Spatial variation of metals and phosphorus in sediments of a river influenced by urbanization Variação espacial de metais e fósforo em sedimentos de rio influenciado pela urbanização

The suspended sediments are a special compartment that characterizes the effects of land use in the watershed. Such compartment is able to trap organic compounds as well as metals in general. Spatial variation of concentrations can be accounted for the land use, soil composition and pollution. In this work, we aimed to assess the spatial variation of metals and phosphorus concentration presented in suspended sediments collected in a river under urban occupation. Using time-integrated suspended sediment samplers, samples were collectec in three sites covering different levels of occupation: Almirante Tamandaré (P1) and two in Curitiba city (Tingui Park -P2 and Barigui Park P3). The sampler was settled to be 20 cm below the water surface and it was removed after 30 days. Twelve samples were collected to examining the following elements Al, As, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, La, Mg, Mn, Na, Nb, Ni, P, Pb, Sc, Sn, Sr, Th, Ti, V, Y, Zn and Zr by ICP-OES. The first site is characterized with only 5.73% of urban area; meanwhile the site P3 has the highest urban area 54.06%. Among the twenty seven elements, nineteen showed a positive variation in concentration between collected sites, giving indication that an enrichment of metals along the river. The following elements showed positive variation: P (151.20%), La (117.98%), Cd (93.33%), Cr (71.43%), Na (68.75%), Ni (68.07%), Y (66.67%), Th (63.64%) and Zn (56.77%). Considering local reference values it is concluded that Cd, Cr, Ni, P, Y and Zn appear associated with urbanization increase in the basin; Co, Sc, Sn, Cu and Mn appear possibly associated with urbanization; Ba, La, Na, Pb and Sr show inconclusive behavior in relation with urban influence ; Al, Fe, Nb, Th, Ti, As, Ca, K, Mg, Zr and Va demonstrate not be associated with urbanization.


INTRODUCTION
Sediments, in general, have the capacity to adsorb and bind metals, organic compounds and nutrients (MIGUEL et al., 2005;TAYLOR;OWENS, 2009).Thus, sediments can be considered as the final destination of some pollutants as well as sources of pollution that can interfere in the biota (BURTON JUNIOR, 2002).Metals, in general, in suspended sediments are often used as indicative of land use changes, for instance increase of urbanization level (HOROWITZ;ELRICK;SMITH, 2008).
Recently, it has been demonstrated that the presence and distribution of metals in suspended sediments can be complex and does not follow a constant variation to associate with land use and occupation.Some authors found no significant variation in metals concentrations (ZGHEIB; MOILLERON; CHEBBO, 2011;GALLO et al., 2013), but others found significant variation (VALTANEN;SILLANPAA;SETALA, 2014;MCKEE;GILBREATH, 2015).The main potential sources of metals in sediments, especially in urbanized areas, are paved and unpaved streets, residential and industrial areas, also sewage discharges (SILVA; SIQUEIRA; LEANDRO, 2015;FRANZ et al., 2014;MIGUEL et al., 2005).
However, many metals and inorganic compounds are natural present in soil constitution of the watershed, thus sometimes is not an easy task to distinguish between polluted and unpolluted condition based on metal distribution (BURTON JUNIOR, 2002).
Researchers had considered other groups of metals, although with lower abundance.Those metals are Ba, Na, P, V and Zr.The natural occurrence of these metals is from 100 to 1,000 ppm.According to Mineropar (2005) La, Nb, Sc, Sn, Sr, Th and Y also are found, but in lower concentrations (lower than 100 ppm).
The aim of this work was to evaluate the spatial variation of metals and phosphorus in suspended sediments of a river influenced by urbanization.

Study area and sampling sites
The study area comprehended the upper part of Barigui River watershed (25º13'24" and 25º38'23" South;49º15'00" and 49º22'29" West).This part has 131 km 2 of drainage area.The annual mean precipitation varied from 1,400 to 1,600 mm (IAPAR, 2013), well distributed along the year.The mean value of river flow, considering the last 30 years, was 5.64 m 3 s -1 , being the maximum 111.0 in January 1995, while the minimum was 0.25 m 3 s -1 in 1982.
In this work, it was selected three sites for sampling in the watershed.The monitoring plan considered suspended sediments sampling every month.(Table 1 and Figure 1).
The discharge of solids into river were 10 to 20 ton day -1 for P1, 20 to 60 ton day -1 for P2 and 10 to 50 ton day -1 for P3.It is highlighted that site P3 retain most of solids due the lake designed for such purpose, in contrast to other sites (CHELLA; FERNANDES; FERMIANO, 2005).

Land use in the study area
The land use was obtained examining the geographic information system (GIS) provided by Instituto das Águas do Parana.This GIS was designed in scale 1:20,000 based on aerophotometric data in 1:30,000 scale.The rural areas on GIS was achieved by restitution of aerophotometric datas for rural areas in lower scale (1:10,000) and thematic mapping (1:20,000) (SUDERHSA, 2000).
Unfortunately, information about land use was quite old for this study, thus it was considered the population growth and urbanization level of the watershed (IPPUC, 2012;IPARDES, 2016).
According to the analysis of GIS and recent datas about population growth and urbanization index, it was detected nine classes of land use: urban, industrial, mining, grazing, reforesting, agriculture, soil without coverage and others (Figure 1).
The sampler consists in a PVC tube with narrow entrance and narrow output (4 mm).The body of sampler has 100 mm as diameter and 100 cm length.Briefly, the reduction of speed flow allows the sedimentation of suspended particles inside the tube.
The samplers were settled approximately 20 cm below the water surface and 1.5 m from the river bank, according the procedures described in the literature (MCDONALD; LAMOUREUX; WARBURTON, 2010; PHILLIPS; RUSSELL;  WALLING, 2000).The width of river and the respective sites were: P1 -9 m, P2 -8 m and P3 -25 m.
The sampler was fixed in a steel support that was then fixed in the riverbed.The sampler were left during 30 days and collected monthly.Before analysis, samples were kept at 4 °C.

Determination of metals and phosphorus concentrations in samples
The samples containing the sediments were filtered (14 µm) under vacuum and then, dried at 40 °C for 24 h.Before further analysis, samples were crushed and pulverized for chemical analysis.For granulometric composition it was considered the fine fraction (< 63 µm) (POLETO et al., 2009;CARTER et al., 2003;DAVIS;FOX, 2009).
The granulometric composition was obtained by a laser granulometer.While the quantification of metals and P was achieved by inductively coupled plasma optical emission spectrometry (ICP-OES).The following elements were analyzed: Al, As, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, La, Mg, Mn, Na, Nb, Ni, P, Pb, Sc, Sn, Sr, Th, Ti, V, Y, Zn and Zr.

Statistical analysis
The principal component analysis (PCA) was used to explain the variance-covariance structure of a set of metals and P through linear combinations.

Land use in the study area
The site P1 comprehend 5.73% of urbanized areas.In this area, even is classified as urban, still exists natural vegetation (86.19%), however, most of them are represented by grazing and vegetation for commercial purpose.In this segment, it is found the largest extension of Barigui River inside the metropolitan area of Curitiba.The P2 site represents 22.30% of land occupied by urbanized area.It is highlighted that natural vegetation corresponds to 57.07%.It can be observed many inputs and discharges from Almirante Tamandaré and some neighborhoods of Curitiba in this segment.While P3 contains 54.05% of urbanized area.In this area, 41.15% is covered by natural vegetation and fields (Table 2).
It is interesting that the urbanization level increase from P1 to P3.The P2 site is the intermediate in despite of urbanization.

Concentration of phosphorus and metals
In order to better understand the great number of metals analyzed in suspended sediments, the metals and P were grouped according the abundance of each element, thus in Table 3 is showed the groups formed and level of concentration found.
The group formed with high concentration presented the following order of concentration Al>Fe>Ca>Mg>K>Ti>P>Na.While the group classified as low concentration presented the distribution Mn>Ba>Zn>V>Zr.Trace elements has Cr in lower concentration and Cd in higher concentration, the sequence is Cr>Cu>La>Sr> Ni>Y>Pb>Co>Sc>Nb>Th>Sn>As>Cd (Table 4).

Metal concentration and land use
Of 27 metals examined, 19 presented positive variability between sites P1 and P3 (Figure 2).This numbers can be resulted from land use as urbanization process (Table 2).In contrast, As, Ca, Cu, K, Mg, Mn, V and Zr showed negative variability.A substantial and negative variation was found for Ca and Mg, 66.11% and 56.19% respectively.
The principal component analysis (PCA) showed that the factor 1 explains 61.20% and factor 2 explains 15.58%, therefore 76.78% of data can be explained (Figure 3).It was observed that factor 1 has a positive correlation with Ca, K, Mg, Zr and a    The metals As, Mn, Cu and V are in the middle part of Figure 3.

Concentration of metals versus background values
The establishment of values that can be used as reference is still under debate (BURTON JUNIOR, 2002).Most of times the comparison are made using international values, although those values do not represent the similar geologic composition of soil, thus misinterpretation is possible.

Figure 1 .
Figure 1.Study area, sampling sites and urban land use.

Figure 2 .
Figure 2. Percentual variability of metals and phosphorus from P1 to P3.
found with Al, Ba, Cd, Co, Cr, Fe, La, Na, Nb, Ni, P, Pb, Sc, Sr, Ti, Y and Zn.The factor 2 did not show any correlation, neither positive nor negative (Figure3).The factor 1 has a strong and positive correlation with Ca, K, Mg and Zr, such metals were grouped in the right side of plot.In fact, those metals showed decrease in concentration pattern from P1 to P3. Lowest level of these metals are associated with increase of urbanization.Such behavior of Ca and Mg was already demonstrated byMineropar (2001).In contrast, component 1 presented a negative correlation with Al, Ba, Cd, Co, Cr, Fe, La, Na, Nb, Ni, P, Pb, Sc, Sr, Ti, Y and Zn.Those metals are concentrated in the left side of Figure3.However, in the figure Sn and Th showed positive variation.Most of results found here agree with soil composition determined byMineropar (2001), thus represent the contribution of soil by erosion, for instance.The elements Al, Ba, La, Na, Ni and Ti did not completely agree with soil composition.

Figure 3 .
Principal component analysis for metals and phosphorus.

Table 1 .
Sampling sites and their characteristics.
* Above the sea mean level.Macedo Neto et al.

Table 2 .
Percentages of land use in the three studied segments.

Table 3 .
Groups of metals and P according the concentration.

Table 5 .
Concentration of metal in suspended sediments and values adopted as background reference for level of pollution.