1

ROCK MAGNETIC CHARACTERIZATION OF THE MINE TAILINGS IN 1 

PORTMAN BAY (MURCIA, SPAIN) AND ITS CONTRIBUTION TO THE 2 

UNDERSTANDING OF THE BAY INFILLING PROCESS 3 

4 

Clara Gómez García1*, Fatima Martin-Hernandez1,2,, José Ángel López García3, Pedro 5 

Martínez-Pagán4, Jose I. Manteca4, Carlos Carmona5 6 

1 Department of Geophysics, Fac. Physics, Complutense University, 28040 Madrid 7 

2 Instituto de Geociencias IGEO (UCM, CSIC), Fac. Physics, Complutense University, 28040 8 

Madrid 9 

3 Department of Crystallography and Mineralogy, Fac. Geology, Complutense University, 10 

28040 Madrid 11 

4 Department of Mining, Geological, and Cartographic Engineering, Universidad Politécnica de 12 

Cartagena (UPCT), Paseo Alfonso XIII, 52, 30203, Cartagena (Murcia), Spain 13 

5 Exploraciones Mineras SA, Avenida Apoquindo 4775, Oficina 602, Las Condes, Santiago, 14 

Chile 15 

* Now at: Instituto de Ciencias de la Tierra Jaume Almera, CSIC, Solé i Sabarís s/n, 08028 16 

Barcelona, Spain 17 

corresponding author: fatima@ucm.es 18 

Abstract 19 

The Portman Bay located in the SW Spain has suffered great changes since the 1950s associated 20 

with the mining activity. The direct discharge into the sea of mining waste has caused the silting 21 

of the bay in a few decades. In this study the magnetic properties of recent sands from a set of 22 

52 samples from the bay and 16 samples from an opencast mine are magnetically analyzed. In 23 

particular, we have measured magnetic susceptibility, hysteresis loops, IRM-acquisition, SIRM-24 

  2

IRM and back-field-coercivity spectra, thermomagnetic curves and FORC diagram and took 25 

SEM micropictures. Data are correlated with classical magnetite content estimations derived 26 

from X-ray diffraction (XRD) in three separated grain sized (coarse, medium and fine grain) 27 

and heavy metal derived from X Ray Fluorescence (XRF) in samples from a parallel study at 28 

the same sampling sites and the same sampling campaign. The results indicate high 29 

concentrations of magnetite (up to 28%) compared to magnetically hard minerals that have not 30 

been quantified. Higher concentrations of magnetite are recorded in the central part of the 31 

beach, as indicated by Ms, SIRM and Kinit. Coercivity, however, indicates smaller values in the 32 

center of the Bay, suggesting a larger size of the magnetic particles. The S-ratio or high 33 

coercivity fraction is higher in older samples. Correlation of the magnetic properties with 34 

magnetite grain size indicates a positive correlation between Ms and medium and fine grain size 35 

fraction, determined by X-ray diffraction. In addition, magnetic properties show a direct 36 

relationship between magnetite content and the heavy metals Pb, Cu and As. Our results 37 

indicate that the deposition of magnetite was fast and properties has not been modified in depth. 38 

However, spatial location in the Bay is significant when we examine distribution of grain sizes. 39 

Also, results suggest that magnetite acts as captor of heavy metals. 40 

Keywords: Iron ore, beach placer deposit, magnetite, Portman Bay, magnetic properties, 41 

hysteresis loops, environmental magnetism 42 

1 Introduction43 

The Portman Bay is located in southeast Spain and belongs to the Cartagena - La Unión Mining 44 

District. The Bay is at the foot of the Cartagena range mountain, named “Sierra Minera”, which 45 

crosses the Mining District from west to east, parallel to the Mediterranean coast (Conesa et al., 46 

2008). The Cartagena – La Unión district covers an area of about 10×5 km in the northeast-47 

southwest direction and includes one of the largest Pb - Zn ore deposits in Spain (López-García 48 

et al., 2011) (Figure 1a). 49 

  3

The mining industry has been the most important economic activity in this region for more than 50 

2000 years. The ore mining extraction around the Bay started before the Roman Empire, first by 51 

the Iberians, followed by the Phoenicians and Carthaginians, and ended definitely in 1991. The 52 

wastes generated by the first miners was then used as a source of minerals in later centuries 53 

when the extraction activity decreased. This continued until the mid-nineteenth century when 54 

the large-scale mining development of the region took place (Conesa et al., 2008; Martínez-55 

Sánchez et al., 2008; Oyarzun et al., 2013). 56 

Mining activity throughout the centuries in the region has produced large geomorphological 57 

changes, particularly at the Portman Bay. In the 1950’s the multinational Peñarroya España S. 58 

A. started opencast mining and built a mineral processing plant “Lavadero Roberto”. This 59 

mineral processing plant dumped over 57 million tons of wastes into the Mediterranean Sea, 60 

through the Portman Bay, from 1957 to 1990 (Martos-Miralles et al., 2001; Manteca Martínez et 61 

al., 2005). As a result, the Portman Bay was completely filled in a few decades affecting 62 

significantly the shoreline, which receded about one kilometre. The mineralogical composition 63 

of the beach iron concentrates has been determined to be 17% of quartz, up to 35% of 64 

magnetite, 12% of hematite and 36% of siderite (Manteca et al., 2014). Furthermore, the Bay, 65 

presently, shows high concentrations of metals such as lead, copper, zinc, cadmium and 66 

mercury as well as sulfur and iron oxides (Martínez Orozco et al., 1993; García et al., 2003). In 67 

particular, Zn has been shown to be the most mobile cation with respect to Pb and As 68 

(Martínez-Sanchez et al., 2008). Heavy metal concentration has already been found to be 69 

anomalously high in several biotic elements of the area (e.g., Auernheimer, C. and S. Chinchon, 70 

1997; Cesar et al, 2009; Benedicto et al., 2008). High concentration of pollutants seems to be 71 

related to accumulation of metals in multitude of life elements suboptimal health status of fish 72 

in the area (e.g., Martinez-Gomez et al., 2012). 73 

Because of significant exposed changes occurred in the Bay within few decades as a result of 74 

filling processes, several remediation actions have been proposed. All of these require a 75 

complete quantification of the material deposited, reliable physical control on the sand 76 

  4

composition and associated geophysical properties (Peña et al., 2013). In this particular aspect, 77 

the tailing has been estimated to contain up to 444 � 103 t of magnetic iron concentrate in the 78 

form of magnetite, hematite and siderite. Finally, if the present beach black sands represent 79 

approximately 1/6 of the total bay infilling, the actual amount of iron in Portman Bay’s could be 80 

much larger, at least on the order of 1 � 106 t (Manteca et al., 2014). 81 

The use of magnetic properties the evaluate iron resources has been gaining importance in 82 

exploration and mineral prospecting in the last decades (e.g., Clark, 1999; Clark, 2014 and 83 

references therein). The work presented here aims to study the magnetic properties of recent 84 

sands from the Portman Bay area. The concentration of magnetic minerals deposited is also 85 

analyzed and we will discuss their spatial distribution along the Bay and the temporal 86 

distribution as a function of the time of deposition. Also, the magnetic parameters will be 87 

compared with estimations of heavy metals derived from a parallel spectrometric analysis in the 88 

same sampling sites (Carmona Manzano, 2012). 89 

It is particularly important to highlight that recently there is a rising interest in the regional 90 

community to environmentaly regenerate the Portman Bay and recover the iron-bearing waste 91 

as economic resource (Murciatoday, 2014). Thererefore the findings from this study are 92 

significant not only in relation to the Portman Bay itself, and future remediation and sustainable 93 

development of the area but also for coastal areas with similar concerns. 94 

2 Geological‐Enviromentalcontext95 

Sierra Minera is located at the eastern end of the Betic Ranges. It is characterized by an overlaid 96 

layer structure of tectonostratigraphic complexes. The basement is formed by the Nevado - 97 

Filábride (Paleozoic to Triassic) and Alpujárride (Permian to Triassic) complexes (Figure 2), 98 

which emerge as superimposed thrust sheets of Alpine age and are affected by metamorphism 99 

decreasing from bottom to top. In the Late Miocene, these basal units suffered a gravitational 100 

collapse accompanied by a significant high-K calc-alkaline volcanism and marine sedimentation 101 

in extensional basins of Miocene to Pliocene age. After the Neogene, there was an important 102 

  5

phase of fracturing followed by volcanic events and uplift of the Sierra, and the currently 103 

erosive dismantling (Manteca Martínez and Ovejero Zapino, 1992; Vera, 2004; López-García et 104 

al., 2011). 105 

The following types of mineralization can be described in the Sierra Minera the following types 106 

of mineralization (Oen et al., 1975; Manteca Martínez and Ovejero Zapino, 1992): (i) mantos: 107 

masses and stratabound bodies found in Nevado - Filábride (second manto) and Alpujárride 108 

(first manto) complexes. Both manto units have two sets of different minerals: (a) greenalite - 109 

magnetite - sulphide - carbonate - silica (silicate manto) and (b) chlorite- sulphide-carbonate-110 

silica (pyrite manto). Most of the lead and zinc produced in the district were extracted from 111 

these manto-type deposits. (ii) Disseminations in the Miocene: irregular bodies elongated in a 112 

NW – SE direction, with respect to the fracturing. They were the most important mineral 113 

resource after the majority of mantos was used up. This is sphalerite-rich mineralization with 114 

pyrite, marcasite and galena as minor phases. (iii) Veins and stockworks associated with 115 

vulcanites and (iv) Monteras or Gossans type deposits are also developed. However the 116 

stockworks (iii) and Gossan (iv) type were not subject to mining by the Peñarroya Company. 117 

The main minerals that have been described are goethite, hematite, silica and, to a lesser 118 

proportion, mineral of the jarosite group and alunite along with carbonates and clayey minerals 119 

(López García, 1985; Carmona Manzano, 2012). 120 

121 

3 Materialsandmethods122 

A total of 52 samples of recent sands of the Portman Bay (labeled as “PP samples”) have been 123 

analyzed. Nine sampling sites have been selected, distributed along the Portman bay coastline 124 

placer and one site at the interior of the bay (Figure 1b). Samples from sites PP1, PP1 - B, PP2, 125 

PP2 – B y PP3 – B have been taken at several depths. The PP3 - A samples and a fraction of 126 

PP5 samples (the PP5 - D) are hardpan-type or duricrust. Hardpan term refers to a dense layer of 127 

that is normally found in the topsoil layer, using here the duricrust term meaning no soil 128 

  6

development is necessary happening. Moreover, the PP4 samples are sulphated facies. In total, 129 

26 samples were collected from the Bay. Furthermore, for each sampled point, two specimens 130 

were taken to check their homogeneity. 131 

A small specimen of rock material was sampled from an opencast mine, the Emilia open pit in 132 

the vicinity of the town El Llano del Beal, 4 km north of Portman (Figure 2). This was divided 133 

into 16 small fragments, due to it being highly heterogeneous in composition (M samples) and 134 

represents the original mined material. This material is made to have a statistical sampling of 135 

the magnetic properties of the original magnetic signal with respect to the mining processed 136 

tailing. 137 

Measurements include magnetic and non-magnetic techniques. Among the first, the magnetic 138 

susceptibility of 25 PP samples was determined in a constant low-field (300 A/m) using a 139 

Kappabridge KLY-4 manufactured by AGICO (Advanced Geoscience Instruments Co.) 140 

(Hrouda et al., 2006). Also, low-field susceptibility has been measured at two frequencies for 52 141 

PP samples using a Bartington Instruments MS2 Magnetic Susceptibility System. The 142 

frequency dependence indicates the presence of ferromagnetic grains that are possessing 143 

superparamagnetism (Tauxe, 2009). 144 

Hysteresis loops have been measured up to 0.5 T with a spectrometer of coercivities developed 145 

by the University of Kazan (Coercivity Spectrometer J-meter) (Jasonov et al., 1998) for 52 PP 146 

samples and 16 M samples. Data have been corrected for paramagnetic/diamagnetic 147 

contribution, to characterize the magnetic parameters in the studied area: paramagnetic 148 

susceptibility (κpara), saturation magnetization (Ms), remanent magnetization saturation (Mrs), 149 

coercive force (Bc) and initial susceptibility (κini). 150 

The magnetic characterization also includes acquisition of Saturation of Isothermal Remanent 151 

Magnetization (SIRM) up to 0.5 T and further demagnetization by static fields in the opposite 152 

direction (back - field IRM) in a Coercivity Spectrometer J-meter. The coercivity of remanence 153 

(Bcr) and total value of SIRM have been computed for all PP and M samples. Also, the amount 154 

  7

of high coercivity fraction has been estimated computing the S-ratio as the difference is 155 

remanence acquired between 100mT and 500mT (Evans and Heller, 2003). Also, the IRM - 156 

acquisition curves were modeled by a series of log-normal distributions of the IRM gradient in 157 

order to infer the coercivity distribution associated to magnetic fractions. The mathematical 158 

fitting has been modeled by the method outlined by Kruiver et al., (2001) and the associated 159 

software. 160 

Thermomagnetic curves were generated through a Petersen Instruments Variable Field 161 

Translation Balance (VFTB) in order to determine the Curie/Neél temperature of the 162 

ferromagnetic phases. Curie/Neél temperatures were determined for a selection of 26 PP 163 

samples and 5 M samples by the maximum in second derivative method (Moskowitz, 1981; 164 

Tauxe, 1998). 165 

Magnetic parameters have been summarized in a Day plots attempting to discriminate domain 166 

state of the magnetic particles (Day et al., 1977; Dunlop, 2002). First-Order Reversal Curve 167 

(FORC) diagrams were measured in material from site PP1 in order to achieve knowledge about 168 

particles interaction and/or domain state (Pike et al., 1999; Roberts et al., 2000). Sand material 169 

was consolidated using non-magnetic glue into a gel cap. FORC diagrams have been processed 170 

with the software and methodology described by Harrison and Feinberg (2008) using a 171 

smoothing factor, SF-6. 172 

Physical characterization and pseudo-compositional estimations have been achieved also by 173 

scanning electron microscopy (SEM), helping in the identification of size and morphology of 174 

the magnetic grains. The SEM was carried out in 6 PP samples (PP1B-A (30-38), PP2-B (25-175 

30), PP3B–A SUP, PP4-A, PP5P-B y PP6C-B) for the study of the physical structure, particle 176 

size, morphology and location within the sandy material of the ferromagnetic particles. 177 

Magnetic properties have been compared with magnetite content estimated in grain size 178 

fractions run in a parallel study, specific details can be found elsewhere (Carmona Manzano, 179 

2012). Recent sand samples have been physically clasified into three fractions, Coase grain size 180 

  8

(C > 0.40 mm), medium (0.40<M<0.20 mm) and fine (F<0.20 mm). Magnetite content in the 181 

different particle sizes has been evaluated by X-ray diffraction. The most relevant heavy metal 182 

and meteloids concentration has been measured by X-ray Fluorescense (XRF), finding 183 

significant concentrations of Pb, Zn, Cu, As, Cd and Hg. 184 

4 Results185 

4.1 Low‐fieldsusceptibility186 

Values obtained for the magnetic susceptibility defined in low-field (300 A/m) are summarized 187 

in Figure 3a for sample PP and Figure 3b for M samples . They both show a large variability of 188 

values between 0.74×10-5 and 14.70×10-5 m3/kg in the case of PP samples and from -0.07×10-5 189 

to 6.35×10-5 m3/kg in M samples. Magnetic susceptibility of the sand samples are characterized 190 

by two relatively equal maxima centered at 1.7 10-5 m3/kg and 12 10-5 m3/kg, which suggests a 191 

bimodal distribution of the low field susceptibility (Figure 3a). The low field susceptibility of 192 

the mine sample, however, is characterized by one maximum value centered at 1.12 10-5 m3/kg 193 

(Figure 3b). 194 

Measurements of the magnetic susceptibility in low-field with two frequencies indicate that 195 

there is almost no frequency dependence. Values vary between 0.01% and 1.66%, therefore the 196 

presence of superparamagnetic material (SP) is considered not to be significan (Hrouda, 2011). 197 

4.2 Hysteresisloops198 

The hysteresis loops all have very similar shapes for the PP samples (Figure 4a) and M samples 199 

(Figure 5a). Magnetization reaches the reversible state at a saturating field of about 250 mT, 200 

with no indications of a high coercivity fraction in the induced magnetization. Hysteresis 201 

derived parameters, however, show slight differences depending on the source (PP and M 202 

samples), as well as sampling site (PP1 to PP6) (Table 1). Remanence parameters (Ms and Mr), 203 

display large variability, even recorded between twin samples form the same level and site, 204 

indicating a highly heterogeneous concentration of magnetic minerals (Evans and Heller, 2003). 205 

  9

Magnetic coercivity is less variable. The largest differences can be found associated to the 206 

sample source, with values between 10.28 and 12.72 mT from the sand samples and values 207 

between 11.52 and 15.49 mT, from the Emilia open pit samples (Table 1). 208 

4.3 IRM‐acquisitionandback‐fieldSIRM209 

SIRM acquisition up to 0.5 T and subsequent back-field SIRM curves of PP (Figure 4b) and M 210 

samples (Figure 5b) present a very similar shape. The beach sand samples do not have an initial 211 

remanence (Figure 4b) while samples from the mine have an original remanence magnetization, 212 

which is observed by an initial value of the SIRM at low fields (Figure 5b). Samples approach 213 

to saturation at about about 200 mT and asymptotically reach an almost constant value at 500 214 

mT . Values of the S-ratio range between 11.72 and 4.15 % in sand beach samples and 7.49 and 215 

2.19 in M samples (Table 1). Subsequent back-field SIRM shows Bcr values between 24.8 and 216 

30.4 mT for samples PP and 22.5 and 32.4 mT for samples M (Table 1). 217 

Both SIRM and back-field SIRM show one inflection point (Figure 4b, Figure 5b) which 218 

suggests the presence of a single significant magnetic mineral responsible for the remanence 219 

within the soft coercivity fraction. 220 

4.4 IRM‐coercivityspectra221 

IRM acquisition curves and first derivative have been mathematically modeled following the 222 

procedure proposed by Kruiver et al.(2001). Figure 4c and Figure 5c show the example of the 223 

derivative of the IRM acquisition curve (Gradient Acquisition Plot, GAP) and its corresponding 224 

fitting using a series of log-normal distribution curves (Kruiver et al., 2001). All the analyzed 225 

samples have a similar behaviour, the gradient of the IRM acquisition curve is characterized by 226 

a main distribution that carries most of the magnetization (about 90% of the total SIRM) which 227 

has been named as Component 1. It has been necessary to include two additional components of 228 

low and high coercivity labeled Component 2 and Component 3, respectively, in order to reduce 229 

the residual of the GAP curve. The low coercivity component (Component 2) provides an 230 

average of 7% to magnetization and, at such low fields, its presence has been attributed to 231 

  10

thermal activation of magnetic particles (Heslop et al., 2004). Component 3 contributes only 3% 232 

to the total SIRM, but suggests the presence of a higher coercivity phase. 233 

The field at which the SIRM acquisition curve reaches its inflexion point (B1/2), or maximum 234 

value of the GAP, is used as an estimation of the average coercivity of the corresponding 235 

magnetic population carrying the remanence. Table 1 summarizes the values of the B1/2 together 236 

with the associated dispersion parameter (DP) of the coercivity distribution. 237 

Both samples PP and M have a very similar coercivity, with a mean B1/2=38.42.9 mT for the 238 

samples collected from the Bay and B1/2=36.28.4 for the samples belonging to the Emilia open 239 

pit. The results confirm the well-constrained coercivity found in the samples derived from the 240 

infilling process with the detachment having a larger variability of the magnetic properties 241 

(Table 1) 242 

243 

4.5 Thermomagneticcurves244 

Thermomagnetic curves for PP samples have very similar characteristics (Figure 4d) during 245 

heating and subsequent cooling magnetization processes. The Curie/Neél temperature ranged 246 

between 550 °C and 580 °C. This Curie/Neél temperature indicates the presence of either 247 

magnetite (Curie temperature is 580 °C), or titanomagnetite with low Ti content (Dunlop and 248 

Özdemir, 1997). The decrease found above 580 ºC suggests the presence of a magnetic mineral 249 

that loses magnetization at 680 ºC, i.e., hematite (Figure 4d). Since the susceptibility of hematite 250 

is lower than that of magnetite, the observation of the magnetization tail attributed to hematite 251 

suggests it contains large concentrations of it (Evans and Heller, 2003). 252 

A closer analysis to the heating curve reveals two minor changes in the curve slope, one 253 

between 250 °C and 350 °C, approximately, and the second minor change around 450 °C and 254 

550 °C, which might indicate the very mild transformation in the samples (Henry et al., 2005). 255 

These two features might be indicative of the presence of iron hydroxides and/or jarosite, that 256 

display a breackdown at about 250-300ºC and often transforms into maghemite-hematite 257 

(Hanesch et al., 2006). The presence of goethite and jarosite has been confirmed by XRD 258 

  11

measurements (Carmona, 2012) and it is also supported by the lower intensity of the cooling 259 

with respect to the warming curve. 260 

The Curie/Neél temperature in M samples (Figure 5d) is also around 580 °C. However, the 261 

relationship between heating and cooling tracks has the opposite behavior. The heating curves 262 

show lower magnetization than the cooling track of the curve. Also, there is no clear evidence of 263 

a change in the slope of the curve for the 250-350 ºC range (Figure 5d). The high temperature 264 

range does not suggest the presence of hematite (Figure 5d). 265 

4.6 Domainstatesummary‐plots:DayplotandFORCdiagram266 

The also called Day plot is a good estimation of the domain state for magnetite and magnetite-267 

like particles (Day et al., 1977). This representation discriminates the domain state separating in 268 

different areas of the diagram the coercivity ratios (Hcr/Hc) as a function of magnetization ratios 269 

(Mr/Ms). Results from this study for samples PP and M are presented in Figure 6a. Both sets are 270 

clustered in the pseudosingle domain state (PSD) of the classical interpretation of the plot (Day 271 

et al., 1977) or a mixture of about 50% singledomain (SD) and 50% multidomain (MD) mixing 272 

curve 2 in the model proposed by Dunlop (2002). 273 

The sand samples deposited on the Bay are magnetically softer with data getting closer to the 274 

MD state than samples from the mine (Figure 6a). Moreover, there is no significant shift toward 275 

the superparamagnetic mineral fields (Dunlop, 2002). Considering only the PP samples in the 276 

Day plot values are grouped according to the sampling site and spatial distribution within the 277 

Bay (Figure 6a). 278 

The FORC diagram for PP sample suggests a distributed particle size in the PSD to SD range, 279 

as can be seen by the spread of the contour lines along the y-axes of the diagram (Pike et al, 280 

1999; Roberts et al., 2000). 281 

  12

4.7 SEM282 

SEM micropictures using the secondary electrons method have been taken for every sampling 283 

site. This technique allows a quick identification of the large metallic fraction, visible as bright 284 

areas in the image (Reed, 2010). Figure 7 shows the images obtained for the six sampling sites. 285 

Big sand grains appear partially covered by smaller bright zones in their corners and edges in all 286 

samples (Figure 7a-f). A closer look into the morphological information of the grains shows 287 

hexagonal structures and planar crystals suggested to be jarosite grains (Figure 7b and 7c). 288 

Pseudocompositional analysis based on spectral identification of scattered electrons reveals 289 

these bright zones are related to the presence of magnetite intergrowths with quartz grain 290 

(Figure 7e-f). No heavy metals have been identified as single grains at this resolution level, 291 

indicating their sizes are smaller than the detection limit of the technique. 292 

5 Discussion293 

The magnetic results provide information needed to answer three key questions in the study of 294 

mining debris: (i) are there a distinct difference in the physical-magnetic properties of samples 295 

from the Bay sands (PP samples) and original samples from the mine (M samples), (ii) is there a 296 

correlation between PP and M samples; and (iii) is there a correlation of the magnetic 297 

parameters with heavy metals content from the same sampling sites. 298 

In order to study the spatial distribution along the Portman Bay, mean values and their 299 

corresponding uncertainties from hysteresis and remanence parameters (Ms, Mrs, Hc, Hcr, Kint, 300 

Kpara, SIRM, B1/2 and S-ratio) of the PP samples have been computed. An arithmetic mean and 301 

standard deviations are graphically displayed in Figure 8. Sampling sites located at the edges of 302 

the Bay (PP1, PP1 – B or PP6) are significantly and statistically different from those sites 303 

located at the central part of the Bay. With these two spatial distributions (ends or central parts 304 

of the Bay) we have a uniform trend for every parameter that is broken by the values of the PP3 305 

– B sample: this sample has widely scattered values with a large error bar that includes values 306 

within the general trend. 307 

  13

The results are consistent because there is a lower content of paramagnetic material (Figure 8a) 308 

in the central part of the Bay where the ferromagnetic contribution is greater (Figure 8b). 309 

Besides, κini, Ms, Mr and SIRM show the same spatial distribution (Figure 8b, c, d and g). Bcr 310 

has the same distribution that these parameters but with a very small variation range of its 311 

values (Figure 8 f). Bc seems to show higher values at the ends of the Bay and lower in the 312 

central parts though, the PP3 - B and PP4 areas have higher values than their adjacent (Figure 313 

8e). However, differences between sampling sites are small and their statistical significance is 314 

limited. B1/2 has its maximum for the PP4 area and the minimum for the PP1 area (Figure 8h). 315 

The S-ratio seems to display a similar trend as that observed for Bcr and B1/2 (Figure 8i). Indeed, 316 

S-ration statistically correlates to Bcr and B1/2 (Figure 9). The correlation between S and Bcr 317 

indicates the higher coercivity might be related to the presence of a high coercivity mineral 318 

(Figure 9a) while a correlation with the B1/2 might be related to variations in the magnetite grain 319 

size (Figure 9b). Results are not conclusive, but statistically, the correlation seems to be stronger 320 

with respect to B1/2, therefore a grain size dependence is expected between different sampling 321 

sites. 322 

We can establish three different classifications according to the variability of the parameters in 323 

the Bay: (a) variation with depth, (b) spatial variation along the Bay and (c) temporal variation 324 

based on the age of the sediments that have been deposited in the Bay by considering its 325 

distribution. 326 

(a) Taking into account the samples that have been collected at different depths in the Bay (PP1, 327 

PP1 - B, PP2, PP2 - B and PP3 - B), the parameters obtained do not vary with respect to depth 328 

(Table 1), at least for the studied depths, which did not exceed 80 cm. This suggests that 329 

alterations, diagenetic processes or transformations due to the sedimentary aging are not 330 

important (Evans and Heller, 2003). One should note, however, that the infilling was very rapid, 331 

so that the upper most layers have not yet started to undergo alteration; this does not preclude 332 

alteration of deeper strata, considering the total thickness of the mineral deposits which is 333 

estimated to be 10 m (Martinez-Sanchez et al., 2008; Manteca et al., 2014). 334 

  14

(b) Figure 8 illustrates the spatial distribution of the hysteresis and remanence parameters but 335 

these may also be related to temporal effects, since the filling of the bay was earlier in some 336 

areas than others. Consequently, minerals have accumulated in some places for longer than in 337 

others. 338 

(c) The filling of the Bay began at the western edge of the bay (next to the discharge point 339 

where the wastes were poured into the sea) (Martínez Orozco et al., 1993), then the next stage 340 

was the waste infilling the bottom of the Bay from the south towards the north. Marine currents 341 

affected the way in which sediments have been deposited and distributed since they play a key 342 

role in differential particles sedimentation processes depending on their grain size. The coarsest 343 

grains have remained on the beach while the finest ones have been deposited at sea bottom. In 344 

addition, there is a dynamic mobility: in good weather conditions, the material has accumulated 345 

in deposits that are partially or totally destroyed during storms, and they are reformed in another 346 

place under more favorable weather conditions. Therefore, the coastline is constantly being 347 

modified and where the emerged beach is considered as unstable (Martínez Orozco et al., 1993). 348 

In view of bathymetric maps (Martínez Orozco et al., 1993) we assume that the minerals in PP4 349 

and PP3 samples are the oldest ones because that sampled area is the farthest from the coastline 350 

and this is observed in the similarities between parameters in PP4 and PP3A. Some of the 351 

magnetic parameters for the PP3 – B samples are significantly different from those measured in 352 

the adjacent PP3 –A area. Although we do not know the reason for this difference and the error 353 

bars associated with PP3 – B samples cover the general trend, perhaps the different geological 354 

properties that may present both samples have some influence: PP3 – A samples are duricrust 355 

while PP3 - B samples are not. It should be noted that the PP3 – B samples also contain 356 

sphalerite and abundance of jarosite (Carmona Manzano, 2012). Although these phases are 357 

paramagnetic at room temperature (Hunt et al, 2013; Majzlan et al., 2000), they may be 358 

associated with ferromagnetic intergrowths, which have not been observed under SEM. 359 

  15

The magnetite content determined by classical XRD has been correlated to the estimation 360 

derived from magnetic hysteresis. Because magnetization curves are mainly controlled by 361 

magnetite, Ms values from hysteresis can be transformed into percentage of magnetite dividing 362 

by the tabulated value of 90 Am2/kg (Dunlop and Özdemir, 1997). Values from each level depth 363 

have been averaged and standard deviation computed from Table 1. Estimations of magnetite by 364 

both methods are recorded in Table 2. The concentration of magnetite in PP samples varies 365 

between 21.83% and 0.97% by magnetic methods and 34.04 and 7.78% by X-ray diffraction 366 

(Table 2). M samples were only studied by magnetic methods, giving a variation range of 367 

magnetite from 5.8% to 0.02% (Table 1). 368 

Magnetite content derived from hysteresis is correlated to κini and Mr parameters (Figure 10a 369 

and b). The Pearson p-parameter for these two magnitudes gives a “strong correlation” with 370 

respect to the magnetite content according to the classification given by Evans (1996) with 371 

values higher than 0.8. Anthropogenic action constrains the range of values of the coercive field 372 

since what we find in Portman Bay (such as the PP samples) are the products of the mine 373 

extraction (such as M samples) (Figure 10c). Correlations are not conclusive, because the 374 

magnetite concentration of M samples is much lower than in PP samples. 375 

Hysteresis and remanence results can be correlated with the magnetic granulometry of the PP 376 

samples and with the heavy metals content that was obtained by X-ray diffraction analysis 377 

(Table 2) (Carmona Manzano, 2012). 378 

Magnetite has been fractioned into three grain-sizes and quantified by XRD: coarse (C), (> 0.40 379 

mm), medium (M), (0.40 to 0.20 mm) and fine (F) (<20 mm) grains. Bulk Ms derived from 380 

magnetic hysteresis has been compared with the grain-size content in Figure 11a-c. The coarse-381 

grain fraction is directly correlated with Ms while the opposite happens with the medium-grain 382 

fraction (Figure 11 b). Fine particles do not exhibited a good agreement with Ms (Figure 11c). 383 

Because magnetite is known to be an effective heavy metals captor (Biswal et al., 2013; Mayo 384 

et al, 2007), Ms was compared with heavy metals content, specifically lead, zinc, iron, copper, 385 

  16

arsenic, cadmium and mercury determined by XRF (Carmona Manzano, 2012) (Table2). We 386 

obtained the best correlations for lead, copper and arsenic (Figure 11d,e and f). Copper and 387 

arsenic (Figure 11e-f) can not be correlated linearly, but we have few samples to try another fit. 388 

Table 2 depicts the correlation found between the magnetite content by X-ray diffraction and 389 

magnetic methods. We observe that the slope of the correlation is nearly 1 and the determination 390 

parameter is high. In view of the results we can say that magnetic methods are faster and very 391 

effective for studies of Portman Bay (Figure 12). Particularly, magnetic methods are more 392 

suitable for low concentrations of magnetite, when XRD is known to have poor resolution. 393 

6 Conclusions394 

We can conclude related to the study of the rock magnetic characterization of the Portman Bay 395 

in four main points: 396 

397 

a) The study of the magnetic properties of the PP samples has shown high magnetite 398 

content. Predominantly high concentration of magnetite in all samples with 399 

concentrations up to 21.8%. 400 

b) Magnetically hard mineral phases such as hematite are detected by magnetic methods 401 

but quantification is not possible. The magnetite is relatively hard with saturating fields 402 

higher than 0.2 T. Despite having strong indications of the presence of iron hydroxides 403 

in the system, they do not contribute significantly to the remanence. 404 

c) We have also observed that the magnetic properties of the PP samples correspond to a 405 

characteristic spatial distribution in the bay, and this distribution may be closely related 406 

to how sediments were deposited in the Bay depending on marine currents. 407 

d) We do not observe significant variations in the magnetic properties with depth, up to 408 

values of about 80 cm, therefore diagenetic alterations, superficial transformations in 409 

maghemite and / or aging alterations can be excluded. 410 

411 

  17

From the results, we observed agreement between the PP samples and M samples , despite the 412 

differences attributed to the transformation in the processing plant. 413 

414 

In view of future research, a much more comprehensive bay sampling would be necessary, with 415 

more samples of the coastline and especially from inside the Bay. Having a dense spatial 416 

distribution of samples from the bay would allow for a better understanding of the interaction 417 

between deposition and latter reworking from currents, including depth effects. It would also 418 

allow for a better statistical evaluation. This would further develop the theory of spatial 419 

distribution along the Bay of the parameters of the magnetic properties, test to a greater extent 420 

the use of magnetite as captor heavy metal, clarify the formation of magnetic minerals, 421 

determine the oxidized phases found in the thermomagnetic curves, etc. 422 

423 

Acknowledgments 424 

Sara Guerrero Suárez is specifically thanked for assisting part of the magnetic measurements in 425 

the UCM-paleomagnetic lab. Also Prof. Eric C. Ferré for running FORC curves and hysteresis 426 

at the USI-Carbondaile rock magnetic laboratory. The manuscript has been improved thanks to 427 

two anonymous reviewers. 428 

References 429 

Auernheimer, C. and S. Chinchon (1997). Calcareous skeletons of sea urchins as indicators of 430 

heavy metals pollution. Portman Bay, Spain. Environmental Geology, 29(1-2), 78-83. 431 

Benedicto, J., C. Martinez-Gomez, J. Guerrero, A. Jornet and C. Rodriguez (2008). Metal 432 

contamination in Portman Bay (Murcia, SE Spain) 15 years after the cessation of mining 433 

activities. Ciencias Marinas, 34(3), 389-398. 434 

  18

Biswal, M., K. Bhardwaj, P. K. Singh, P. Singh, P. Yadav, A. Prabhune, C. Rode and S. Ogale 435 

(2013). Nanoparticle-loaded multifunctional natural seed gel-bits for efficient water 436 

purification. Rsc Advances, 3(7), 2288-2295. 437 

Carmona Manzano, C. (2012), Estudio mineralógico del placer costero de magnetita de la Bahía 438 

de Portmán ( Cartagena ), 71 pp., Universidad Complutense de Madrid. 439 

Cesar, A., A. Marin, L. Marin-Guirao, R. Vita, J. Lloret and T. A. Del Valls (2009). Integrative 440 

ecotoxicological assessment of sediment in Portman Bay (southeast Spain). Ecotoxicology and 441 

Environmental Safety, 72(7), 1832-1841. 442 

Clark, D. A. (1999). Magnetic petrology of igneous intrusions: implications for exploration and 443 

magnetic interpretation. Exploration Geophysics, 30(1/2), 5-26. 444 

Clark, D. A. (2014). Methods for determining remanent and total magnetisations of magnetic sources – 445 

a review. Exploration Geophysics, 45(4), 271-304. 446 

Conesa, H. M., R. Schulin, and B. Nowack (2008), Mining landscape: A cultural tourist 447 

opportunity or an environmental problem?, Ecol. Econ., 64(4), 690–700, 448 

doi:10.1016/j.ecolecon.2007.06.023. 449 

Day, R., M. Fuller, and V. A. Schmidt (1977), Hysteresis properties of titanomagnetites: grain-450 

size and compositional dependence, Phys. Earth Planet. Inter., 13, 260–266. 451 

Dunlop, D. J. (2002), Theory and application of the Day plot (M-rs/M-s versus H-cr/H-c) 2. 452 

Application to data for rocks, sediments, and soils, J. Geophys. Res. Earth, 107(B3), -, 453 

doi:Artn 2056 Doi 10.1029/2001JB000487. 454 

Dunlop, D. J. and Ö. Özdemir (1997). Rock Magnetism. Cambridge, Cambridge Univ. Press. 455 

Evans, J. D. (1996). Straightforward statistics for the behavioral sciences. Pacific Grove, Brooks/Cole 456 Pub. Co. 457  458 

  19

Evans, M. E., and F. Heller (2003), Enviromental Magnetism: Principles and Applications of 459 

Enviromagnetics, edited by R. Dmowska, J. R. Holton, and H. T. Rossby, Accademic 460 

Press, London. 461 

García, G., H. Conesa, and Á. Faz (2003), Contaminación por metales pesados de los suelos y 462 

sedimentos mineros del distrito Minero de Cartagena–La Unión (Murcia), Patrim. 463 

Geológico y Min. y Desarro. Reg. Inst. Geológico y Min. España, Madrid, Spain, (In: 464 

Rábano, I., Manteca, I., García, C. (Eds.)), 295–299. 465 

Hanesch, M., H. Stanjek and N. Petersen (2006). Thermomagnetic measurements of soil iron 466 

minerals: the role of organic carbon. Geophysical Journal International, 165(1), 53-61. 467 

Harrison, R. J. and J. M. Feinberg (2008). FORCinel: An improved algorithm for calculating 468 

first-order reversal curve distributions using locally weighted regression smoothing. 469 

Geochemistry Geophysics Geosystems, 9. 470 

Henry, B., D. Jordanova, N. Jordanova and M. Le Goff (2005). Transformations of magnetic 471 

mineralogy in rocks revealed by difference of hysteresis loops measured after stepwise 472 

heating: theory and case studies. Geophysical Journal International, 162(1), 64-78. 473 

Heslop, D., G. McIntosh, and M. J. Dekkers (2004), Using time- and temperature-dependent 474 

Preisach models to investigate the limitations of modelling isothermal remanent 475 

magnetization acquisition curves with cumulative log Gaussian functions, Geophys J Int, 476 

157(1), 55–63. 477 

Hrouda, F., M. Chlupacova, and J. Pokorny (2006), Low-field variation of magnetic 478 

susceptibility measured by the KLY-4S Kappabridge and KLY-4A magnetic susceptibility 479 

meter: Accuracy and interpretational programme, Stud. geoph. geod., 50, 283–298. 480 

Hrouda, F. (2011). Models of frequency-dependent susceptibility of rocks and soils revisited 481 

and broadened. Geophysical Journal International, 187(3), 1259-1269. 482 

  20

Hunt, C. P., B. M. Moskowitz and S. K. Banerjee (2013). Magnetic Properties of Rocks and 483 

Minerals. Rock Physics & Phase Relations, American Geophysical Union: 189-204. 484 

Jasonov, P. G., D. K. Nougaliev, B. V Burov, and F. Heller (1998), A modernized coercivity 485 

spectrometer, Geol. Carpathica, 49, 224–225. 486 

Kruiver, P. P., M. J. Dekkers, and D. Heslop (2001), Quantification of magnetic coercivity 487 

components by the analysis of acquisition curves of isothermal remanent magnetisation, 488 

Earth Planet. Sci. Lett., 189(3-4), 269–276. 489 

López García, J. A. (1985), Estudio Mineralógico, textural y geoquímico de las zonas de 490 

oxidación de los yacimientos de Fe, Pb, Zn de la Sierra de Cartagena, Murcia, Universidad 491 

Complutense de Madrid. 492 

López-García, J. A., R. Oyarzun, S. López Andrés, and J. Manteca Martínez (2011), Scientific, 493 

Educational, and Environmental Considerations Regarding Mine Sites and Geoheritage: A 494 

Perspective from SE Spain, Geoheritage, 3(4), 267–275, doi:10.1007/s12371-011-0040-2. 495 

Majzlan, J., P. Glasnák, R. Fisher, M. White, M. Johnson, B. Woodfield and J. Boerio-Goates 496 

(2010). Heat capacity, entropy, and magnetic properties of jarosite-group compounds. 497 

Physics and Chemistry of Minerals, 37(9), 635-651. 498 

Manteca Martínez, J. I., and G. Ovejero Zapino (1992), Los yacimientos de Zn, Pb, Ag - Fe del 499 

distrito minero de La Unión - Cartagena, Bética oriental, Recur. Miner. España (García 500 

Guinea, J., Martínez Frías, J, Coords), CSIC, Madrid, 1085–1102. 501 

Manteca Martínez, J. I., M. A. Pérez de Perceval Verde, and M. . López-Morell (2005), La 502 

industria minera en Murcia durante la época contemporánea, pp. 123–134. 503 

  21

Manteca, J. I., J. Á. López García, R. Oyarzun and C. Carmona (2014). The beach placer iron 504 

deposit of Portman Bay, Murcia, SE Spain: the result of 33 years of tailings disposal 505 

(1957-1990) to the Mediterranean seaside. Miner Deposita, 49(6), 777-783. 506 

Martinez-Gomez, C., B. Fernandez, J. Benedicto, J. Valdes, J. A. Campillo, V. M. Leon and A. 507 

D. Vethaak (2012). Health status of red mullets from polluted areas of the Spanish 508 

Mediterranean coast, with special reference to Portman (SE Spain). Marine Environmental 509 

Research, 77, 50-59. 510 

Martínez Orozco, J. M., F. V. Huete and S. G. Alonzo (1993). Environmental problems and 511 

proposals to reclaim the areas affected by mining exploitations in the Cartagena mountains 512 

(southeast Spain). Landscape and Urban Planning, 23, 195-207 513 

Martínez-Sánchez, M. J., M. C. Navarro, C. Pérez-Sirvent, J. Marimón, J. Vidal, M. L. García-514 

Lorenzo, and J. Bech (2008), Assessment of the mobility of metals in a mining-impacted 515 

coastal area (Spain, Western Mediterranean), J. Geochemical Explor., 96(2-3), 171–182, 516 

doi:10.1016/j.gexplo.2007.04.006. 517 

Martos-Miralles, P., A. Sansano Sánchez, P. Baños Páez, J. A. Navarro Cano, and T. Méndez 518 

Pérez (2001), Medio Ambiente y Empleo en la Sierra Minera de Cartagena–La Unión, Ed. 519 

Fund. SierraMinera. La Unión, 256 pp. 520 

Mayo, J. T., C. Yavuz, S. Yean, L. Cong, H. Shipley, W. Yu, J. Falkner, A. Kan, M. Tomson 521 

and V. L. Colvin (2007). The effect of nanocrystalline magnetite size on arsenic removal. 522 

Science and Technology of Advanced Materials, 8(1-2), 71-75. 523 

Moskowitz, B. M. (1981), Methods for estimating Curie temperatures of titanomaghemites from 524 

experimental Js-T data, Earth Planet. Sci. Lett., 53, 84–88. 525 

  22

www.murciatoday.com (07/01/2014) – “Offshore soundings as Portman Bay regeneration 526 

project moves forward”. (http://murciatoday.com/offshore-soundings-as-portman-bay-527 

regeneration-project-moves-forward_19778-a.html) 528 

Oen, I. S., J. C. Fernández, and J. I. Manteca (1975), The Lead-Zinc and Associated Ores of La 529 

Union Sierra de Cartagena , Spain, Econ. Geol., 70, 1259–1278. 530 

Oyarzun, R., J. I. Manteca Martínez, J. A. López García, and C. Carmona (2013), An account of 531 

the events that led to full bay infilling with sulfide tailings at Portman ( Spain ), and the 532 

search for “ black swans ” in a potential land reclamation scenario, Sci. Total Environ., 533 

454 - 455, 245 – 249. 534 

Peña, J. A., J. Manteca, P. Martínez-Pagán and T. Teixidó (2013). Magnetic gradient map of the 535 

mine tailings in Portman Bay. J. Appl. Geophys., 95(0), 115-120. 536 

Pike, C. R., A. P. Roberts and K. L. Verosub (1999). Characterizing interactions in fine 537 

magnetic particle systems using first order reversal curves. J. Appl. Phys, 85(2), 6660-538 

6667. 539 

Reed, S. J. B. (2010). Electron Microprobe Analysis and Scanning Electron Microscopy in 540 

Geology. Cambridge, Cambridge University Press. 541 

Roberts, A. P., C. R. Pike, and K. L. Verosub (2000), First-order reveral curve diagrams: a new 542 

tool for characterising the magnetic properties of nagtural samples, J. Geophys. Res., 102, 543 

28461–28475. 544 

Tauxe, L. (1998), Paleomagnetic principles and practice, edited by G. Nolet, Kluwer Academic 545 

Publishers, Dordrecht. 546 

Tauxe, L. (2009), Essentials of Paleomagnetism, University of California Press. 547 

Vera, J. A. (2004), Geología de España, edited by J. A. Vera, , 688. 548 

  23

549 

Figure Caption 550 

Figure 1. Location of the Portman Bay in the Iberian Peninsula (white dot in the upper and right panel) and the 551 

collected samples along the coastline of the Bay (lower and left panel. Source: Google Earth). 552 

Figure 2. Geology and mineral deposits of the Cartagena-La Unión Mining District (simplified after Manteca 553 

Martínez and Ovejero Zapino (1992)) 554 

Figure 3. Histograms of the low field magnetic susceptibility, κ at (300 A / m) for PP (a) and M (b) samples. 555 

Figure 4. Summary of analyzes with PP1B – A (0.20 – 0.30) sample. (a) Measured hysteresis loops (black) and after 556 

correction of the paramagnetic/diamagnetic contribution (red). (b) SIRM acquisition (up to 0.5 T) and SIRM back-557 

field curves. (c) Gradient Acquisition Plot (GAP) for the raw data, the computed 1, 2 and 3 components and the sum 558 

of components. (d) Thermomagnetic curve with the warming-up (red) and cooling-down (blue) values. Arrows 559 

indicate the direction in which the process occurs. The inset shows a zoom into the higher temperature range of the 560 

warming up curve. 561 

Figure 5. Summary of analyzes with M11 sample. (a) Hysteresis loops without paramagnetic/diamagnetic correction 562 

(black) and with such contribution (blue). (b) SIRM acquisition (up to 0.5 T) and SIRM back-field curves. (c) 563 

Gradient Acquisition Plot (GAP) for the raw data, the computed 1, 2 and 3 components and the sum of components. 564 

(d) Thermomagnetic curve with the warming-up (red) and cooling-down (blue) values. Arrows indicate the direction 565 

in which the process occurs. The inset shows a zoom into the higher temperature range of the warming up curve. 566 

Figure 6. (a) Day plot (Day et al., 1977) of PP and M samples and the corresponding theoretical mixing curves 567 

proposed by Dunlop (2002). The inset shows zoom of PP samples (b). FORC diagram of material from site PP1. 568 

Figure 7. Selection of images from PP samples obtained by SEM. Detection method of secondary electrons for a, b, c 569 

and f panels and backscattered electrons method for d and e panels. The samples were in each case: PP1B-A (30- 38), 570 

PP2-B (25-30), PP3B-A SUP, PP4-A-B and PP6C PP5P-B. The scales are at the lower and rigth part of each panel. 571 

Figure 8. Spatial distribution along the Bay of the hysteresis parameters κpara (a), κini (b), Ms (c), Mr (d) and Bc (e) and 572 

the parameters related with IRM acquisition Bcr (f), SIRM (g), B1/2 (h) and S-ratio (i). Red dots represent mean values 573 

for clustered samples in the Bay. Black bars represent the standard deviation. 574 

Figure 9. Correlation between the S-ratio (%) and Bcr (a) and B1/2 (b). Red squares represent mean values for PP 575 

samples and blue diamond represents mean value of the M samples, all with their corresponding standard deviation. 576 

  24

Correlation has been computed only for PP samples and the figures display the R2 is the coefficient of determination 577 

and p is the Pearson correlation coefficient. 578 

Figure 10. Comparison and correlation between hysteresis parameters κini (a), Mr (b) and Bc (c) of PP (red) and M 579 

(blue) samples. Black lines indicate the linear regression, R2 is the coefficient of determination and p is the Pearson 580 

correlation coefficient. PP samples represent mean values for clustered samples in the Bay. Black bars represent the 581 

standard deviation. Red dots represent mean values for clustered samples in the Bay. Black bars represent the 582 

standard deviation. 583 

Figure 11. Comparisons and correlations for Ms of PP samples. Comparisons with coarse (a), médium (b) and fine 584 

(c) fractions, in percentage, and with Pb (d), Cu (e) and As (f) contents in parts per million. Black lines: linear 585 

regression, R2: coefficient of determination and p: Pearson's correlation coefficient. 586 

Figure 12. Correlation between the two methods used to calculate the concentration of magnetite in PP samples. 587 

Black line: linear fit. R2: coefficient of determination and p: Pearson's correlation coefficient.. 588 

589 

Table Caption 590 

TABLE 1: Summary of the rock magnetic properties and magnetic characterization of samples from the Portman 591 

Bay Area (PP samples) and Mina Emilia (M samples). This includes saturation magnetization (Ms), saturation 592 

magnetization of the remanence (Mr), coercivity (Bc), coercivity of remanence (Bcr), initial low field susceptibility 593 

(Kinit), saturation of isothermal remanent magnetization (SIRM), median destructive field of the main modelled 594 

coercivity spectra (B1/2) and its corresponding dispersion parameter (DP) and S-ratio (%). 595 

TABLE 2: Percentage of magnetite (Mt) derived from magnetic measurements and comparison with the estimation 596 

derived from XRD in the total fraction and the grain sized fraction (coarse, medium and fine). Heavy metal content 597 

when measured 598 

Portman Bay

SPAIN

POR

TUG

AL

FRANCE44°0

42°0

40°0

38°0

36°0

44°0

42°0

40°0

38°0

36°0

-10°0 -8°0 -6°0 -4°0 -2°0 -0°0 2°0 4°0

-10°0 -8°0 -6°0 -4°0 -2°0 -0°0 2°0

El GorguelCreek Mediterranean Sea

Portman Bay

La Unión

Portman

Sierra de Cartagena

Emilia Open Pit

Cabezo RajaoMine

Mediterranean Sea

Portman Bay

La Unión

Portman

Sierra de Cartagena

Emilia Open Pit

La CrisolejaMine

: Fault: Road: Town

El GorguelCreek

Stratigraphic column

: Mines, open pit or underground

: Lower Nevado Filábrides (Paleozoic): Upper Nevado Filábrides (Permian-Triassic): Alpujárrides (Permian-Triassic): Dykes and felsic domes (Upper Miocene): Marine Sediments (Tortonian and Messinian)

: Coastline

1 km : Quaternary Sediments

(a) (b)

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

2.5

-500 0 500

m A( M

2 /kg)

B (mT)

-0.25

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

-500 0 500M

(A m

2 /kg)

B (mT)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.1 1.0 10.0 100.0 1000.0

tneidarG

B (mT)

Raw data

Sum ofcomponents

Component 1

Component 2

Component 3

0

2

4

6

8

10

12

14

0 200 400 600 800

M (A

m2 /k

g)

T (°C)

Warm-up

Cool-down

(b) (a)

(c) (d)

0

1

2

400 600 800

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-500 0 500

M (A

m2 /k

g)

B (mT)

-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

-500 0 500M

(A m

2 /kg)

B (mT)

0.0

0.1

0.1

0.2

0.2

0.3

0.1 1.0 10.0 100.0 1000.0

tneidarG

B (mT)

Raw data

Sum ofcomponents

Component 1

Component 2

Component 3

0

1

2

3

4

5

6

0 200 400 600 800

M (A

m2 /k

g)

T (°C)

Warm-up

Cool-down

(a) (b)

(c) (d)

0

1

2

400 600 800

2.2 2.4 2.6 2.82.0 3.00.08

0.10

0.12

0.14

Hc (T)

0.00 0.04 0.08 0.12

x10-28

4

0

-4

-8

Hu (

T)

SD

PSD

MD

SF=6

(a) (b)

-4

-2

0

2

4

6

8

PP1

PP1

BP

P2P

P2B

PP3

AP

P3B

PP4

PP5

PP6

kpa

ra(1

0-7 m

3/k

g )

02468

10121416

PP1

PP1

BP

P2P

P2B

PP3

AP

P3B

PP4

PP5

PP6

� ini

(10-

5 m

3/k

g)

0

4

8

12

16

20

PP1

PP1

BP

P2P

P2B

PP3

AP

P3B

PP4

PP5

PP6

Ms

(Am

2 /kg)

0.0

0.4

0.8

1.2

1.6

2.0

PP1

PP1

BP

P2P

P2B

PP3

AP

P3B

PP4

PP5

PP6

Mr(A

m2 /k

g)

10

11

12

13

PP1

PP1

BP

P2P

P2B

PP3

AP

P3B

PP4

PP5

PP6

Bc

(mT)

22

24

26

28

30

32

PP1

PP1

BP

P2P

P2B

PP3

AP

P3B

PP4

PP5

PP6

Bcr

(mT)

0.0

0.4

0.8

1.2

1.6

2.0

PP1

PP1

BP

P2A

PP2

BP

P3A

PP3

BP

P4P

P5P

P6

SIR

M (A

m2 /k

g)

30

34

38

42

46

PP1

PP1

BP

P2A

PP2

BP

P3A

PP3

BP

P4P

P5P

P6

B1/

2(m

T)

(a) (b) (c)

(d) (e) (f)

(g) (h)

0 2 4 6 8

10 12 14

PP1

PP1

BP

P2A

PP2

BP

P3A

PP3

BP

P4P

P5P

P6

S-ra

tio(%

)(i )

y = 0.666x + 22.18 R = 0.8376

p=0.8373

0

5

10

15

20

25

30

0 2 4 6 8 10 12 14

Bcr

(mT)

S-ratio (%)

y = 1.7427x + 23.861 R = 0.8647

p=0.9299

0 5

10 15 20 25 30 35 40 45 50

0 2 4 6 8 10 12 14 B

1/2

(mT)

S-ratio (%)

(a) (b)

R² = 0.9549

0

4

8

12

16

0 10 20 30

� ini

( 10-

5 m

3/k

g)

Magnetite %

PP M

R² = 0.9873

0.0

0.4

0.8

1.2

1.6

2.0

0 10 20 30

Mr (

Am2 /k

g)

Magnetite %

PP M

10

12

14

16

0 10 20 30

Bc

(mT)

Magnetite %

PP M

(a) (b) (c)

p = 0.9911 p = 0.9640

R² = 0.7154

01020304050607080

0 10 20 30

C %

Ms (A m2/kg)

R² = 0.5016

0

10

20

30

40

50

60

70

0 10 20 30M

%Ms (A m2/kg)

0

5

10

15

20

25

30

35

0 10 20 30

F %

Ms (A m2/kg)

R² = 0.6287

0

1000

2000

3000

4000

5000

0 10 20

)mpp( bP

Ms (A m2/kg)

0

50

100

150

200

250

0 5 10 15 20

Cu

(ppm

)

Ms (A m2/kg)

0

500

1000

1500

2000

2500

0 5 10 15 20

As (p

pm)

Ms (A m2/kg)

(a) (b) (c)

(d) (e) (f)

p = -0.8458

p = 0.7082

p = 0.7929