MAPPING OF HEAVY METAL CONTAMINATION IN ALLUVIAL SOILS OF THE MIDDLE NILE DELTA OF EGYPT

Areas contaminated by heavy metals were identified in the El-Gharbia Governorate (District) of Egypt. Identification used remote sensing and Geographical Information Systems (GIS) as the main research tools. Digital Elevation Models (DEM), Landsat 8 and contour maps were used to map physiographic units. Nine soil profiles were sampled in different physiographic units in the study area. Geochemical analysis of the 33 soil samples was conducted using X-ray fluorescence spectrometry (XRF). Vanadium (V), nickel (Ni), chromium (Cr), copper (Cu) and zinc (Zn) concentrations were measured. V, Ni and Cr concentrations exceeded recommended safety values in all horizons of the soil profiles, while Cu had a variable distribution. Zn concentrations slightly exceeded recommended concentration limits. Concentrations were mapped in each physiographic unit using the inverse distance weighted (IDW) function of Arc-GIS 10.1 software. Pollution levels were closely associated with industry and urban areas.


Introduction
Sustainable agriculture is mainly related to environmental, agronomic, ethical and socio -economic issues (Abd Elgawad et al. 2007). One aspect of sustainability is accumulation of heavy metals in soils, which may cause serious problems, if certain levels are exceeded. In recent years, much concern has been articulated over problems of soil contamination with heavy metals. These metals can accumulate in plants and animals and then in humans through the food chain (Govil et al. 2001;Lu, Bai;2010;Romic, Romic, 2003). Thus, heavy metals may damage human health and the environment (Jankaite, Vasarevičius, 2005). The Nile Delta (area ~20,000 km 2 ) represents only 2.3% of the area of Egypt, but it has ~46% of the total cultivated area (55,040 km 2 ) and accommodates ~45% of Egypt's population (Fanos 2002), with densities ≤1600 inhabitants per km 2 (Zeydan 2005). On the Nile Delta ~63% land is agricultural, due to suitable soil properties and the presence of irrigation systems (Dawoud 2004). The River Nile divides into two branches, the Rosetta and Damietta, and the Delta region is located between them (Dumont, 2009). The Nile Delta (area 404,686 ha) depends on drainage water for irrigation (Abu Khatita 2011). There are three major layers in the middle Delta aquifer (Atwia et al. 2006). The uppermost layer is composed of clay deposits, the second layer is formed from sandy clay deposits and the third layer is composed of saturated sand and gravel. Thus, the thin clay layers and presence of sandy clay lenses facilitate percolation of sewage water to the aquifer. Many activities, including agricultural development and industrial activi ties and inadequate rural sanitation, have impacts on eutrophication and contamination status, ecological value and environmental conditions in the Nile Delta (Zeydan 2005). Heavy metal contamination of soil may present risks and hazards to humans and the ecosystems through: direct consumption or contact with contaminated soil, the food chain (soil-plant-human or soil-plant-animal-human), drinking of contaminated ground-water, decreased food quality (safety and marketability) via phytotoxicity, reduction in land usability for agricultural production causing food insecurity, and land tenure problems (McLaughlin et al. 2000;Ling et al. 2007). Huge amounts of fertilizers are regularly added to soils in intensive farming systems to provide sufficient nitrogen (N), phosphorus (P) and potassium (K) for crop growth. The compounds used to supply these elements contain trace amounts of heavy metals which, after continued fertilizer application, may significantly increase soil metal contents (Raven et al. 1998). Integration of remote sensing information within a GIS database can quickly provide detailed soil survey information at low cost. GIS databases can also help derive Digital Elevation Models (DEM), which can help derive landscape attributes utilized in landform characterization (Brough 1986, Dobos 2000. It is critical to analyse the distribution and concentration of metals. This will enable identification of contamination levels and assess associated impacts, on both the environment and human health. Soils are a vital sink for these metals,

Study area
The study area occupies the Middle part of the Nile Delta of Egypt. It is bounded by 30°45′20″-31°10′50″E and 30°35′10″-31°10′05″N, and covers an area of 1927.4 km 2 (Fig. 2). Based on the US Soil Taxonomy (USDA 2010) the soil temperature regime of the study area is Thermic and the soil moisture regime is Torric. The mean annual temperature reaches its maximum in June, July and August and often exceeds 30°C. The mean minimum temperature (11.2°C) usually occurs in January, February or March at Tanta Meteorological Station (Climatologically Normal for Egypt 2011). Precipitation is unequally distributed through the rainy season. Annual rainfall is very low and mostly falls in winter; with a mean 3.8 mm/year. Rain mainly falls in the cold season (November-March) and the minimum amount is in June and September. The area belongs to the late Pleistocene era, which is evidenced by the deposits of the Neonile, which are composed of medium and fine silts (Said 1993).

Digital image processing and physiographic mapping
Digital image processing was completed for two Landsat 8 satellite images (path 177/row 38 and path 177/row 39), with a spatial resolution of 30 m, acquired in May 2014. The images were pre-processed, including radiometric correction (used to modify digital values of pixels to remove noise). Images were geometrically rectified using the Universal Transverse Mercator (UTM) co-ordinates, with the World Geodetic System datum (WGS 1984) and then maps were constructed. Images were atmospherically corrected using the FLAASH module (ITT 2009). Data were calibrated to radiance using the inputs of image type, acquisition date and time. Images were subject to linear stretching by 2%, smooth-filtered, and their histograms were matched, adopting the procedures of Lillesand and Kiefer (2007) and mosaicked using ENVI 5.1 software. The extraction of landform units used high spatial resolution images, so the spatial resolution of satellite image was enhanced using the data merge function of Envi5.1 software. Merging is performed by using multispectral bands (~30 m) as low spatial resolution, and band 8 (panchromatic band) with ~15 m resolution. Landforms were extracted using contour maps (scale 1:25,000) and enhanced satellite images. Both enhanced satellite images were processed with DEM in ERDAS Imagine 8.7, to extract the landform information (Dobos et al. 2002). The initial landform maps were ground-truthed using field observations.

Spatial distribution of heavy metals
Spatial interpolation is widely used when data are collected at distinct locations (e.g. soil profiles) for producing continuous information (Ali, Moghanm 2013). Inverse distance weighted (IDW) is an interpolation method which uses measured values surrounding the prediction location. The measured values closest to the prediction location have more influence on the predicted value than those farther away, thus giving greater weight to poin ts closest to the prediction location, and the weights decrease as a function of distance (Shepard 1968). Geostatistical relationships among the known points (IDW) of Arc-GIS 10.1 software were used to interpolate heavy metal concentrations in the study area. The spatial interpolation method (IDW) was used with 12 neighbouring samples for estimation of each grid point. A power of two was used to weight the nearest points.

Assessment of contamination risk
The Geoaccumulation Index (Igeo) was originally used to evaluate bottom sediment contamination. However, it has been successfully used to evaluate soil contamination (Gowd et al. 2010). The Igeo Index means the assessment of contamination depends on comparing heavy metal concentrations in soils to background values. The calculation of the Geoaccumulation Index uses the equation: Cn 1.5Bn (1) Where C n = the measured concentration of the element in soil. B n = the geochemical background concentration of the heavy metal.
The Geoaccumulation Index (Igeo) is shown in

Soil analysis
Soil samples were collected from nine profiles in El-Gharbia Governorate. The selected profiles represent the different soil units. Pedological descriptions of profiles were conducted using the procedures of FAO (2006) ( Table  2). About 1 kg was collected from each horizon of each profile. Soil samples were air-dried and large stones and organic debris were removed before sieving. Samples were gently ground, homogenized, sieved through a 2.0 mm sieve and then crushed to a fine (<125 μm) powder. Oven-dry samples were ignited at 375•C for 16 hours (overnight), adopting the procedures of Ball (1964). Subsamples of 8.5 g of soil powder were added to 1.5 g of wax (Lico waxc micropowder PM, Hoechst wax)) and then compressed under 12 tonnes pressure by a semi -automatic hydraulic press to make a pellet. The geochemical composition of soil pellets were analysed using an XRF spectrometer model Epsilon3 XLE. XRF analyses were performed at the University of Wolverhampton, UK.

Physiographic map of the study area
The satellite images show that the study area is a flood-plain and includes high terraces (12.04% of area), moderately high terraces (22.41%), low terraces (21.67%), high decantation basins (2.05%), low decantation basins (12.26%), high overflow basins (12.68%), low overflow basins (10.71%), river levees (5.38%) and swales (0.77% ). The main physiographic soil units of the study area are reported in Table 3 and Fig. 3. 1: River terraces: these soils represent the late Pleistocene deltaic plain and occur at the edge of decantation basins (these are basins in which sedimentation, particularly of silt and clay, occur during floods). The soils are formed on terraces at various heights above the valley floor.
2: Basins: these are artificially enclosed areas of a river or harbour, designed so that water levels are unaffected by tides.
3: River levees: these are a type of dam that runs along the banks of rivers or canals. Levees reinforce the banks and help prev ent flooding. By confining the flow, levees can also increase water velocity. 4: Swales: these are low tracts of land, usually consisting of moist and marshy lands. The term can refer to both natural and ar tificial landscape features. Artificial swales are often designed to manage water runoff, filter pollutants and increase rainwater infiltration.  Table 4. For the metals Te, Mo, As and Pb, results are not reported, because their concentrations were below detection limits. Spatial interpolation maps (Figs. 4, 5, 6 and 7) of heavy metal concentrations were prepared using the IDW function (inverse distance weighted) interpolation method in Arc GIS 10.1.

Vanadium
The concentrations and the interpolation map for V in the soil samples are given in Table 4 and Fig. 4. V concentrations ranged from 194.0-744.4 mg/kg with a weighted mean ranging from 206.79-450.58 mg/kg (Table 5). The highest measured concentration of V was in the upper horizon of Profile 2, which represents a swales unit and is located 270 m north of Mansuriyyat Al-Farastaq village, ~6.5 km south-west from the centre of the town of Kfr Elzayat (population in 2015 was 448,965). The Igeo Index showed that all soil samples were in the uncontaminated/moderately contaminated categories, except for first horizon of Profile 2 in the swales mapping unit, which is classified as moderately/strongly contaminated ( Table 6). The high deposition of V might be caused by the numerous local factories. V concentrations are higher than the permissible limits (90 mg/kg), recommended by Bowen (1979) in all soil profile horizons ( Table 5). The spatial interpolation shows an increasing trend from northeast to south-west. The highest weighted mean (weighting concentration by representative area) (450.58 mg/kg) was found in 0.77% of the study area. From the interpolation map of V in the study area (Fig. 4) we can conclude that the order of concentration ascending in the mapping units is: low decantation basin (D2), high overflow basin (OB1), low overflow basin (OB2), moderately high terraces (T2), high decantation basin (D1), levees (L), high terraces (T1), low terraces (T3) and swales (S).

Fig. 4. Spatial interpolation of the weighted mean of vanadium
Chromium Anthropogenic sources of Cr include alloys, chrome plating, pigments, chemical catalysts, dyes, tanning, wood impregnation and refractory bricks (Reimann, de Caritat, 1998). The highest concentration of Cr (519 mg/kg) was in the top-soil of Profile 2, which may be due to the many local factories. The lowest (140.3 mg/kg) was in the top -soil of Profile 6, which represents a high over-flow basin ( Table 4). The mean weight of Cr concentrations ranged from 152.84-314.73 mg/kg (Table 5). All concentrations exceeded the recommended values given by Bowen (1979) and, according to the Igeo Index, most soil samples are in the uncontaminated/moderately contaminated category (Table 6). Cr concentrations increased from east to west and south of the study area (Fig. 5). The highest Cr concentrations tended to be in the swales unit and the lowest in the high overflow basin unit.

Fig. 5. Spatial interpolation of the weighted mean of chromium
Nickel Baghdady, Sippola (1984) reported that the mean total Ni content in Egyptian alluvial soils is 64.4 mg/kg, ranging from 20-74 mg/kg, while the mean NH 4 OAC-EDTA extractable Ni is 1.9 mg/kg, ranging from 1.0-2.2 mg/kg. However, in this study, Ni concentrations ranged from 60.60-267.30 mg/kg (Table 4), with mean weight ranging from 69.60-148.39 mg/kg (Table 5). Ni concentrations were higher in alluvial soils than previous studies and exceeded the permissible limit (50 mg/kg) (Bowen 1979). According to the Igeo Index, all samples are in the uncontaminated and moderately contaminated categories (Table 6). Table 4 reports Ni concentrations in soil profiles and Fig. 6 shows the spatial trends, which increased from north to south and west. The highest concentrations were in swales, which occupy 14.89 km 2 of the study area. The interpolation of Ni shows high spatial variability, with the lowest values in the high decantation basin units.  Table 4 and Fig. 7 report Cu concentrations and the spatial interpolation of weighted mean Cu concentration s, respectively. Cu contents in horizons ranged from 0-288.9 mg/kg and mean weight ranged from 28.90-152.92 mg/kg (Table 5). All concentrations exceeded the permissible limit of 25 mg/kg (Bowen 1979), except for the second horizon of Profile 4 and the first, second and the third horizons of Profile 8 (Table 4). In addition, in Profile 9 the deepest horizon exceeded the limit, whereas concentrations in the upper layer were 0. This is probably due to percolation and illuviation of Cu associated with irrigation water. These profiles represent moderately high terraces, levees and high terraces, respectively. The Igeo Index showed that soil samples were in three contamination categories (uncontaminated/moderately contaminated, moderately contaminated and moderately/strongly contaminated) (Table 6). Two important sources of Cu in the Nile Delta are: (i) applications of Cu-based liquid fungicides, and (ii) use of CuSO 4 as an algicide in treating and controlling problematic macro-algal blooms in the Nile, especially during summer (Abdel-Moati, El-Sammak 1997). The lowest Cu concentrations were in the river levees units and the highest values were in the swale units. Fig. 7. Spatial interpolation of the weighted mean of copper Zinc Zn concentrations slightly exceed the permissible concentration limit of 90 mg/kg (Bowen 1979) (Table 5).
Exceptions include the upper layers of Profiles 2 and 4, where concentrations greatly exceeded permissible limits (377.6 and 308.0 mg/kg, respectively). The highest concentrations were in the upper horizon, but in Profile 8 the highest concentration was in the subsurface (Table 4). This could be caused by infiltration of irrigation water through the profile. The mean weight of Zn ranged between 88.26-201.65 mg/kg. According to the Igeo Index, the Zn concentrations of all soil samples fell into the uncontaminated category, except for the first horizons of Profiles 2 and 4, which were moderately contaminated ( Table 6). The spatial interpolation of Zn is presented in Fig. 8. The highest concentration was in the south-west of the study area, which is located 270 m north of Mansuriyyat Al-Farastaq village. This could be due to atmospheric deposition, originating from local industrial plants. The highest Zn concentrations were in the swales top-soil and moderately high terrace units.

Major Soil oxides
Soil samples were analysed for heavy metals and major oxides (Table 7). Results for SiO 2 , Al 2 O 3 , P 2 O 5 , K 2 O, CaO, MgO, Na 2 O and Fe 2 O 3 were compared with average concentrations of major oxides in soil (Bohn et al. 2001) ( Table  5). SiO 2 concentrations varied from 51.40-56.55% and were all less than the representative average value of 72.64% (Bohn et al. 2001). Al 2 O 3 concentrations varied from 16.28-24.42% and the mean concentration of profiles ranged from 18.81-22.9%. Al 2 O 3 concentrations in all samples exceeded the representative mean value of 13.22%. K 2 O concentrations ranged from 1.05-1.62% with weighted mean values ranging from 1.14-1.35%, near the 1.2% representative mean. CaO concentrations ranged between 2.58-6.69% and the mean weighted value ranged between 3.70-5.95%. CaO and Na 2 O concentrations exceeded the representative means of 1.44% and 0.99%, respectively. Fe 2 O 3 concentration ranged between 9.73-12.23%, whereas the representative mean is 5.77%. P 2 O 5 concentrations ranged from 0.15-0.49%. The weighted mean concentrations of P 2 O 5 samples exceed the 0.18% representative mean (Table 8). Thus, these deltaic soils are predominantly siliceous, with slight enrichment of the alumina component.

Relationships between trace and major elements
V, Cr, Ni, Cu and Zn concentrations are significantly correlated (Table 9). There are no significant correlations between major elements and heavy metal concentrations, except for V, Ni and Zn. V and Ni have significant positive medium and strong correlations with Fe, respectively. This may indicate the sorption of these elements by Fe hydroxides. Ni has a strong positive association with Al, whereas a strong significant association was found between Zn and K 2 O and a medium correlation with P 2 O 5 . Such associations indicate strong affinity for these elements for Fe, Al and K oxides. Al 2 O 3 and Fe 2 O 3 display a strong positive significant correlation. In addition, P 2 O 5 is significantly correlated with K 2 O. There are strong significant negative correlations between Al 2 O 3 and CaO, between CaO and Fe 2 O 3 and a medium significant negative correlation between Na 2 O and Fe 2 O 3 .

Conclusions
This study shows that concentrations of V, Ni and Cr exceeded recommended limits in the soils of the Middle Nile Delta. Cu concentrations were very variable. Zn concentrations slightly exceed the recommended limit. V, Cr, Ni, Cu and Zn concentrations are significantly correlated. There are no significant correlations between major elements and heavy metal concentrations, except for V, Ni and Zn. V and Ni have significant positive medium and strong correlations with Fe, respectively. The Igeo Index of V showed that all soil samples were in the uncontaminated/moderately contaminated categories, except for the first horizon of Profile 2, which is classified as moderately/strongly contaminated, while the Igeo Index for Cr showed most soil samples are in the uncontaminated/moderately contaminated category. The Igeo Index for Ni reveals that all samples are in the uncontaminated and moderately contaminated categories. For Cu, soil samples were in three contamination categories (uncontaminated/moderately contaminated, moderately contaminated and moderately/strongly contaminated). All Zn concentrations were in the uncontaminated category, except for the first horizons of Profiles 2 and 4, which were moderately contaminated. The highest heavy metal concentrations dominate the south-west of El-Gharbia Governorate and is mainly attributed to human activities, especially pollution from the Kfr Elzayat urban area (population in 1995 was 448,965). In terms of the distribution of heavy metals in the different physiographic units, the swales unit contained the highest values, as this is in Kfr El Zayat, which has many factories. We recommend that heavy metal contamination be studied within entire soil profiles and not just top-soils, because these metals affect soil and crop quality and can cause ground-water pollution. Protection against this hazard is vital for sustainable land management. Precise measures and efficient methods to improve soil and water quality must be conducted, in order to prevent soil and water pollution and to avoid the need for costly remediation in the future.