Paleoclimatic implications of micromorphic features of a polygenetic soil in the Monegros Desert ( NE-Spain )

Pedofeatures can be repositories of information about soil forming factors such as climate. The aim of this work is to provide a model of interpretation of a polygenetic soil in the Monegros desert (Ebro Basin, NE-Spain) and its relationship to environmental changes during the Quaternary. To achieve this goal, the physical, chemical, mineralogical and especially the micromorphic pedofeatures of this profile were studied. Carbonate accumulations extend into all of the horizons of the profile. The paleosol has a thick petrocalcic horizon at the top, with a massive-laminar structure comprising layers of micrite and sparite that sometimes form pendants. Towards its base, the petrocalcic horizon contains a spaced framework of orthic micrite nodules packed between relatively pure micritic laminar bands. Below the petrocalcic horizon, coatings and infillings of microcrystalline calcite occur in old channels, and soft concretions (some of them geodic) indicate an in situ accumulation process (Bkc, calcic horizon). Another calcic horizon with orthic nodules of calcite, impregnative and diffuse (Ckc), is present at the bottom part of the profile. Between the two nodular calcic horizons, two recarbonated argic horizons are found (Btkc and Btk) with coarse orthic nodules of dense micrite superimposed on textural pedofeatures. These textural micromorphic pedofeatures are: (1) interbedded microlaminated clay pockets not associated with current or past pores and (2) microlaminated clay and silt (dusty clay) present as weakly oriented coatings on channel walls. Reduction pedofeatures are associated with textural ones: (1) coatings of manganese oxides around pore channels and cracks, and (2) nodules of manganese and iron oxides within the peds. The presence of calcic horizons alternating with argic horizons, all positioned below the petrocalcic horizon, confirm fluctuations in paleohydrological conditions in the Pleistocene. Its presence indicates that the oldest soil corresponds to a Calcic Luvisol-like pedotype, which is overlain by an Haplic Calcisol-like pedotype and this, in turn, by a Petric Calcisol-like pedotype. This superposition of profiles indicates, within the mentioned climatic changes, a tendency towards increasing dryness during the Pleistocene in the semiarid Ebro Valley. AUTHORS


Introduction
During the Quaternary, the landscape has been subjected to a wide variation of environmental conditions which have left distinct footprints in soils as a result of past processes.This is the case of polygenetic or polycyclic soils, a composite of horizons from more than one soilforming event (Chesworth 2008;Fedoroff et al. 2010).Polygenetic soils can be also considered a complex of soils formed in pre-Holocene environments different from today, subsequently subjected to processes of burial by younger sediments and exhumation by erosion (Nettleton et al. 2000).The study of paleosols is a tool in the reconstruction of the paleoclimate and regional landscape dynamics on the basis of the processes inferred from soil properties (Hamer et al. 2007;Brock and Buck 2009;Markovic et al. 2011).Paleopedology is a useful proxy to reconstruct paleoenvironmental conditions but is still an immature field (Sheldon and Tabor 2009).Moreover, there are few studies on paleopedogenic processes in semiarid Mediterranean regions because the soils are generally weakly developed and rarely preserved due to low plant cover and high erosion rates (Zielhofer et al. 2009).
The Ebro Basin, the northernmost semiarid region in Europe, is an area with high sensitivity to environmental changes that contains excellent morphosedimentary records of Quaternary climate oscillations (Lewis et al. 2009;Sancho et al. 2011).Previous research on paleoenvironmental and paleoclimatic reconstruction in continental deposits of Ebro Basin has been focused on sedimentological and paleontological studies (González-Sampériz et al. 2010;Sancho et al. 2011;Moreno et al. 2012).However, the significant wealth of information that can be acquired from processes in paleosols from semiarid regions is only now starting to come to light (Hamer et al. 2007;Brock and Buck 2009;Meléndez et al. 2011).In this sense, some soil processes such as secondary carbonate precipitation and clay illuviation can be repositories of information about past soil forming factors (Brock and Buck 2009).Petrocalcic horizons (calcrete, also known by popular names such as caliche, mallacán, calicanto, cervell de gat, etc.) are very frequent in semiarid regions such as the Ebro Basin, forming on old stable surfaces where soils display progressive and systematic patterns of carbonate accumulation over time (Gómez-Miguel 2005;Sancho et al. 2008;Lewis et al. 2009).Although petrocalcic formation has generally been understood to be a continuous process with no interruptions by sedimentation or erosion (Gile et al. 1966;Machette 1985;Nettleton et al. 1991), non-linear models have also been proposed (Huggett 1998;Meléndez et al. 2011).In calcretes evolved on Early Pleistocene terraces of the Segre River, a tributary of the Ebro, alternating layers of carbonates and illuvial clay provide unequivocal evidence of the various phases of successive secondary carbonate precipitation, and it has been interpreted as the reflex of climatic changes during the Quaternary (Badía et al. 2009a).Beside carbonation, clay illuviation is another common pedofeature in semiarid to subhumid regions in relatively stable landforms (Dorronsoro and Alonso 1994;Khormali et al. 2003;Ortiz et al. 2002;Roquero et al. 1999;Dan et al. 1981). However, Nettleton et al. (2000) suggested that smectitic argillic horizons in the Argid sub-order (Soil Survey Staff 2010) are Pleistocene relicts related to a wetter climate.Clay translocation, also called argilluviation or lessivage, is traditionally considered to be promoted by a low organic matter content, coarse to medium textured soils, a smectite-dominated clay fraction and, especially, a soil pH of 5.5 to 6.5 with no Ca 2+ , Mg 2+ or Al 3+ to act as binding agents (Khormali et al. 2003;Chesworth 2008;Quénard et al. 2011).Therefore, decarbonation and desaturation seem to be preliminary steps to clay transfer from an elluvial horizon to an illuvial one, although Poch et al. (2012) suggested that spatial differentiation of soil environments in the profile could allow the temporal coexistence of clay illuviation and calcification.
The aim of this work is to provide a model of a polycyclic soil in the Monegros desert (Ebro Basin, NE-Spain) which formed during the Pleistocene.To achieve this goal, the physical, chemical, mineralogical, and morphological properties of this complex profile were studied.Particular emphasis is placed on the micromorphology, particularly calcification and textural pedofeatures, and its relationship to environmental changes.

Geologic setting and background
The studied paleosol was sampled in cliff edges of the Saso (= mesa) de las Fitas landform (Figure 1), near the village of Sariñena (Monegros county, NE-Spain).The region has a semiarid Mediterranean climate, with a mean annual precipitation (MAP) about 400 mm and a mean annual temperature (MAT) of 14.5 °C; the ratio of rainfall to mean annual reference crop evapotranspiration (P/ETo) is about 0.35.The paleosol is placed at the north-central Tertiary Ebro Basin, the southern foreland basin of the Pyrenean mountain range.The sedimentary regime of this Basin persisted until the end of the late Miocene (between 12.5 and 8.5 Ma) when the internal drainage opened due to aggradation of lacustrine environments and extension into the western Mediterranean basin (García-Castellanos et al. 2003).From then, erosive activity and fluvial incision in the Ebro basin has persisted throughout the Quaternary, as shown by the extensive systems of stepped terraces along its rivers (Lewis et al. 2009).The studied paleosol has been de- veloped on the highest and oldest alluvial fan (Qt1) of a well-preserved stepped terrace system of the Alcanadre River.The Qt1 strath terrace is now an encrusted surface (calcretised), 170 m above the present channel (relief inversion) and it has a broad fan shape with a mean slope reaching 1%.It forms the water divide between the Alcanadre and Cinca Rivers.This highest surface is considered to be related to the initial stages of the configuration of the Ebro Basin's Quaternary exorrheic fluvial network (Meléndez et al. 2011).The Qt2 to Qt4 terraces are missing in this study area (Figure 1), but are clearly recognised in northernmost sectors (Rodríguez 1986).Paleomagnetic analysis has shown the Matuyama-Brunhes (779 ka) transition to be located between terraces Qt2 and Qt3 whilst terrace Qt1 is tentatively related to the Jaramillo (around 1000 ka) magnetozone (Lewis et al. 2009).The soil parent materials, coming from the External Pyrenees, are successive fine calcareous detrital sediments (above 4.5 m thick) overlying rounded gravels of limestones, sandstones and conglomerates embedded in a sand matrix.
The paleosol was described in the field according to FAO procedures (FAO 2006), and detailed morphological data were recorded for each horizon.Bulk horizon samples were collected, air-dried, crushed and sieved to 2 mm to remove coarse fragments.These bulk samples were used for physical, chemical and mineralogical characterization.In addition, undisturbed oriented clod samples were carefully extracted to maintain the original structure of the soil, placed in containers and transported to the laboratory for thin section preparation.Peds were also airdried before being heated to 90 ºC, impregnated under vacuum with epoxy resin, mounted on glass slides and ground to 30 μmm.Thin sections were studied using a petrographic microscope under both plane-and cross-polarized light.The study will be conducted following a systematic approach where types of pedofeatures will be assigned to a given soil formation event following a hierarchy (Fedoroff et al. 2010).

Chemical and physical properties analysis
Particle size distribution was determined using the pipette method, after the removal of organic matter using H 2 O 2 and with Na-hexametaphosphate used as a dispersing agent (Gee and Bauder 1986).Using detailed texture data (USDA subdivisions) from all horizons, all fractions (%) of the fine earth were recalculated removing clay; if one horizon deviated more than 5% from another, a lithological discontinuity was diagnosed.
Soil pH (1:2.5 ratio in H 2 O) was determined using a glass electrode (McLean 1982).Total carbonate content was measured volumetrically (with a calcimeter) after treatment with 6N hydrochloric acid (Nelson 1982).The electrical conductivity (ECe) and the sodium adsorption ratio (SAR) were measured in the saturated paste extract (Rhoades 1982a).Total soil organic C was determined by the wet oxidation method using the van Bemmelen factor (1.724) to estimate organic matter (Nelson and Sommers 1982).The cation-exchange capacity (CEC) was determined with the NH 4 + ion retention after washing the soil sample with a solution of NH 4 OAc 1N at pH 7 (Rhoades 1982b).

Mineralogical characterization
The mineralogical composition of the clay fraction (≤ 2 μm) of selected horizons (Ah, 2Bkc, 2Btk 4Ckc and 5C) was identified by X-ray powder diffraction (XRD) on a Philips X'Pert diffractometer with graphite-monochromated Cu Kα radiation.XRD patterns were obtained from each sample using: a) random powder mounts, and the following oriented aggregates; b) air dried, c) ethylene glycol-solvated, d) heated at 300 ºC and e) 550 ºC for 3 h.Semiquantitative estimates of quartz, goethite and total phyllosilicates were obtained from the random powder patterns, integrating the area of the maxima diffraction at 0.424, 0.416 and 0.444 nm, respectively, and using the mineral intensity factors reported by Schultz (1964).Approximate abundances of illite, smectite and chlorite were obtained in a similar way, using the oriented aggregate patterns (peak areas at 9.98 nm for illite, 1.65 nm for smectite in ethylene glycol-solvated aggregate and 1.42 nm for chlorite in 550 ºC heated oriented aggregate).

Results
This section gives the description of the field characteristics, chemical and physical properties, clay mineralogy and the micromorphological features of the paleosol.

Soil morphology
The Ah horizon, up to 20 cm thick, has a brown color in contrast to the underlying horizons: nearly white for Bkm horizons and very pale brown for the rest of horizons with the exception of Bt horizons that become to reddish yellow (Table 1).In particular, the petrocalcic horizon consists of 2 subhorizons, a laminar one on the top (Bkm1) and underneath a nodular one (Bkm2), also named pisolithic according to Meléndez et al (2011).The upper petrocalcic subhorizon (Bkm1) is composed of individual laminae (some mm thick) that range in color (dry) from white (N8) to pinkish white (7.5YR 8/2), very pale brown (7.5YR 8/3) and light brown (7.5YR 6/3).In contrast Bkm2 has a very pale brown (10YR 8/3) overall color with white veins, and occasional aggregates with light brown (10YR 6/4) color.The limits between horizons are nearly planar or slightly wavy and clear, with the exception of the petrocalcic horizon that has abrupt limits.The type of structure in Ah horizon is granular, massive in Bkm (petrocalcic) horizons, blocky in Bk (calcic) and prismatic in Bt (argic) horizons, overall with a strong grade or development.Argic horizons have slickensides on the faces of the aggregates.The deepest soil layers (5C, 6C) are structureless and non-coherent because of the sandy texture and stoniness (Table 1).

[ PALEOCLIMATIC IMPLICATIONS OF MICROMORPHIC FEATURES OF A POLYGENETIC SOIL
IN THE MONEGROS DESERT (NE-SPAIN) ]

Chemical and physical properties
The soil pH is basic in all horizons, increasing with depth up to 9.0 in bottom horizons.The soil profile is highly calcareous throughout with a maximum in petrocalcic horizons where is higher than 80% w/w.The electrical conductivity (ECe) and sodium adsorption rate (SAR) are very low on the topsoil reaching a maximum at 4 m depth (ECe=17 dS/m; SAR=8 (mmol/L) 0.5 ).Organic matter content is negligible below the Ah horizon.
The coarse element content is very low and they have a different lithology: reworked calcrete clasts in the Ah-horizon, hard concretions in different Bkc and Ckc horizons and only gravels in bottom C-layers, the content of which increases with depth (Table 2).Pebble and cobble-sized rounded gravels with intercalated sand lenses are only found at the bottom of the profile (below 5-m depth) and the upper sediments are mainly silt and clay particles, which suggests a progressive reduction of the sediment transport capacity.The abrupt changes in the stoniness, in the sand and silt fractions, and in the clay mineralogy allowed us to define at least 6 periods of parent material deposition (lithological discontinuities).The textural class is loam in the Ah horizon, silt loam in 2Bkc, 3BC, 4Ckc, 4Ck horizons, silty clay loam in the 2Btkc and 2Btk horizons, and it becomes sandy in the deepest C-layers (Table 3).
Table 2. Some chemical and physical properties of the paleosol

Mineralogical characterization
The semi-quantitative mineralogical composition (Table 4) indicated that dioctahedral mica (illite) is by far the most abundant phyllosilicate (ranging from 40 to 70%) reaching the highest content in the upper horizon Ah).Smectite and chlorite are present in many horizons in small proportions (for both cases not more than 12%).Goethite has also been detected in some horizons but in a very small proportion (traces, ≤ 2%), although enough to redden the horizons.
The mineralogical composition of the 2Bkc and 2Btk horizons was very similar, but not the rest of horizons.This confirms the presence of the above-mentioned discontinuities corresponding to different sedimentary processes (Table 4).
XRD results of the ≤ 2 μm fraction of selected horizons showed the absence of fibrous minerals (palygorskite and sepiolite) in all samples.
The XRD patterns of the 2Bkc horizon are shown as an example (Figure 3).lllite is the most common mineral component of the clay fraction in soils of the region (Badía et al. 2009b).Illite and chlorite phyllosilicates are considered to be inherited from the parent material.Smectites could be either inherited or result from the transformation of illite in an environment with poor drainage, high pH and abundance of basic cations (Wilson 1999).

Micromorphology
Some common trends can be found in all horizons, such as the absence of organic components below the Ah horizon and the crystallitic b-fabric due to the presence of micrite and microsparite (Table 5).Calcite accumulations, textural pedofeatures, and manganese coatings are frequent along the profile (Table 6).Petrocalcic horizons (stage V, Machette 1985) have an average thickness of 255 cm and they were divided into two subhorizons in the field based on their hardness and their morphology.The Bkm1 horizon (about 1-m thick) is extremely hard and dense (2.3-2.4 g/cm 3 ) because CaCO 3 plugs the horizon, with the highest carbonate content (ca 88% w/w) of the profile, and a massive-laminar structure (Figure 4A).Under the microscope, the laminar petrocalcic horizon consists of an alternation of gently undulating bands of micrite and sparite that sometimes form layered pendants, also known as vadose microstalactitic calcite cements (Figure 4B).Pendant laminae are mostly vertical, consistent with the direction of gravity.The pendants show an internal lamination, lighter than the outer coating.These morphologies are commonly recognized in petrocalcic laminar horizons of the zone (Meléndez et al 2011).Although no EDAX analyses have been performed, the layering seems to correspond to different crystal size owing to different precipitation conditions.Within the laminar petrocalcic horizon there are variety of features such as spherical peds (Figure 4C).These peds comprise a thin coating of clay-sized material around a core, which includes pores and mineral grains sometimes with laminar structure.These are not considered to be ooids or pisoliths, since these peds are more clayey that the surrounding carbonate matrix, and are not formed as concretions (precipitation around a nucleus), but by some type of argilloturbation (internal granostriation).
The Bkm2 horizon (150 cm thick) is more porous (2.0 g/cm 3 ) and has a slightly lower carbonate content (ca 84% w/w) than the overlying laminar Bkm1 horizon; it shows a massive appearance with nodular structures inside (Figure 4D).The Bkm2 horizon has a spaced framework of orthic micrite nodules packed between relatively pure micritic laminar layers (Figure 4E).This type of nodule, orthic and without a concentric internal fabric, suggests a progressive enrichment of calcium carbonate in fairly constant conditions.Underlying those petrocalcic horizons is a calcic horizon (2Bkc).Different calcite accumulation types appear in the 2Bkc horizon such as coatings, infillings and orthic nodules, which indicate an in situ calcite accumulation process.The connectivity of rounded gravel sized nodules of calcite suggests that they were formed around roots (and thus are also named rhizoconcretions or rhizoliths, Figures 5A and 5B).Under this nodular calcic horizon, two subhorizons (2Btkc and 2Btk) show evidence of strong clay enrichment.
Both have about 30% of clay particles (1.25 times higher than the upper horizon), a reddish colour, goethite traces and an angular blocky (Figure 5A) or prismatic structure (Figure 5C), defining characteristics of an argic horizon (argillic by STS).Argic horizons are recarbonated, as shown by the presence of micrite accumulations in nodular forms (Figure 5B), especially in the upper subhorizon (2Btkc).The presence of smectites and chlorites (both around 8%) could account for the strong prismatic structure, soil cracking and some argilloturbation.A common feature in the argic horizons, and also the underlying horizons (from 315 to 425 cm depth), is the presence of Fe/Mn accumulations (Figures 5D, E) that, in the field, have been described with percentages varying from about 5% to 40% by volume.Micromorphologically, two types of Fe/ Mn accumulations (gleyed horizons) were distinguished: (1) hypo-and quasi-coatings related to pore channels, and (2) dendritic aggregate nodules within the peds (Figure 5D).These morphologies indicate moderate to strong redox conditions.Two kinds of textural pedofeatures can be distinguished in the argic horizons: (1) interbedded microlaminated clay pockets (Figure 5F), sometimes fragmented and distorted by micrite accumulations and (2) few fine clay coatings along vertical cracks and around calcite channel infillings.
A new calcic horizon appears at the bottom of the profile (4Ckc), again with rhizoconcretions but in this case not as rounded (Figure 6A) as in the upper nodular calcic horizon (2Bkc) and the recarbonated argic horizon (2Btkc).Moreover, coatings of microlaminated clay and silt (dusty clay) can be found until 4.5 m soil depth in some macropores.These dusty clay coatings appear covered by micrite (Figure 6B), which demonstrates the recarbonatation process occurred after clay illuviation.et al. 2011).Moreover, in this case, the petrocalcic nodular horizon is underneath a nodular calcic horizon.The morphology of the nodules (orthic and without internal fabric) and their arrangement in columns suggest that they were formed around roots (rhizoconcretions or rhizoliths) in fairly constant conditions.In any case these conditions of pedogenesis would alternate with erosion and sedimentation events (Meléndez et al. 2011).
The presence of spherical peds within the laminar petrocalcic horizon could be evidence of that succession of processes; their formation according to Alonso-Zarza and Jones ( 2007) requires an initial fragmentation of the groundmass and a posterior coating formation by the biogenic activity of roots and associated microorganisms.Their origin is also interpreted to be old excrements of earthworms or molluscs, which were later cemented by calcite (Kooistra and Pulleman, 2010).
Below the petrocalcic and calcic horizons a thick argic horizon (with two subhorizons: 2Btkc and 2Btk) has been described, according to the macroscopic or microscopic characteristics.At present this horizon has a high carbonate content, both in the fine earth fraction and in the clay fraction (Tables 2 and 4).Because polyvalent cations cause flocculation, the argic horizon must have formed after decarbonation and therefore it is most likely to be a paleofeature related to a wetter climate than present (Khormali et al. 2003;Quénard et al. 2011)  subhorizontal areas not associated with current or past pores.They can be better interpreted as original, non-recarbonated zones formed as the result of a more recent decarbonation process.
Similarly, sparite and recarbonated argillans were found in the oldest terrace (Early Pleistocene) of the Segre River (Badía et al. 2009a) that were interpreted as pedological paleofeatures caused by climate changes during the Quaternary.Analogously, Khormali et al. (2003) found clay coatings superimposed on carbonate coatings in argic horizons of calcareous soils.
The clay coatings, in turn, were covered by calcite, indicating recalcification.They considered that argic horizons were formed in a previous less arid climate than the current semiarid regions of southern Iran (with a MAP about 340 mm/yr and a P/ETo ratio between 0.3-0.5).
In the field, shiny surfaces considered to be clay skins were seen in the recarbonated argic horizons (Table 5).However, in thin sections, microlaminated clay coatings rarely can be found.The main factor responsible for the absence of clay coatings in old argic horizons is their physical disturbance by shrinking and swelling in soils with high clay contents and periodic moist and dry phases (Khormali et al. 2003;Mermut and Pape 1971).Instead, intercalative microfacies are abundant in the recarbonated argic horizons.Intercalative microfacies formation can be related to soil formation process in sodic soils (Rodríguez et al. 1990;Badía et al. 2011), the possibility of which must be discarded in this soil environment, or to waterlogging events (Fedoroff and Courty 2012), and is more likely in this situation.According to these authors, water with high energy percolates through the surface horizon along macropores, which produces a turbulent flow and therefore a micro-erosion or degradation at the top of the argic horizon, especially if this horizon is shallower (ie by previous exhumation).This energy decreases with depth, leading to progressive settling of microlaminated pockets, especially in old cracks related to the swelling activity of the smectites during dry/dessication periods.Moreover although carbonate may prevent the dispersion of clay, the movement of already dispersed clay (and silt) can take place if pores are large enough (Pazos 1990).
Another interpretation that may explain the presence of intercalative microfacies is related to the decrease of redox potential in wet conditions that transforms the ferric iron into ferrous iron that was removed in pseudo-soluble forms.Destruction of the ferric bindings leads to a structural collapse resulting in a loose packing of soil constituents.The partial ferric ion removal would favour mass internal displacement.In calcareous environments, however, this process would be more seldom.
Reducing conditions had to be in the soil at some time that resulted in a Fe 2+ and Mn 2+ ion increase; these conditions could coincide with a decrease in soil reaction and calcite depletion, allowing clay illuviation.Afterwards, the pedoclimate changed to oxidized, with warmer and drier conditions that oxidated Fe 2+ ions via solution to form goethite that gives the horizons their reddish yellow colour.With new well-aerated conditions, also Mn solubility decreases and Mn coatings and nodules can be formed.Moreover, Mn solubility decreases with an increase in soil pH by a factor of 10 for each pH unit.This increase in pH would be given by aggradation of calcareous sediments above soil or new inputs of calcite saturated water solution, as we mentioned earlier (Ortiz et al. 2002;Lequy et al. 2012).

Environmental conditions and time for paleosol formation
The carbonate enrichment in the subsoil resulting in nodular calcic horizons or petrocalcic horizons represents again an unequivocal indicator for autochthonous soil formation (Zielhofer et al. 2009).The large thickness of both overlaid petrocalcic horizons, its high carbonates content (> 80%), and its strong degree of induration could express hundreds-of-thousands to millions of years of development in arid environments (Machette 1985;Brock and Buck 2009;Monger et al. 2009).Because the depth of the calcic horizons is correlated with the depth of water penetration (Dan et al. 1981;Sheldon and Tabor 2009), the relatively shallow formation of Bkm horizons could indicate that the climate was progressively drier.Some authors (Birkeland 1999;Dan et al. 1981) consider that the presence of calcrete infers a paleoprecipitation below 500 mm yr -1 .
But if the unique source of calcium carbonated accumulated in current thick subsurface horizons is the leaching from topsoil horizons, the current Ah horizon (20 cm thick) cannot explain the intense carbonate accumulation.Erosion of former upper horizons as well as inputs of allochthonous carbonates in the alluvial fan, through dust inputs (Lequy et al. 2012) or the subsurface flow (Ortiz et al. 2002;Poch et al. 2012), should be also considered.In the studied paleosol, the occurrence of vertical layered pendants or vadose microstalactitic calcite cements reveals a vertical and slow water flow more than an oblique flow (Figure 4B).
Beside carbonation, clay illuviation is another common pedofeature in semiarid to subhumid regions in relatively stable landforms (Dan et al. 1981;Dorronsoro and Alonso 1994;Roquero et al. 1999;Gallardo et al. 2002;Ortiz et al. 2002;Khormali et al. 2003).Hamer et al. ( 2007) considered that argic horizon formation (Luvisol-like pedotype) was dominant within the Late Oligocene-Early Miocene in the Ebro Basin; they estimated that an argic horizon (Bt) occurred over 10 ka with a MAT about 10-14 ºC (± 4 ºC) and a MAP of 560-830 mm yr -1 (± 200 mm yr 1 ).Similarly, Quénard et al. (2011) estimated that the complete decarbonation and desaturation (prior to argilluviation or lessivage) of loess (with 20% carbonate content) needs over 13.6 ka for an annual effective rainfall (rainfall-evapotranspiration) about 150 mm.After that, the lessivage takes place during a similar period to form an argic horizon (Finke and Hutson 2008).Also in alluvial sequences of the mid-Medjerda floodplain (Northern Tunisia), Zielhofer et al. (2009) found argic horizons with many distinct clay films on ped faces above nodular calcic horizons on Late Pleistocene surfaces (Calcic Luvisols, Chromic); they estimated that between 10 and 40 ka are required to develop this pedotype.Intense processes of clay illuviation, rubification and carbonation, successively superimposed, has been found in Early Pleistocene surfaces in Spain (Roquero et al. 1999;Gallardo et al. 2002;Ortiz et al. 2002).Some Petric Cutanic Luvisols (Petrocalcic Palexeralfs by STS) can be still found in some preserved Early Pleistocene glacis of Somontano, in the north of the Monegros Desert and with a wetter current pedoclimate (Badía et al. 2009b It is deduced that soils converge towards very specific pedofeatures, and therefore pedotypes (Table 7), which depend on the duration of pedogenic processes, geomorphological stability, and climate changes during the Pleistocene in Mediterranean areas (Gónzalez-Sampériz 2004;Markovic et al. 2011;Moreno et al. 2012).
The described paleosol, with a Calcic Luvisollike pedotype buried by a Petric Calcisol-like pedotype, suggests that, despite climatic periodic oscillations, the initially moderately humid climate tends to a progressively drier one.

Model of soil evolution
The pedofeatures and the organization of soil horizons within this paleosol provide information on the Quaternary environmental changes required for its formation, and allow the proposal of a model of polygenetic soil evolution (Figure 7).These are still to be dated.
In the first stages of organisation of the exorheic fluvial network of the Ebro Basin (Pliocene), high sedimentation rates deposited coarse gravels and sand without signs of pedogenesis

Acknowledgements
This project was supported by the Instituto de Estudios Altaoragoneses (Diputación de Huesca).
The suggestions of editors and anonymous reviewers have greatly improved the original manuscript.Also, we thank José Antonio Cuchí who guided us in the first profile visits and stressed the importance of its study many years ago.

Figure 1 .
Figure 1.Location of the studied profile, developed on the highest and oldest alluvial fan (Qt1) of a stepped terrace system of Alcanadre River (Ebro Basin, NE-Spain).

[
PALEOCLIMATIC IMPLICATIONS OF MICROMORPHIC FEATURES OF A POLYGENETIC SOIL IN THE MONEGROS DESERT (NE-SPAIN) ]

Figure 2 .
Figure 2. Section of the selected profile of Saso (= mesa) de las Fitas showing the arrangement of the different horizons from petrocalcic (at the top) to the gravels (at the bottom) and picture showing the spatial continuity of soil horizons and gravel layers.

Figure 4 .Figure 5 .
Figure 4. Pedofeatures in the upper part of the paleosol.A: Section of upper petrocalcic horizon (Bkm1) showing small-scale wavy layers (laminar one or bacon steak like).B: Photomicrograph of the alternation of micrite layers and vadose microstalactitic sparite cement in Bkm1 horizon.C: Photomicrograph of spherical peds in the laminar petrocalcic horizon (Bkm1).D: Section of lower petrocalcic horizon (Bkm2), nodular type.E: Spaced framework of orthic micrite nodules packed between relatively pure micritic layers in Bkm2 horizon.F: micritic laminar bands within sparite intercalations in Bkm2 horizon.All photomicrographs (B, C, E, F) in plane-polarized light.

Figure 6 .
Figure 6.Pedofeatures in the lower part of the profile.A: Irregular carbonate rhizoconcretions in 4Ckc horizon.B: Photomicrograph in plane-polarized light showing a macropore coated of microlaminated clay and silt (dusty clay) covered by micritic calcite (5C horizon).

Figure 7 .
Figure 7. Proposed evolution model for the Saso de las Fitas paleosol with main soil processes and features.

Table 5 .
Main micromorphological properties of the groundmass of the studied horizons

Table 6 .
Micromorphological pedofeatures of the studied horizonsCoatings and infillings of microcrystalline calcite in old channels and orthic nodules and soft concretions (some geodic), which indicate an in situ accumulation process.Scarce hypo-and quasi-coatings of manganese oxides around pore channels and cracks.Some decarbonated zones.Microlaminated pockets of clay intercalations fragmented by vertical cracks or distorted by nodules of micrite.Decarbonated zones associated with sorted intercalations.Coarse orthic nodules of dense micrite superimposed on textural pedofeatures.Hypo-and quasi-coatings of manganese oxides around pore channels and cracks.
2Btk 365Interbedded microlaminated clay pockets, not associated with current or past pores Impregnative nodules of micrite, rounded and lobulated, filling old channels.Hypo-coatings of manganese and dendritic aggregate nodules of manganese and iron oxides.3BC405Microlaminatedclay and silt as coatings on channel walls.Sorted wavy intercalations, with a subhorizontal orientation.Dendritic aggregate nodules of manganese within the peds and hypo-and quasi-coatings of manganese oxides around pore channels. ).