Soil catena along gypseous woodland in the middle Ebro Basin : soil properties and micromorphology relationships

Gypsisols, mainly distributed in arid lands, support a key economic activity and have attracted a lot of scientific interest due to their particular physical and chemical properties. For example, Gypsisols show a high erodibility, low fertility and a variable water holding capacity that can be attributed to different gypsum particle sizes. This study aims to describe some representative Gypsisols from the middle Ebro Basin. Five representative soil profiles (mainly Gypsisols by WRB) were selected and sampled at different positions along a hillside where soils where developed on gyprock. Furthemore, it links micromorphological properties with soil water retention. Soils have a dominant loamy texture, more rarely stoney. Gypsum is abundant in all soil profiles, ranging from 6 to 84% with minimum values in Ah horizons and maximum in By and Cy. The soils have a low level of salinity and a very low cation exchange capacity (CEC). The soil organic matter (SOM) is medium or abundant in the Ah horizons, otherwise it is low. Soil aggregate stability (SAS) is related significantly and positively with SOM and porosity, which is also positively related with moisture retention at field capacity and saturation humidity. However, there is no significant correlation between porosity and permanent wilting point (PWP). Soil water retention is dependant on the gypsum percentage and textural class. Low levels of gypsum have no influence on water retention, but high gypsum levels (> 60%) enhance the field capacity (FC) and decrease PWP, especially when the gypsum is microcrystalline. Gypsum levels between 40 and 60% also increase available water contents (AWC) due to a decrease in PWP. There is a positive and significant correlation between PWP and FC in Gypsisols, except for those which are loamy and have gypsum values over 40%. The higher available water capacity (AWC) than expected is related to microcrystalline gypsum, predominant in the studied soils. These high AWC values are counteracted by an increasingly irregular pore space not suitable for root growth. All these cited characteristics result in a low fertility, influenced by the weather and the human impact, which deforested the highest part of these mountains in the Middle Ages.

Gypsisols have some peculiarities in relation to their physical and chemical behaviour, causing low fertility and lack of structure (Mashali 1996).
From the viewpoint of soil moisture retention, some authors report very low values (Herrero 2005), while others report the opposite.This can be attributed to different gypsum particle size and to the degree of mixing of the gypsum infilling with the groundmass (Poch et al. 1998).
The objective of this study is to describe some representative Gypsisols (IUSS 2007) from this catena (Gypsic Haploxerept in the Soil Taxonomy System -SSS 2010) using a chemical and physical approach, and to examine the water retention capacity of the horizons using micromorphological analysis.

Study Area
The studied soils were located at different hillslope positions in the Castejón Mountains (NE-Spain).These mountains lie in a NE-SW direction, and divide the basins of the Gállego River in the East and the Arba River in the West.The study was carried out in Western part of Castejón Mountains, on the left side of Ebro River (Figure 1).The geomorphological context consists of ravines, which run from the top of the mountain to the flat areas where sediments are deposited.These ravines have been mainly eroded by water.
Between the ravines, the geomorphology unit found is the slope, with similar characteristics to the one selected in this study.
The climate is characterized by two long dry periods in summer and winter.The average rainfall is 428 mm/year, the average annual temperature is 13.9 ºC and the evapotranspiration is 1244 mm according to DGA (2004).Wind events with gusts over 30 m s −1 are common in the area (Cuadrats Prats 2004).The soil temperature regime is mesic (Soil Survey Staff 2010) while the moisture regime is xeric in head-slope and aridic in the rest of the profiles (Jarauta and Porta 1990).
and Mediterranean false-brome (Brachypodium retusum (Pers.)Beauv.).There are some protection zones for birds close to this area (SPA, ES0000293) and the environment is protected (SCI, ES2430080) and included on "Natura 2000" network.

Materials and Methods
Five profiles were sampled along a slope and labeled according to their position as: head-slope, shoulder-slope, back-slope, foot-slope and toeslope profiles.
Soil samples were collected for physical and chemical analyses.The morphological properties of each horizon were described following FAO methodology (2006): color (dry and moist), consistence and accumulations.The laboratory analyses were carried out using the fine earth fraction (< 2 mm).Air-dried samples of the soils were gently sieved to separate 1-2 mm macroaggregates, which were used to measure Soil Aggregate Stability (SAS); it was assayed by wet-sieving with the single sieve method (Kemper and Koch 1966).Porosity was calculated by way of bulk density, obtained with the paraffin method (Blake and Hartge 1986).Water availability at a permanent wilting point (PWP) (-1500 kPa) and at field capacity (FC) (-33 kPa) were measured using a volumetric pressure plate ex-tractor (Richards 1947).The water holding capacity (WHC, as mm/profile) was calculated as the difference in water retention between field capacity and permanent wilting point (USDA 2012).
Particle size determination in (hyper)gypseous soils cannot be performed accurately due to the lack of clay dispersion when gypsum is present in the soil (Vieillefon 1979).Laser diffraction provides at least some results that can be compared with field texture determinations and in our experience it has provided acceptable matches.Particle size distribution was therefore measured with a Malvern Mastersizer 2000, which uses laser method.This method underestimates clays in favor to fine silt, so the clay value was corrected according to the equation: y = 3.089x -2.899 (Taubner et al. 2009), where "x" is the clay value obtained with laser method and "y" is the corrected value to standardize with pipette method.Textural class is shown in the USDA system.
The pH was determined in a 1:2.5 ratio in H 2 O, total carbonate content by calcimetry, total soil organic C by wet oxidation (organic matter was estimated using the van Bemmelen factor, 1.724), Cation Exchange Capacity (CEC) by extraction with AcONH 4 , soil salinity by checking Electrical Conductivity (ECe) of the extract at 25 ºC, and soluble ions were measured in the extract (Page et al. 1982).Total N was obtained for each horizon with Kjeldahl method.Gypsum content was measured by gravimetry according to Vieillefon (1979).The sodium adsorption ratio (SAR) was measured according to the US Salinity Laboratory Staff (1954).
Soil thin sections of selected horizons were prepared using standard techniques (Benyarku and Stoops 2005).Their micromorphological description was done according to Stoops (2003) using a polarizing microscope.

Morphological properties
The main field morphological characteristics of the profiles are summarized in Table 1.In general, top horizons show vermiform gypsum accumulations in a calcareous matrix, while subsurface horizons (By) are whitish, massive, with generalized flour-like gypsum accumulations.Dry consistency is classified as soft for all Ah horizons and all the horizons of toe-slope but it is getting harder in depth.

Chemical properties
Soil pH is basic for all the horizons due to the presence of CaCO 3 , except the 2Az horizon in toe-slope, which is very basic because of sodium and magnesium carbonates (Table 2).These results are similar to others obtained in Gypsisols (Herrero 1991;Machín and Navas 1993;Artieda 1996;Florea and Al-Joumaa 1998;Cantón et al. 2003).Gypsum content ranges from 6 (Ah) to 84% (By

Origin of the buried horizon
The values of SOM, salinity, SAR and pH in the horizon 2Az of the toe-slope are very different to those found in the other horizons.This is because the horizon 2Az was buried, unlike the rest of the horizons.The 2Az horizon was probably buried as a result of increased erosion between the XVI and XIX centuries.During this period, the Spanish landscape was particularly subject to erosion be- A document dated from 1270 reports that there was a forest in the plateau atthe top of these mountains (Giménez-Soler 1922).The document regulated the exploitation of the area, preserving trees.Today there is no forest on this plateau (including head-slope and shoulder-slope of this catena).In this area, Braun-Blanquet and De Bolos (1957) point out the dominance of Artemisia herba-alba Asso and Salsola vermiculata L., which show evidence of past cultivation.Evidently, deforestation of huge areas of the Ebro Basin played an important role in Ebro Delta formation, which was accelerated between the XV and XVII centuries (Fatoric and Chelleri 2012).Constante and Peña-Monné's ( 2009) study showed sediment accumulations in a closed area, even on the left margin of the Ebro River.They cite a sediment accumulation, which is similar to our toe-slope profile over the 2Az, because of the position, gravel percentage and depth.This sediment dates back between the XVII and XVIII centuries.

Physical properties
Loam textural classes predominate in these soils, but on the head-slope sand is of greater importance.Gravels are negligible on the backslope and foot-slope; however they make up more than a half of the head-slope and the By and Cy horizons of toe-slope.The high concentration of gravels at the top of the slope is because the bedrock is very close to the surface; whereas their accumulation in the lower part is due to gravity.
The toe-slope in the studied slope is also part of a main slope.This explains why there are no gravels in the soils of medium studied slope.
Horizons show mainly a 10YR hue (Table 1), with values ranging from 6 to 8 and chroma from 1 to 3 (light gray and light brownish gray).This is due to the high gypsum content and low SOM content for most of the horizons.The value shows a significant correlation with gypsum content (R=0.84;n=15; p<0.001) and with SOM (R=-0.74;n=15; p<0.002).This strong relationship between color, soil component and SOM was previously reported (Badía et al. 1998).Furthermore, analyzing some Gypsisols (those with texture, gypsum percentage, PWP and FC reported) that were described by other authors (Herrero 1991;Olarieta et al. 1991, Badía et al. 2006;Badía et al. 2008) and the horizons of this paper, a relationship is found between PWP and FC.For most of these horizons (blue circles in Figure 4b) the relationship can be described as PWP (%) = 0.470 (FC %) -1.749 (r=0.84;n=44; p<0.001).However, some horizons (n=11, red triangles in Figure 4b) do not follow this regression.All of them are loamy horizons with a gypsum content higher than 40%, which shows that a gypsum percentage of more than 40% provokes a difference in water retention for loamy horizons.However, other loamy horizons with less gypsum (< 40%) show a similar behaviour to the rest of the Gypsisol horizons (n=44) defined by the previous equation.
In the studied horizons of this catena (n=15), where gypsum levels range from 40% to 60%, a decrease of PWP (and also FC in sandy-loam horizons) is observed.However, these horizons have more sand, which also influences AWC.The sand percentage affects AWC according to the following regressions:      Lenticular gypsum is found as nodules or coatings that are interpreted as recrystallizations of primary gypsum.Furthermore, we find some iron oxy-hydroxide nodules, which are considered relict from a past seasonal soil flooding because they are found inside soil aggregates (Figures 5  and 7).The By horizon also has lenticular gypsum crystals and isles fabric, but microcrystalline gypsum is more general (Figures 8 and 9).Microcrystalline gypsum found in the studied soils is due to gypsum rock weathering in the process explained by Herrero et al. (1992), while lenticular gypsum is either the result of precipitation from a gypsum-rich solution, or from the reprecipitation of microcrystalline gypsum.The main processes are dissolution and precipitation, together with biotic ones in microcrystalline gypsum (Herrero 1991).
Herrero (1991) reports high AWC values in microcrystalline gypsum horizons, which hold more water than lenticular ones due to the smaller porosity and the association between gypsum and some organic materials.However, roots have problems in using this water because the growth of gypsum crystals (as loose infillings) creates irregular, discontinuous packing pores where roots cannot penetrate (Poch and Verplancke 1997); see also Figures 5 and 8.This could explain, together with low rainfall and low nutrient level (Table 6), why the vegetation is so scarce in the area, in spite of having a high AWC.their properties.The soil profile in the headslope is classified as Haplic Gypsiric Leptosol while the others are classified as Gypsisols, belonging to various units depending on the gypsum content, stoniness, etc. (Table 7).

Conclusions
The studied soils have a high gypsum content, together with low salinity and basic pH values due to the presence of carbonates.At the top of the slope, soils show the lowest levels of soil organic matter, soil aggregate stability, cation exchange capacity (CEC) and available water contents (AWC); these values increase for the rest of the slope.Soils have poor chemical fertility due to abundance of gypsum and lime, which form soils with low CEC because of the low clay content and low organic matter in this arid environment.Also, physical fertility is poor due to the pore characteristics features in gypsum-rich horizons, which are not suitable for root penetration.
In loam horizons, AWC increases when microcrystalline gypsum contents are high (> 40%).This increase is higher in horizons with gypsum content > 60% because an increase in field capacity (FC) occurs together with a decrease in permanent wilting point (PWP).However, in horizons with gypsum content between 40-60% also the increase in AWC is only due to a decrease in PWP.Field capacity is notably reduced by an increased sand percentage.Microcrystalline gypsum, mainly due to gyprock weathering, can form lenticular gypsum by dissolution and reprecipitation; both gypsum forms are secondary.
These high AWC values are counteracted by an increasingly irregular pore space not suitable for root growth, making it difficult for roots to develop in these horizons, as is shown in the field.This behaviour should be studied in future researches with Gypsisols with other textures than loam.

Aknowledgements
Government of Aragón (Spain) and la Caixa sponsored this work throughout the project GA-LC-055/2011.We want to thank Clara Martí´s comments about the first manuscript.Also the revision of Martinho Martins was gratefully received.• Gutiérrez M, Sancho C, Desir G, Sirvent J, Benito G, Calvo A. 1995.Cuantificación de la erosión hídrica y procesos geomorfológicos en terrenos arcillosos y yesíferos de la Depresión del Ebro.Zaragoza (Spain): Ministerio de Agricultura -Universidad de Zaragoza.389 p.
• Herrero J. 2004.Revisiting the definitions of gypsic and petrogypsic horizons in Soil Taxonomy and World Reference Base for Soil Resources.Geoderma 120:1-5.

Figure 2 .
Figure 2. Relation between PWP , CEC and value color with gypsum content, and relation between organic matter and SAS.All show a p<0.001.
these soils is mainly microcrystalline, silt-sized and with a flour-like consistency in the field.Powdery gypsum, made of sandsized lenticular gypsum crystals (Poch et al. 2010), was not observed in the field.Flour-like gypsum predominates in all the horizons, which is in agreement with the micromorphological observations of the Ah2 and By horizons of the back-slope profile (Figures5-9), where this gypsum size is found in the micromass and also as pedofeatures.Microcrystalline gypsum occupies almost all the horizon volume in By (Figure8).

Figure 4 .
Figure 4. a) Relation between PWP and FC for all the horizons of the studied toposequence.Four groups have been established according to their position and features.b) Relation between PWP and FC for horizons of this work and also for other horizons described by other authors (see text) in Gypsisols of the Ebro Valley.Two groups were established, one which holds the loamy horizons with a gypsum content higher than 40%, and a second one including the rest of the horizons.

Figure 8 .
Figure 8. Micromorphology of By horizon (Back slope), in XPL (left) and in PPL (right).GL) Lenticular gypsum.N) Nodule of iron oxy-hydroxides.A) Fragment of original marl.R) Root sections.Almost all the volume is occupied by microcrystalline gypsum, which also appears filling the pores.Frame length: 6 mm.

Figure 7 .
Figure 7. Micromorphology of Ah2 horizon (Back-slope) magnified from Figure 6, XPL (left) and PPL (right).MG) Nodule of microcrystalline gypsum.Note the almost isotropy of the nodule due to the random packing of silt-size gypsum crystals.A) Aggregates of calcite, fine silt and clay.P) Pores.N) Nodules of iron oxy-hydroxides.R) Root section.Frame length: 1.2 mm in the main picture and 0.3mm for the box placed down-left.

Figure 9 .
Figure 9. Microphotographs of By horizon (Back slope) developed on gypsum rock: sand-sized lenticular gypsum infilling pores and void spaces, surrounded by a mass of microcrystalline gypsum, in XPL (left) and in PPL (right).LG) Lenticular gypsum.MG) Microcrystalline gypsum.P) Pores.A) Fragment of original marl.N) Nodule of iron oxy-hydroxides.Frame length: 6 mm in the main picture and 1.2 mm in the smallest one.

Table 3 .
Pedotransfer equations relating SOM, clay, gypsum and carbonates with CEC

Table 2 .
Main properties of the profile.HS: Saturation Humidity.FC: Field Capacity.PWP: Permanent Wilting Point.AWC: Available Water Capacity.*SAR: Sodium Adsorption Ratio

Table 5 .
Water retention at different tensions for loamy horizon and relation between water tension and gypsum content

Table 7 .
Soil forming processes, horizons and diagnostic properties, and classification of the soils studied in accordance with the WRB (IUSS 2007) taxonomy system and Soil Taxonomy System(Soil Survey Staff 2010) Gomes L, Arrúe JL, Sterk G, Richard D, Gracia R, Sabre M, Gaudichet A, Frangi JP. 2003.Wind erosion in a semiarid agricultural area of Spain: the WELSONS project.Wind erosion in Europe.Catena 52:235-256.