the ability of a sediment deposit to allow fluids to pass through it. It depends on the local types of
Received 12 September 2016 sediments. When the fluid is oil, high permeability implies greater interaction with the site and more extensive
Received in revised form 1 December 2016 damage, which makes recovery most difficult. Knowledge of permeability oscillations is necessary to understand
Accepted 7 December 2016
oil behavior and improve cleanup techniques. The goal is to determine oil permeability variations on lagoon sand
Available online 10 December 2016
beaches. Oil permeability tests were performed at the beach face, using a Modified Phillip Dunne Permeameter
Keywords:
and parameters were sampled. Permeability of lagoon beaches is driven by grain diameter and roundness, soil
Permeability compaction, and depth of the water table. Factors that enhance permeability include: sand sorting, vertical dis-
Hydraulic conductivity tribution of sediments and gravel percentage. High permeability on lagoon beaches is related to polymodal dis-
Crude oil tribution, to the sediment package, and to the system's low mobility.
Patos-Guaíba system © 2016 Elsevier Ltd. All rights reserved.
MPD permeameter

1. Introduction
Permeability is defined as the ability of sediment to allow fluids to
flow through it. This is measured by a hydraulic conductivity coefficient,
which depends on the fluid and environmental characteristics. When
the fluid is oil, permeability determines infiltration and the contaminant's
residence time in the substrate. According to Stern (2007), the longer
the oil's residence time in an area, the higher its interaction with the
environment, and the stronger will be its impact, making recovery
more difficult.
High substrate permeability causes subsurface oil deposits, which
worsens the situation due to a few factors: First, in the subsurface oil
degradation by bacteria is limited, due to low oxygen availability
(Rowland et al., 2000). Second, the contaminated volume may be
underestimated, since visual measures are ineffective (Owens, 1988).
And finally, the variations in the beach sediments package may form in-
terleaved layers of sand and oil, which can be released through time
(NOAA, 1992).
Because it depends on sediment characteristics, hydraulic conduc-
tivity will vary according to its location (ocean, lake, lagoon, etc). More-
over, sedimentary changes in time may alter permeability, leading to
different infiltration rates, causing problems during the cleaning pro-
cess, besides not corresponding to pre-existing oil sensitivity rankings
for the area (Tsouk et al., 1985).
Soil K value can either be measured or predicted. Most natural soils
have spatially variable hydraulic properties. This implies that many K
⁎ Corresponding author.
E-mail address: elaine.bo@hotmail.com (E.B. Oliveira).
data are necessary to adequately characterize the field's K value. Most
projects do not have a budget to perform many field and laboratory per-
meability tests, since those are time-consuming and more costly than
predictions. This is why simple methods are used to predict the hydrau-
lic conductivity (Chapuis, 2012).
The official Brazilian method to determine oil sensitivity (MMA,
2007) establishes standard behaviors of oil flow based, mostly, on sub-
strate structure, where permeability is crucial to this definition. Howev-
er, permeability estimates in the current methodology is based on grain
size, which in turn is visually determined by the researcher on the
beach, excluding any variation. In addition, oil behavior is based on an
ocean beach, but considered the same for beaches located in any other
system (like lagoons).
Actual oil behavior does not always match predictions. This is why
methodology improvement is necessary and justified. Some efforts to
improve oil sensitivity have been proposed by Moe et al. (2000),
Castanedo et al. (2009), Fataal et al. (2010), Marques and Nicolodi
(2015), and Marinho (2015). In this context, and in order to increase
knowledge of oil behavior and permeability and advance the existing
methodology, this article aims to determine oil permeability on lagoon
beaches, as well as define the main parameters that determine the flow.
2. Methods
2.1. Study area
The area chosen for the study comprises four lagoon beaches in the
“Patos-Guaíba” System in Rio Grande do Sul (RS), Brazil, and can be
seen in Fig. 1. These beaches are Arambaré Norte (UTM 452819,

http://dx.doi.org/10.1016/j.marpolbul.2016.12.020
0025-326X/© 2016 Elsevier Ltd. All rights reserved.
 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163
Fig. 1. Location of the studied area in the Patos-Guaíba System, Rio Grande do Sul, Brazil.
6580432), Arambaré Sul (UTM 452250, 6578548), Praia Alegria in
Guaíba city (UTM 469869, 6665665), Ipanema in Porto Alegre (UTM
477865, 6665860) and Mostardas (UTM 503033, 6567971). Arambaré
was divided in two samples (Norte and Sul) due the differences be-
tween the sediments caused by the Velhaco River which flows out
onto the beach.
The Patos-Guaíba system is located on the coastal plain of Rio
Grande do Sul. On the northern end of the coastal plain the adjacent
highlands consist of Paleozoic and Mesozoic sedimentary and volcanic
rocks of the Paraná Basin that locally reach 1000 m. On the southern sec-
tion, igneous and metamorphic rocks of the Pre-Cambrian shield form
lower highlands. At present, all sandy sediments eroded from these
highlands and carried by rivers to the coast are trapped in the coastal la-
goons and other back-barrier environments, and none reaches the
ocean shoreline (Tomazelli et al., 2000).
The coastal plain is dominated by a bimodal high-energy wind re-
gime. The dominant wind comes from NE and is more active in spring
and summer months. The secondary W-SW wind becomes more impor-
tant in the autumn-winter months. The coast is dominated by
microtidal waves, with semidiurnal tides averaging 0.5 m (Motta,
1969; Calliari et al., 1998; Tomazelli et al., 2000).
The wind pattern has great influence on the velocity of groundwater
flow, as it defines the direction of surface flow and the water percolation
rate in the sediment. The driving force generated by the lagoon's surface
flows moves continental groundwater toward the sea. However, a re-
verse process, caused by the concentration gradient, transports back to-
ward the lagoon (Machado, 2007).
The Patos Lagoon is described by Toldo Jr. (1994) as a shallow and
wide body of water ranging in width from 10 to 60 km. The total length
is 240 km. The sediments on the western part of the lagoon are poorly
sorted ranging in size from fine sand to gravel, while the sediments
along the eastern bank clearly show sorting and fine grain size
(Martins et al., 1989).
155
3. Methods
The methodology adopted included oil infiltration tests and param-
eters (compaction and water table depth) and sediments sampling. The
permeability tests, compaction, and water table tests were performed
during a fieldwork on each beach.
The permeability tests were conducted with the Modified Phillip-
Dunne Infiltrometer (MPD). The MPD was installed at the beach face
of the beaches studied. Six tests were performed for each beach in two
perpendicular profiles, named 1 and 2. Point A is located next to the wa-
terline, Point C is on the berm, and Point B is halfway between Points A
and C, as shown in Fig. 2. The tests were authorized by the proper Envi-
ronmental Agency in charge (authorization 127/2015-DL).
The fluid used in the tests was crude oil (API 33.54 and density
0.8549 g/cm3) donated by Alberto Pasqualini Refinery (REFAP). In
each test, the MPD was filled with 750 ml of oil and the fluid dripping
time was recorded every 0.5 cm. After infiltration was complete, the
MPD was removed and a stratigraphic cut was conducted for subse-
quent image capture. The infiltration plume (corresponding to the
blob formed by the oil in the sand) was photographed and measured.
After the test, the contaminated sand was removed and stored in bar-
rels. The barrels were taken to LOG – the Geological Oceanography Lab-
oratory of FURG – Federal University of Rio Grande, where they were
stored until final destination.
In the laboratory, dripping time and height variation were input
onto a macro to calculate hydraulic conductivity using Philip's (1993)
theory. The macro was developed by St. Anthony Falls Laboratory of
the University of Minnesota. The macro creation tutorial is available in
Ahmed et al. (2011) and Nestingen (2007). The data was also input in
other empirical hydraulic conductivity equations in order to identify
the effectiveness of each equation in the “Patos-Guaíba” system.
The compaction tests were conducted with a manual penetrometer
whereas the penetration resistance to 40 kg/cm2 and 50 kg/cm2 loads
156 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163

B C
A

Fig. 2. Permeability tests. A) Modified Phillip Dunne Infiltrometer (MPD); B) Figure representing the profiles and points. At the top is a cross-shore profile and below a plan view; C)
Permeability test being performed.

was inserted in spreadsheets. The water table depth was measured with
a ruler in a hole drilled by a hand auger.
The granulometric analysis procedure was performed by manual
sieving using ¼ phi range sieves. The morphometric analysis was per-
formed by visual comparison using Krumbein (1941), Rittenhouse
(1943) and Bigarella et al. (1955) letters. The stratigraphy was analyzed
by the substrate cut. The statistic results were calculated using the
rysgran package (Gilbert et al., 2014) from R software (R Core Team,
2016).
The sediment and soil results were compared to the permeability
through scatter plots in R software in order to understand the behavior
of oil in the environment and find possible correlations. Polynomial
trendlines were generated in R software for each variable in order to de-
termine the occurrence of a significant correlation between the param-
eters and the hydraulic conductivity. Linear trendlines were generated
between the MPD and the empiric equations and the Pearson's coeffi-
cient was calculated.
4. Results
Table 1 shows permeability and parameter results at each point.
Through the table it is possible to notice the water table behavior. The
highest depths are in the berm (point C) and the lowest in the waterline
(point A). The water table is deeper in Arambaré Norte and Mostardas
and shallowest in Arambaré Sul.
Water table behavior, which is similar on all beaches, was found by
other authors like Oliveira and Vinzon (2003) and Alves et al. (2004):
a water table depth gradient in the backshore direction, being the gradi-
ent intensity dependent on the beach characteristics and dynamics. It is
worth mentioning that water table profiles represent specific and in-
stant data. Water table depth varies according to precipitation, humidity
and other parameters, such as tide and waves, causing changes to beach
morphology and fluid percolation.
Penetration resistance results show low compaction of the substrate
with average penetration of 30 cm. Exceptions occurred due to local

Table 1
Location, permeability and parameter results.

Permeability Parameters

Infiltration Depth (cm) Infiltration width (cm) K (cm/s) Time HH:MM:SS Water table depth (cm) Compaction RP 50 (cm)

Arambaré N 1A 9 23 2.40E-02 00:02:17 20 31

Arambaré N 1B 19 19 1.91E-03 00:07:34 57 43
Arambaré N 1C 17 20 3.25E-02 00:02:06 93 34
Arambaré N 2A 14 24 4.03E-02 00:00:58 20 40
Arambaré N 2B 21 20 2.33E-03 00:09:03 57 37
Arambaré N 2C 20 19 1.44E-02 00:05:12 93 43
Arambaré S 1A 16 20.5 2.63E-03 02:07:00 15 19
Arambaré S 1B 4 38.5 1.02E-03 01:30:11 35 25
Arambaré S 1C 16 15 2.23E-03 01:30:00 62 28
Arambaré S 2A 0 – 3.12E-03 02:00:00 15 34
Arambaré S 2B 15 20 8.54E-03 01:03:25 35 28
Arambaré S 2C 7 18.5 7.37E-04 02:54:39 62 34
Guaíba 1A 13 30 2.34E-02 00:05:24 23 47
Guaíba 1B 23 21 1.37E-02 00:10:13 50 40
Guaíba 1C 29 19 1.90E-02 00:28:04 63 45
Guaíba 2A 23 31 1.02E-02 00:04:29 23 34
Guaíba 2B 19 18 1.42E-02 00:05:45 50 65
Guaíba 2C 29 18 6.21E-03 00:35:14 63 43
Ipanema 1A 14 24 4.49E-03 00:06:54 15 28
Ipanema 1B 18 20 1.32E-02 00:05:44 46 28
Ipanema 1C 20 30 1.33E-03 00:51:56 52 31
Ipanema 2A 14 23 1.24E-04 00:07:35 15 27
Ipanema 2B 18 20 9.22E-03 00:06:08 46 27
Ipanema 2C 28 20.5 5.65E-03 00:23:42 52 31
Mostardas 1A 10 40 1.91E-03 02:24:59 10 32
Mostardas 1B 12 19 1.26E-03 00:26:43 62 31
Mostardas 1C 13.5 14 3.15E-03 01:59:06 76 31
Mostardas 2A 5 12 5.15E-06 03:16:43 10 40
Mostardas 2B 12 19.5 3.58E-03 01:34:25 62 32
Mostardas 2C 10 16 1.94E-03 01:25:21 76 28
 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163 157

Table 2
Granulometric and morphometric sedimentary results.

Granulometric Morphometric

Grain size Classification Median Sorting Classification Skewness Kurtosis Gravel Sphericity Roundness

ANA 0.2347222 Coarse sand -0.0006667 1.170827 Poorly sorted 0.28072617 0.9109187 13.4422976 Medium Angular/subrounded
ANB 0.725555 Coarse sand 0.5883333 0.9873283 Moderately sorted 0.18726963 0.8749109 1.975776 Medium Angular/subrounded
ANC 0.3191667 Coarse sand 0.2463333 0.9465694 Moderately sorted 0.1090234 1.0438827 8.4160212 Medium Angular/subrounded
ASA 0.8395556 Coarse sand 0.6643333 1.4071515 Poorly sorted 0.14153973 0.6140434 7.5405843 Medium Angular/subrounded
ASB 1.8919444 Medium sand 2.3268333 1.0459596 Poorly sorted −0.64038034 2.2361929 2.720277 Medium Subrounded
ASC 0.4025556 Coarse sand −0.045 1.6984848 Poorly sorted 0.30343535 0.5370139 30.4262215 Medium Very angular
GA 0.012 Coarse sand 0.2431667 0.9901351 Moderately sorted −0.31211236 0.9252216 19.2162031 Medium Subangular/subrounded
GB 0.499667 Coarse sand 0.6421667 0.6746199 Moderately well sorted −0.36909611 1.2079379 5.6277392 Medium Subangular/subrounded
GC 0.7498333 Coarse sand 0.8036667 0.6715972 Moderately well sorted −0.05235487 1.3081821 0.7666306 Medium Subrounded
IA 0.0647778 Coarse sand 0.1608333 1.1839735 Poorly sorted −0.0774328 0.6591727 26.1553533 Medium Subangular/subrounded
IB 0.747722 Coarse sand 0.8448333 0.8138813 Moderately sorted −0.2780106 0.9926406 5.7978579 Medium Subrounded
IC 0.3909444 Coarse sand 0.582 1.0376831 Poorly sorted −0.21781425 0.82782 12.7837252 Medium Subrounded
MA 2.204389 Fine sand 2.2888333 0.2776111 Very well sorted −0.38284585 0.8226636 0 High Rounded
MB 2.213889 Fine sand 2.292 0.2800101 Very well sorted −0.27965839 0.8081275 0 High Rounded
MC 2.182222 Fine sand 2.2603333 0.2611061 Very well sorted −0.37202219 0.6351102 0 High Rounded

particularities. The compaction was high in Arambaré Sul, point 1C, due
the traffic of vehicles on the beach.
Granulometric and morphometric results are shown in Table 2 and
the frequency histogram is shown in Fig. 3. Table 2 shows the differ-
ences in each point's sedimentary results and will be used later to dis-
cuss hydraulic conductivity.
The histogram shows that the beaches on the western bank of Patos
Lagoon (Arambaré Norte, Arambaré Sul) and the Guaíba lake beaches
(Guaíba, Ipanema) are typically bimodal, formed by coarse to medium
sand and low sorting, while the eastern bank of the lagoon (Mostardas)
has ocean similarities: fine sand with high sorting. The main difference
between A, B, and C points is the amount of gravel, which is higher in A
and C, and lower in B.
The sedimentary differences between the eastern and western
banks of Patos Lagoon/Guaíba lake banks can be explained by sediment
maturity. The western lagoon bank and the Guaíba banks are next to the
sediment source area (the Precambrian Shield), which has reduced sed-
iment reworking time and its maturity. The eastern bank, on the other
hand, is part of the lagoon's ocean barrier, an area known as Restinga
da Lagoa dos Patos – the Patos Lagoon Beach Ridge. The beach ridge is a

a b

c d

Fig. 3. Berm point histogram. The Y axis shows the Frequency (%) and the X axis shows the Grain size (Φ). a) Arambaré Norte; b) Arambaré Sul; c) Guaíba; d) Ipanema; e) Mostardas.
158 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163

A B
C
Fine sediments

Platform front Holocene Coarse sediments

Rock

Fine sediments

Fig. 4. Stratigraphy. A) Irregularity caused by a rock; B) Impermeable layer formed by a Holocene platform front; C) Lenses (layers of fine and coarse sand interleaved).

multiple barrier complex whose origin dates back to the relative varia- Through the 30 plumes generated in the tests, it was possible to
tions in sea level during the Quaternary period. The reworking process identify four patterns (Fig. 6) of plume infiltration, each related to spe-
through variations in sea level caused high sediment maturity. cific parameters.
The roundness and sphericity results can be split into three groups: Pattern 1 (Fig. 6A,B): Lateral spreading greater than vertical infiltra-
angular/subangular; subangular/subrounded; rounded. The rounded tion. Rectangular plume. Occurrence: 12 tests. Found in waterline and
group occurs at Mostardas, on the eastern bank of the lagoon. The on the Patos Lagoon banks. Pattern associated with shallow water
subrounded/subangular grains occur at Ipanema, Guaíba, and Arambaré table and high humidity.
Norte, on the Guaíba Lake banks and on the western bank of Patos La- Pattern 2 (Fig. 6C,D): Lateral spreading similar to vertical infiltration.
goon. Arambaré Sul beach has shown low rounding and belongs to an Bulb, triangle, or lamp-shaped plume. Occurrence: 10 tests. Found at
angular/subangular group. point C (berm) and point B (beach face) on Patos Lagoon banks. Pattern
The stratigraphy was homogeneous on most beaches, without
layers/strata or irregularities. However, rocks, vegetation, and other
solid structures (Fig. 4A,B) were found at a few points, blocking fluid A B
percolation, and hindering oil behavior prediction. Besides, some
beaches have layers of fine and coarse sand interleaved in the form of
lenses (Fig. 4C). These lenses occur due to wave energy oscillations dur-
ing strong winds or storm surges, and do not have spatial patterns
defined.
The heterogeneous distribution package of sediment sets a barrier to
fluid percolation when the discontinuity is solid (like a rock or an im-
permeable layer of silt). However, when the discontinuity is caused by C D
a biological process (excavation by animals and propagation of roots)
the fluid percolation may be favored because the heterogeneity can pro-
vide a preferred path for the flow.

4.1. Permeability

The oil infiltration depth and width are shown in Table 1. In general,
the lowest infiltration values were found in point A (waterline), while
the highest were founded in point C (berm). This behavior is shown in E F
Fig. 5.
Guaíba, Ipanema, and Arambaré Norte have the highest infiltration
values. These beaches have coarse sand and moderately to poor sorting.
Arambaré Sul and Mostardas presented the lowest infiltration values.
Mostardas has fine sand and high sorting while Arambaré Sul has high
sand compaction, which reduces the amount of pores and hampers
fluid percolation.
G H

Fig. 6. Plume patterns. A & B) Pattern 1: lateral spreading (LS) greater than vertical
infiltration (VI) – Rectangular plume; C & D) Pattern 2: (LS) equal (VI) – plume shaped
triangle or bulb; E & F) Pattern 3: (VI) greater than (LS) – rectangular plume; G & H)
Fig. 5. Infiltration depth (cm) for each point. Pattern 4: Irregular plume – no shape defined.
 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163
Fig. 7. Hydraulic conductivity coefficient (K) for each point.
observed at locations with coarse to medium sand and moderate
sorting.
Pattern 3 (Fig. 6E,F): Vertical infiltration greater than lateral spread-
ing. Rectangular plume. Occurrence: 3 tests. Found at point C and B on
the Guaíba Lake banks. Pattern associated with coarse sand and moder-
ate to poor sorting.
Pattern 4 (Fig. 6G,H): Irregular plume. Occurrence: 4 tests. Found in
places with heterogeneous stratigraphy caused by lenses, rocks or bio-
turbation. These plumes have no pattern, and their occurrence depends
on aspects of the area.
Fig. 8. Scatter plots between K and parameters. The graphics show the polynomial trendline and the corresponding multiple R2 value, calculated with R software.
159
Hydraulic conductivity (K) is shown in Table 1 and Fig. 7. The highest
K values are found in Arambaré Norte, Guaíba, and Ipanema, while the
lowest are in Arambaré Sul and Mostardas.
Fig. 8 shows the scatter plots between hydraulic conductivity and
factors (grain size, median, sorting, gravel %, roundness, sphericity,
compaction and water table depth). The polynomial trendline and the
multiple R2 were calculated using R software through script commands.
The multiple R2 found were 0.457 for grain size, 0.44 for median,
0.248 for skewness, 0.54 for kurtosis, 0.445 for sorting, 0.57 for gravel
%, 0.483 for roundness, 0.389 for sand compaction, 0.344 to sphericity,
0.5899 for the water table and 0.119 for sphericity.
The scatter plots show that hydraulic conductivity has great sedi-
ment variability from coarse to medium size, and a linear response
only occurs in fine sand. As the lagoon beaches consist mainly of coarse
sand and poorly sorted sediment, hydraulic conductivity varies widely,
and does not correspond to a set pattern. The coefficient of variation
(CV) was 114% for hydraulic conductivity, indicating that the standard
deviation is much higher than the average conductivity.
The hydraulic conductivity can be calculated from several different
equations from Ahmed (2011). These equations are divided in two:
those using test sampled parameters (Darcy's Law (Darcy, 1856);
Ródio, 1965; Caputo, 1988) and empirical equations, which use
160 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163
secondary parameters derived from sedimentary analysis (Hazen,
1911; Kotyakhov, 1949; NAVFAC, 1974; Shahabi et al., 1984; Chapuis,
2004).
The spatially variable hydraulic properties of natural soils must be
considered in permeability to find the spatial variability of oil infiltration
and provide the best response possible. According to Chapuis (2012),
empirical equations are important because they can provide more K
data for a lower cost and are less time consuming. To test the efficiency
of these equations, scatter plots were created comparing the Ahmed
(2011) equation results with other equations. The R2 value and the lin-
ear trendline are shown on the plots (Fig. 9).
A significant correlation was found for the Hazen (1911), Caputo
(1988), Ródio (1965) and Darcy's Law (Darcy, 1856) equations.
5. Discussion
Low infiltration values found in the waterline are explained by pore
saturation. At this location, the oil percolates down to the barrier formed
by the water table and from that depth it spreads laterally. As the water
table depth increases, the number of pores also increases, which con-
tributes to fluid percolation. The maximum depth reached by the oil at
points B and C was limited by the amount of oil used (750 ml). During
an oil spill, where the amount of fluid is high, the oil will infiltrate all
the way down to the water table depth. According to Hayes and
Michel (1999), the depth of oil penetration is limited by spill volume,
by water table depth, and by the presence of impermeable layers.
However, despite blocking vertical infiltration, the water table can
transport the oil to backshore, through tide oscillations and subsequent
variations of water table level. The oil, initially concentrated on the
beach face, moves to upper sediments as the tide floods. With the ebb
tide, the water table releases oil on the beach face while residual oil is
trapped by upper sediments (McLaren, 1985). Furthermore, soluble hy-
drocarbon components (BTEX, MTBE) dissolve in the water table, con-
taminating the water (Halmemies et al., 2003).
According to Petrobras (2006) and the Environment Ministry
(MMA) (2007), the studied beaches (fine to medium sand beaches,
both exposed and protected) have moderate permeability, with a max-
imum penetration depth of 25 cm. These results, however, indicate that,
even using a small and limited amount of oil (750 ml), the fluid perco-
lation was higher than expected. With a higher amount of oil (as in a
spill accident), the fluid would infiltrate all the way down to the water
table surface, which in this case was between 60 and 95 cm.
The methodology to determine oil sensitivity (Petrobras, 2006;
MMA, 2007) uses the contaminated substrate volume and ease of
cleaning to measure environmental sensitivity, and, to this end, estab-
lish expected behaviors for the contaminant (maximum percolation
depth). But these results show that predetermined values cannot be
generalized, since the local characteristics of this beach may modify its
substrate permeability, and fail to match pre-existing classifications.
Therefore, specific characteristics of oil permeability in each place
must be studied and inserted in existing methodology.
The plume patterns found relate to intermediate, unweathered oil
(API 33). Since permeability depends on fluid characteristics, oils with
different viscosities will infiltrate at different rates. Another important
point is the weathering process. After being released in the environ-
ment, the oil undergoes a weathering process, which includes evapora-
tion, oxidation, emulsification, degradation, etc. This alters its chemical
composition and physical characteristics over time, modifying the
contaminant's behavior. Weathered oil, according to tests made by
Gandra (2005), has lower infiltration values due its high viscosity.
The scatter plots show that hydraulic conductivity (K) has great var-
iability in coarse to medium size sand, whereas a linear correlation only
occurs with fine sediment. As the lagoon beaches are typically bimodal,
K decreases as grain size decreases. This result was expected, because
the larger the grain size, the greater the amount of pores and higher
the percolation.
The negative correlation between K and roundness/sphericity is ex-
plained by the grain packing: well rounded grains packs with a mini-
mum amount of pores, decreasing infiltration. With the increase of
angularity, the porous space and permeability rise. Roundness is related
to maturity, and, consequently, depends on geological inheritance and
distance of the source area.
The positive correlation to penetration resistance depth (RP50)
shows that K decreases with the increase of sand compaction. Vieira et
al. (2004) studied the impact of vehicle traffic on Cassino beach, in Rio
Grande do Sul, Brazil, and found that permeability was significantly
lower in high compaction areas.
The positive correlation for the water table depth indicates that
the deeper the water table, the higher the permeability. This is a
usual behavior because, as previously noted, water causes pore satu-
ration creating a barrier to fluid percolation. However, with spatial
and temporal variations of the water table, fluid behavior may
change. According to CETESB (2007), this barrier is clear in seasonal
terms, since one location can be more or less vulnerable to the spatial
extent of the contamination, according to the periodic changes in
water levels. In addition, variations in water table depth cause
beach morphology changes that will affect leakage, as a secondary
effect. According to Turner and Nielsen (1997), rapid water table
elevation rates cause sediment remobilization on the beach and, as
a result, fluctuation levels can be considered a significant factor in
changing the arrangement of sediments in the swash zone.
Rodriguez et al. (2015) found that the structural arrangement,
which is different for each soil, may cause high variation in coeffi-
cient permeability in samples considered homogeneous.
The positive linear correlation of sorting and the gravel %, although
not too significant, is an important result because the data contradict
past studies from Beard and Weyl (1973), Selley (1998) and Stern
(2007). According to these authors, soils poorly sorted allow intergran-
ular interstices to be filled by fine sediments (very fine sand, clay and
silt), blocking the pores and, consequently, reducing porosity and per-
meability. The data obtained here, however, show that permeability in-
creases with the decrease in sorting – in other words, poorly sorted
samples (bimodal) with low or zero concentration of fine sediment,
tend to have higher permeability values than well sorted samples.
This feature was also noticed by Pereira and Calliari (2005) and Costi
and Calliari (2008) at the Concheiros ocean beach on the Rio Grande do
Sul coast. According to the authors, Concheiros has the highest perme-
ability of the coast (K 0.0098 cm/s and infiltration of 20 cm), which is as-
sociated with the bimodal characteristics of the area (beach formed by
biodetrical gravel and variable granulometric sand). The oceanic values,
however, are one order of magnitude lower than the values found for la-
goon beaches (highest value 0.04). The lower permeability of ocean
beaches can be explained by the high mobility of the beach face caused
by wave energy, which moves the sediment structure, affecting the
amount of pores. The sediment structure, conditioned by the arrange-
ment of particles, defines the existing gaps that are possible paths of
seepage flow (Rodriguez et al., 2015). Therefore, high energy waves at
ocean beaches alter the particles disposal at the beach face, changing
permeability.
The relations between permeability and parameters (grain size, me-
dian, roundness, sphericity, compaction and water table depth) are in
accordance with literature and indicate that permeability in the Patos-
Guaíba system is governed, to a great extent, by the same parameters
as the ocean beaches studied. The sorting and gravel % results, on the
other hand, show that the bimodal characteristics of lagoon beaches
cause a permeability increase in systems ruled by fluvial forcing. Pryor
(1973) reports that porosity in fluvial sand bars may exhibit an in-
creased porosity and a decreased sorting due to the different type of
grain packing which occurs in these areas. The author also says that
the presence of layers and stratification are important porosity factors
in such environments. An increased porosity can be accompanied by
an increased permeability.
 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163 161

Fig. 9. Scatter plots between the results obtained with Phillip theory (Ahmed, 2011) and other equations.

An important characteristic of the flow at lagoon beaches is, therefore, frequency histogram, but also the gravel percentage (which favors infiltra-
the standard deviation or the sorting of the sample. The definition of per- tion), the clay percentage (which decreases infiltration) and the grain
meability in these environments should consider not just the granulometric arrangement and structure (presence of layers and lenses).
162 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163
5.1. Permeability and oil sensitivity
Permeability in lagoon systems is governed not just by grain size and
median as at ocean beaches, but also by the system's bimodal character-
istics and the sediment packing. Bimodal characteristics favor the flow
at locations where the percentage of gravel is high and of fines is low.
A heterogeneous vertical packing (formed by interlayer lenses) can
modify fluid seepage. The presence of a lens formed by coarse grains ac-
celerates the flow's velocity in this layer, causing the oil to deposit at
lower depths than expected.
Both characteristics worsen an oil spill because they increase sub-
strate permeability, causing greater contamination by increasing the
volume of contaminated sand, environmental impacts, and difficulty
in the cleaning process. However, even if they change oil behavior,
these characteristics are not included in the oil sensitivity methodology
for a couple of reasons. First, the methodology is based on an ocean en-
vironment, where such characteristics are not relevant. Second, the sed-
imentary characteristics of an area are established visually by the
researcher on the field, making the inclusion of the mentioned charac-
teristics difficult.
Two parameters that are merely theoretically considered in oil sen-
sitivity, but are important for response plans, are sand compaction and
water table depth. Water table depth is considered maximum oil infil-
tration depth (McLaren, 1985; Oliveira and Vinzon, 2003; Cetesb,
2007). Due to spatial variability and seasonal behavior, water table
depth at the contaminated location is crucial information. During dry
months, the water table is deeper and the contaminated volume is
higher than in rainy months. Soil humidity is connected to water table
depth. As said before, the higher the humidity, the higher the pores sat-
uration and the lower the water table depth. According to Halmemies et
al. (2003), humidity has a great effect on oil seepage. High sand compac-
tion reduces the amount of pores, and reduces permeability. A high
compaction substrate may be used as deflation area for oil slicks, as
the oil will be retained on the surface.
The findings have permitted creating a guideline with field proce-
dures and laboratory techniques that should be adopted before the
final oil sensitivity ranking:
• Sediments sampling for granulometric analysis (including grain size,
sorting and percentage of gravel and clay);
• Stratigraphic cut to determine the presence of lenses and hard struc-
tures such as rocks, impermeable layers, etc;
• Sand compaction measures;
• Measuring water table depth and its behavior though seasons;
• Permeability prediction by the best empiric equation for the studied
area.
This guideline will make oil behavior and contaminant residence
time easier to predict, ensuring that the final oil sensitivity ranking of
a beach contemplates the actual sensitivity of the environment. More-
over, the adoption of laboratory techniques will reduce generalization
of oil behavior, since the specific aspects of the beach will be considered
during the ranking.
6. Conclusions
Oil permeability on the beaches of the “Patos-Guaíba” system has
shown hydraulic conductivity values (K) between 0.00124 and
0.04 cm/s and infiltration depth between 5 and 30 cm. The K values fit
in the values expected for the sand granulometric class (0.001 to
0.0001 cm/s). The equations that use measured data to calculate K
showed similar and satisfactory results, while the empirical equations
(which estimate K from granulometric parameters) showed low corre-
lation. Only the Hazen (1911) equation presented satisfactory results, as
it is the equation indicated to predict lagoon permeability.
The highest permeability values occurred on beaches with angular
coarse to medium size grains, poorly to moderately sorting, low sand
compaction and a deep water table. These characteristics indicate that
permeability in lagoon systems is determined, mainly, by grain size,
roundness and sand compaction, and is increased by the bimodal distri-
bution of the grains (poorly sorted samples). The increase in permeabil-
ity with the decrease in sorting diverges from literature. Normally,
poorly sorted substrates show low K values, because the intergranular
interstices are filled with fine grains that block the pores, reducing per-
meability. In the Patos-Guaíba system, however, bimodal beaches had
higher permeability. This relation is explained by the absence of fine
grains (silt and clay), the high concentration of gravel and the low re-
mobilization of the sediments in the lagoon system. Low remobilization
allows the grains to be deposited in intercalated layers of fine and coarse
sediments, creating a vertical heterogeneous package, which alters the
fluid flow.
Other important factors are geological inheritance and grain maturi-
ty. On ocean beaches and at Mostardas (on the lagoon barrier, in an area
known as Restinga da Lagoa dos Patos – the Patos Lagoon beach ridge)
the wave energy during sea level variations during the Quaternary Peri-
od caused sediment reworking, which led to grains with a high maturity
(spherical and rounded grains, plus well sorted sediment) which in turn
led to a reduced permeability. Sediment on the western bank of Patos
Lagoon and on the banks of Guaíba Lake beaches, however, originated
from the highlands. The proximity between the source area and the de-
posit area has reduced sediment reworking time, causing low maturity
on these beaches (angular grains and poorly sorted sediment) and this
favors fluid flow.
The official method to define oil sensitivity is based on behavior and
permanence of the contaminant in the substrate, assuming typical
values for the permeability. According to the methodology, lagoon
beaches (with coarse, medium, or fine sand) have a maximum oil perco-
lation of 25 cm in depth. These results, however, show that, even using a
limited amount of contaminant, infiltration was higher than the rate set
by the method. This difference is due to a number of factors. First, the
methodology is based on an ocean environment (sediments with high
maturity, great mobility, and well to moderately sorted) but is applied
to river and lagoon environments (poorly sorted sediments, with low
maturity and mobility). Second, the methodology does not consider
the sample's standard deviation. Grain size is visually defined by the re-
searcher on the field, and the substrate is ranked as medium, fine, or
coarse sand. However, lagoon beaches are typically bimodal and, be-
cause of this, the researcher's ranking is not consistent with the actual
characteristics of the environment. Third, the vertical sediment distribu-
tion is not considered. At lagoons, intercalated layers of fine and coarse
sediment are common due to variations in the environment energy and
low system mobility. This heterogeneous package changes the oil flow,
altering the expected behavior of the contaminant.
In order to contemplate lagoon permeability in oil sensitivity of an
area, the sedimentary characteristics of the environment, including
grain size, sorting, gravel and fine percentages, package (layers and
hard structures), sand compaction and water table depth must be
included at the final ranking of sensitivity value.
Acknowledgements
We gratefully acknowledge Agência Nacional do Petróleo (ANP) and
its Human Resources Program (PRH) for the financial support offered
during this work, REFAP for offering the oil used in the tests and FEPAM
for giving the authorization license.
References
Ahmed, F., 2011. The “dirt” on soils, water, and infiltration. Stormwater U.S. Education
Program: Bioretention Systems. St. Anthony Falls Laboratory. University of Minnesota
– College of Science, USA.
 E.B. Oliveira, J.L. Nicolodi / Marine Pollution Bulletin 115 (2017) 154–163
Ahmed, F., Gulliver, J.S., Nieber, J.L., 2011. Performance of low impact development prac-
tices on stormwater pollutant load abatement. Project Report No. 560. St. Anthony
Falls Laboratory. University of Minnesota. Minneapolis, Minnesota (33 p).
Alves, E.S., Rodrigues, S.A., Pezzuto, P.R., Shimizu, G.Y., 2004. Caracterização ambiental da
porção subaérea de praias arenosas expostas com diferentes características
morfodinâmicas no litoral centro-norte de Santa Catarina, Brasil. Notas Técnicas. 8.
FACIMAR, pp. 1–24.
Beard, D.C., Weyl, P.K., 1973. Influence of texture on porosity and permeability of uncon-
solidated sand. Am. Assoc. Pet. Geol. Bull. 57 (2), 349–369.
Bigarella, J.J., Hartkopf, C.C., Sobanski, A., Trevisan, N., 1955. Textura superficial dos grãos
em areias e arenitos. Arq. Biol. Tecnol. 10 (11), 253–275.
Calliari, L.J., Tozzi, H.A.M., Klein, A.H.F., 1998. Beach morphology and coastline erosion as-
sociated with storm surges in Southern Brazil – Rio Grande to Chuí. Acad. Bras. Cienc.
70 (2), 231–247.
Caputo, H.P., 1988. Mecânica dos solos e suas aplicações. sixth ed. 1. Editora LTC, Rio de
Janeiro (244 p).
Castanedo, S., Juanes, J.A., Medina, R., Puente, A.L., Fernandez, R., Olabarrieta, M., Pombo,
C., 2009. Oil spill vulnerability assessment integrating physical, biological and socio-
economical aspects: application to the Cantabrian coast (Bay of Biscay, Spain).
J. Environ. Manag. 91, 149–159.
CETESB, 2007. Ambientes Costeiros contaminados por óleo: procedimentos de limpeza.
Manual de orientação. Governo do Estado de São Paulo. Secretaria do Meio Ambiente.
Companhia de Tecnologia de Saneamento Ambiental, São Paulo, p. 120.
Chapuis, R.P., 2004. Predicting the saturated hydraulic conductivity of sand and gravel
using effective diameter and void ratio. Can. Geotech. J. 41, 787–795.
Chapuis, R.P., 2012. Predicting the saturated hydraulic conductivity of soils: a review. Bull.
Eng. Geol. Environ. 71, 401–434.
Costi, J., Calliari, L.J., 2008. Estimativa do volume sedimentar potencialmente contaminado
em casos de derrame de óleo em condições de verão para praias arenosas do Rio
Grande do Sul. RIO OIL & GAS EXPO AND CONFERENCE. Anais… Rio de Janeiro (8 p).
Darcy, H., 1856. Les fontaines publiques de la ville de Dijon. Victor Dalmont, Paris.
Fataal, P., Maanam, M., Tillier, I., Rollo, N., Robin, M., Pottier, P., 2010. Coastal vulnerability
to oil spill pollution: the case of Noirmutier Island (France). J. Coast. Res. 26 (5),
879–887.
Gandra, M.S., 2005. Efeitos do petróleo sobre a associação de macroinvertebrados
bentônicos de praias arenosas do extremo sul do Brasil. Dissertação (Mestrado) –
Programa de Pós Graduação em Oceanografia Física, Química e Geológica. Instituto
de Oceanografia. Universidade Federal do Rio Grande/FURG, Rio Grande, RS (P. 86).
Gilbert, E.R., Camargo, M.G., Sandrini-Neto, L., 2014. rysgran: Grain size analysis, textural
classifications and distribution of unconsolidated sediments. R package version 2.1.0 .
https://CRAN.R-project.org/package=rysgran.
Halmemies, S., Gröndahl, S., Nenonen, K., Tuhkanen, T., 2003. Estimation of the time pe-
riods and processes for penetration of selected spilled oils and fuels in different
soils in the laboratory. Spill Sci. Technol. Bull. 8 (5–6), 451–465.
Hayes, M.O., Michel, J., 1999. Factors determining the long-term persistence of Exxon
Valdez oil in gravel beaches. Mar. Pollut. Bull. 38 (2), 92–101.
Hazen, A., 1911. Discussion of “Dams on sand formations” by A.C. Koenig. Trans. Am. Soc.
Civ. Eng. 73, 199–203.
Kotyakhov, F.I., 1949. Relationship Between Physical Parameters of Sandstones. 12. Neft
Khoz, pp. 29–32.
Krumbein, W.C., 1941. Measurement and geological significance of shape and roundness
of sedimentary particles. J. Sediment. Petrol. 11 (2), 64–72.
Machado, C.S., 2007. Balanço hídrico superficial da Lagoa dos Patos e sua influência sobre
a água subterrânea. (Master's Dissertation – Graduate Program in Ocean Engineer-
ing). Federal University of Rio Grande/FURG, Rio Grande, RS, p. 141.
Marinho, C., 2015. Integração de parâmetros geomorfológicos e biológicos no
desenvolvimento de uma proposta de Índice de Sensibilidade do Litoral (ISL).
(Master's Dissertation – Graduate Program in Ocean Engineering). Federal University
of Rio Grande/FURG, Rio Grande, RS (P. 114).
Marques, V.C., Nicolodi, J.L., 2015. Análise sedimentológica na definição de sensibilidade
ao óleo na Lagoa dos Patos e Lago Guaíba, RS.
Martins, I.R., Villwock, J.A., Martins, L.R., Benvenuti, C.E., 1989. The Patos Lagoon Estuarine
Ecosystem (RS, Brasil). Research 3, 5–44 (Porto Alegre).
McLaren, P., 1985. Behavior of diesel fuel on a high energy beach. Mar. Pollut. Bull. 16 (5),
191–196.
163
Ministry of the Environment – MMA, 2007. Especificações e normas técnicas para
elaboração de cartas de sensibilidade ambiental para derramamentos de óleo.
Ministério do Meio Ambiente, Secretaria de Mudanças Climáticas e Qualidade
Ambiental, Gerência de Qualidade (107 p).
Moe, K.A., Skeie, G.M., Brude, O.W., Loyas, S.M., Nedrebo, M., Weslawiski, I.M., 2000. The
Svalbard interdital zone: a concept for the use of GIS in applied oil sensitivity, vulner-
ability and impact analysis. Spill Sci. Technol. Bull. 6, 187–206.
Motta, V., 1969. Relatório-diagnóstico sobre a melhoria e o aprofundamento do acesso
pela Barra de Rio Grande. Instituto de Pesquisas Hidráulicas, UFRGS, Porto Alegre.
NAVFAC, 1974. Soil mechanics, foundations, and earth structures. Naval Facilities Engi-
neering Command design manual DM7. U.S. Government Printing Office, Washing-
ton, D.C.
Nestingen, R.S., 2007. The Comparison of Infiltration Devices and Modification of the Phil-
ip-Dunne Permeameter for the Assessment of Rain Gardens. (MSc Dissertation –
Graduate School). University of Minnesota, Minneapolis (58 p).
NOAA, 1992. An Introduction to Coastal Habitats and Biological Resources for Oil Spill Re-
sponse. National Oceanic and Atmospheric Administration, Seattle, Washington.
Oliveira, F.A.M., Vinzon, S.B., 2003. Dinâmica do escoamento na face da praia. IX
CONGRESSO DA ASSOAÇÃO BRASILEIRA DE ESTUDOS DO QUATERNÁRIO. Anais… Re-
cife (P.3).
Owens, E.H., 1988. Estimating and quantifying oil contamination on the shoreline. Mar.
Pollut. Bull. 18 (3), 110–118.
Pereira, P.S., Calliari, L.J., 2005. Permeabilidade das praias oceânicas do Rio Grande do Sul
(RS) em relação a eventuais derrames de óleo. III CONGRESSO BRASILEIRO DE P&D
EM PETRÓLEO E GÁS. Anais…Salvador (p.6).
PETROBRÁS, 2006. Manual de avaliação de óleo em linha de costa. Centro de Pesquisas e
Desenvolvimento Leopoldo A. Miguez de Mello (CENPES). Gerência de Avaliação de
Impacto e Monitoramento Ambiental, Rio de Janeiro (P. 90).
Philip, J.R., 1993. Approximate analysis of falling-head lined borehole permeameter.
Water Resour. Res. 29, 3763–3768.
Pryor, W.A., 1973. Permeability-porosity patterns and variations in some Holocene sand
bodies. AAPG Bull. 57 (1), 162–189.
R Core Team, 2016. R: A language and environment for statistical computing. R Founda-
tion for Statistical Computing, Vienna, Austria. https:www.R-project.org/.
Rittenhouse, G., 1943. A visual method of estimating two-dimensional sphericity.
J. Sediment. Petrol. 13 (2), 79–81.
Ródio, S.A., 1965. Measures des permeabilities en place. S·I: Cooperation Technique
A.S.T.E.F., Stage de Perfectionnement, Geoconseil.
Rodriguez, T.T., Weiss, L.A., Teixeira, R.S., Branco, C.J.M.C., 2015. Permeabilidade de solo
laterítico por diferentes métodos. Semina: Ciências Exatas e Tecnológicas 36 (2),
17–32.
Rowland, D.K., Lindley, A.P., Hall, G.P., Rossal, M.J., Wilson, D.R., Benham, D.G., Harrison,
A.F., Daniels, R.E., 2000. Effects of beach sand properties, temperature and rainfall
on the degradation rates of oil in buried oil/beach sand mixtures. Environ. Pollut.
109, 109–118.
Selley, R.C., 1998. Elements of Petroleum Geology. 2ª edição. Academic Press, London (470 p).
Shahabi, A.A., Das, B.M., Tarquin, A.J., 1984. An empiral relation for coefficient of perme-
ability of sand. Fourth Australia-New Zealand Conference on Geomechanics Perth.
Anais… Australia, pp. 54–57.
Stern, A.G., 2007. Parâmetros texturais de sedimentos para elaboração de cartas de
sensibilidade ambiental ao derramamento de óleo Cartas SAO. (Master's Dissertation
– Institute of Geosciences). University of São Paulo/USP, São Paulo (103 p).
Toldo Jr., E.E., 1994. Sedimentação, predição do padrão de ondas e dinâmica sedimentar
da antepraia e zona de surfe do sistema lagunar da Lagoa dos Patos. (Doctoral thesis.
Graduate Program in Geosciences). Institute of Geosciences – UFRGS (143 p).
Tomazelli, L.J., Dillenburg, S.R., Villwock, J.A., 2000. Late quaternary geological history of
Rio Grande do Sul coastal plain, Southern Brazil. Rev. Bras. Geosci. 30 (3), 474–476.
Tsouk, E., Amir, S., Goldsmith, V., 1985. Natural self-cleaning of oil pollute beaches by
waves. Mar. Pollut. Bull. 16 (1), 11–19.
Turner, I.L., Nielsen, P., 1997. Rapid water table fluctuations within the beach face: impli-
cations for swash zone sediment mobility? Coast. Eng. 32, 45–59.
Vieira, H., Calliari, L.J., Oliveira, G.P., 2004. O estudo do impacto da circulação de veículos
em praias arenosas através de parâmetros físicos: um estudo de caso. Engevista 6 (3),
54–63.