Soil crust distribution and development in a semi-arid landscape dominated by Macrochloa tenacissima and Anthyllis cytisoides A field work research in the Rambla Honda Basin, Almería Province, SE Spain

(1)

Soil crust distribution and development in a

semi-arid landscape dominated by Macrochloa

tenacissima and Anthyllis cytisoides

A field work research in the Rambla Honda Basin, Almería Province, SE Spain

Annabel Isarin - 11031786 BSc Thesis Project

Future Planet Studies – major Earth Sciences

University of Amsterdam – Institute for Biodiversity and Ecosystem Dynamics (IBED) Supervisor: Dhr. dr. L.H. (Erik) Cammeraat

July 3rd, 2018

(2)

Abstract

The Mediterranean region is threatened by desertification. In order to reduce the expected effects of desertification, it is necessary to understand the processes that influence infiltration and erosion at a local scale. One of the components that influences infiltration and runoff are soil surface crusts, which are likely to develop on bare soil patches in semi-arid regions such as SE Spain. In this research the distribution and development of soil surface crusts at Macrochloa tenacissima and Anthyllis cytisoides on two geomorphological units in the

Rambla Honda in SE Spain have been studied. Crust samples have been taken to determine the soil organic matter content, organic carbon content, inorganic carbon content, total carbon content and CND in the laboratory. Furthermore, crust descriptions and photos have been made both in the field and in the laboratory. It was found that sieving crusts are

abundant in the Rambla Honda and that their organic matter content and organic carbon content is significantly higher at the canopy border of the Anthyllis cytisoides compared to its bare patch. The total carbon was also found the be higher at the Anthyllis cytisoides

compared to the Macrochloa tenacissima as is the amount of plant material on the crust. Thin crusts and crusts with much coarse material were found to occur more on the alluvial fan in comparison to the hill. No significant differences were found for the CND between the hill and the fan, the Anthyllis cytisoides and the Macrochloa tenacissima and the bare patch and the canopy border which can be attributed to the extremely low CND values. In addition, thin crusts and crusts with much coarse material were found to occur more on the alluvial fan in comparison to the hill. Combining these results with results on infiltration and runoff in the Rambla Honda is needed to assess the influence of the distribution and development of sieving crusts on the potential erosion in the Rambla Honda. Moreover, further research should focus on the distribution and development of biocrusts and crust field description methods which may result in more accurate and complete conclusions.

Keywords: crust development, sieving crust, soil degradation, erosion, vegetation cover, Anthyllis

(3)

Introduction

Desertification is the process of land degradation reducing land productivity due to adverse human actions and climatic variations. In the 1970s it was recognised as a global problem (UNCOD, 1977). In Europe, especially the Mediterranean region is threatened by this process. Desertification is expected to cause many ecological and socio-economic problems in the Mediterranean region (Cammeraat et al., 2002). In order to reduce the expected effects of desertification, it is necessary to understand the processes that influence infiltration and erosion at a local scale. Then, management decisions can be influenced which could prevent the degradation of vulnerable land (Cammeraat et al., 2002, Huang et al., 2016; Rodriguez-Caballero et al., 2017). One of the components that influences infiltration and runoff are soil surface crusts (Belnap, 2006; Thompson et al., 2010). Many semi-arid regions such as SE Spain show a patchy vegetation pattern due to the lack of available water. On the bare components of these two-phase mosaics these soil surface crusts are likely to develop (Contreras et al., 2008; Valentin and Bresson, 1992).

Different types of soil crusts can occur in semi-arid regions. They can be divided in biocrusts and physical crusts (Casenave and Valentin, 1992). Biocrusts are composed of organisms such as algae and cyanobacteria (Belnap and Eldridge, 2001). In semi-arid regions they strongly influence hydrology, erosion and soil properties (Belnap, 2006; Lazaro et al., 2008). Biocrusts generally reduce infiltration and increase runoff (Eldridge and Greene, 1994). Furthermore, these crusts increase water storage and the amount of available nutrients which improves soil fertility (Cantón et al., 2004; Harper and Belnap, 2001; Harper and Pendleton, 1993; Lazaro et al., 2008). Consequently, biocrusts may influence the establishment and performance of vegetation and landscape formation in semi-arid ecosystems (DeFalco et al., 2001; Escudero et al., 2007). They are often classified according to the organisms they are dominated by.

Physical surface crusts are formed by geomorphological processes. They also generally, reduce infiltration rate, and increase runoff (Casenave and Valentin, 1992). However, they may reduce water storage in the soil. These physical crusts can be classified as structural, depositional or erosion crusts (Valentin and Bresson, 1992). A type of physical crust that is very common in the Rambla Honda is the sieving crust (Contreras et al., 2008). When they are well developed they consist of three well sorted layers. The uppermost layer is composed of loose coarse grains, the middle one is composed of fine grains and the deepest layer is a

Figure 2: Patchy vegetation pattern. Photo taken with drone by Thijs de Boer and Etienne de Jong during field work, Rambla Honda.

Figure 3: Patchy vegetation pattern. Photo taken with drone by Thijs de Boer and Etienne de Jong during field work, Rambla Honda.

(5)

plasmic layer of fine particles with reduced porosity (Valentin and Bresson, 1992). They primarily develop due to the impact of waterdrops which compact the soil. The textural differences between the different layers are caused by mechanical sieving so that the finest particles are deposited the deepest. Moreover, clay translocating through the coarse top layer forms the plasmic horizon (Valentin and Bresson, 1992).

In semi-arid regions, the soil surface layer largely determines the erodibility of the soil (Dunne et al., 1991). This is a measure of the sensitivity of the soil to detachment and transport by wind and water (Lal, 1994). Sieving crusts, which are common in the Rambla Honda, may act as physical barriers that restrict the translocation of water to deeper layers and consequently, increase runoff (Contreras et al., 2008). This increased runoff can cause both a positive and a negative feedback interaction between crust formation and desertification. If the increased runoff resulting from physical crusts cannot be taken up by plants, the presence of physical crusts reduces plant water availability. Furthermore, it may increases soil erodibility which is likely to enhance desertification (Belnap, 2006; Cerdà, 1997; Kröpfl et al., 2013).

However, if the increased runoff can infiltrate in vegetated sites, soil surface crusting can mitigate the effects of desertification (Ludwig et al., 1994; Valentin and d’Herbés, 1999). Moreover, the removal of soil surface crusts can cause plant mortality in desert systems (Valentin and d’Herbés, 1999). It can be concluded from this that physical crusts, such as the sieving crusts in the Rambla Honda, can both enhance and reduce desertification. The amount of physical crusts should be balanced to maximize their positive influence and minimize their negative influence. As these physical soil crusts influence the infiltration rate and erodibility of the bare components they form an important factor for vegetation growth and landscape formation and functioning (Calvo-Cases et al., 1991; Contreras et al., 2008; Verheijen and Cammeraat, 2007). Therefore, it might be important to study the distribution and development of physical soil crusts in areas that are threatened by desertification such as SE Spain (Assouline et al., 2015).

The presence of biocrusts may also have important consequences for geomorphic processes and vegetation growth as their erodibility is lower than that of physical soil crusts, and influence runoff and infiltration. Knowing whether and in what amounts these biocrusts are present may be important for the sensitivity of the Rambla Honda to degradation (Belnap, 2006; Escudero et al., 2007; Lazaro et al., 2008; Maestre et al., 2011; Rodríguez-Caballero et al., 2017.).

In addition, soil erodibility is influenced by soil properties such as its aggregate stability and organic carbon content of the soil (Dunne et al., 1991; Cantón et al., 2009; Cerdà, 1998). Moreover, vegetation influences soil erodibility. Vegetation enhances infiltration and increases organic carbon content of the soil through litter input and forms a protective cover (Cerdà, 1998; Oades, 1993). Moreover, vegetation may provide a more suitable microclimate for flora and fauna under the plant cover which will result in a higher aggregate stability (Cerdà, 1998; Oades, 1993). It can be concluded from this that vegetation cover, soil organic carbon content, aggregate stability and crust formation influence each other and the erodibility of the soil which may have implications for the rate of desertification in the Rambla Honda. Researching these soil properties may be important in decision making on the mitigation of desertification (Assouline et al., 2015).

Finally, knowing how Macrochloa tenacissima or Anthyllis cytisoides influence the surrounding soil will be useful in vegetation recovery strategies after disturbances such as forest fires or the abandonment of agricultural land. This could also enhance the mitigation of land degradation (Cerdà, 1998).

(6)

This research aims to assess the distribution and development of soil surface crusts in the Rambla Honda in southeast Spain. Through analysing data that was collected in the field and laboratory, this research will answer the following research question: ‘How do soil surface crusts spatially distribute and develop in a semi-arid landscape dominated by Macrochloa tenacissima and Anthyllus cytisoides on two geomorphological units?’ In order to answer the main research question three sub-questions have been formulated:

1) Which types of soil surface crusts occur on the hillslope and the alluvial fan in the Rambla Honda catchment?

2) How is the spatial development of the soil surface crusts related to the presence of Macrochloa tenacissima and Anthyllis cytisoides?

3) What is the difference between soil crust formation on the hillslope and on the alluvial fan?

The goal of answering these questions is to get an overview of the crusts present in the Rambla Honda and their development in relationship to the presence of Macrochloa tenacissima and Anthyllis cytisoides and the location on the hillslope. This will gain more insight in the erodibility of the soil and consequently, the sensitivity of the area to

degradation. This research will focus on sieving crusts as they are quite common in the field work area.

(7)

Field work area:

The field work site is located in the Rambla Honda in southeast Spain. This rambla drains an area of approximately 30 km2 _{on the southern slope of the Sierra de los Filabres} (Puigdefábregas et al., 1999). This is a mountain range which mainly consists of Precambrian to Triassic metamorphic rocks. The lithology consists of dark grey, fine grained mica schists with garnets and graphite, crossed by quartz veins with thin phyllite layers (Contreras et al., 2008). The climate is semiarid Mediterranean which means the summers are long and hot while the winters are mild. The annual average temperature is 16 degrees Celcius. The annual rainfall is approximately 300 mm per year. The soils show little pedogenetic development and consist mainly of loamy sands and fine sandy loams. (Contreras et al., 2008).

The field site includes only a small part of the Rambla Honda of approximately 500 m2. The field work area can be divided in the hillslopes, the fan they are connected to and the rambla itself. The upper hillslopes are dominated by Macrochloa tenacissima which grow on very shallow soils. The alluvial fan system is dominated by Anthyllis cytisoides on the upper part and Retama sphaerocarpa on the lower part (Puigdefábregas et al., 1999). Much research has already been done at this site. However, sieving crusts have not been studied that much in the Rambla Honda before, despite the fact that they are quite abundant in the area.

(8)

Method

Field methods

The crusts present upslope of Macrochloa tenacissima and Anthyllis cytisoides plants have been classified and sampled. They were located on both the hill and the alluvial fan of the Rambla Honda. Five transect perpendicular to the slope have been divided over the three shoulders and the fan they are connected to. For each transect, one Macrochloa tenacissima and one Anthyllis cytisoides was chosen on both the hill and the alluvial fan. This resulted in a total of ten observations of Macrochloa tenacissima and ten observations of Anthyllus cytisoides. For each plant the crusts located near the canopy border and approximately half a meter uphill were classified. Under the canopy itself no surface crusts were present. This resulted in a total of 40 samples that were classified.

At each crust first, the surface was described in terms of the amount of rocks, the vegetation cover and the amount of dead plant material that was present. Then a brush was used to determine the amount of microhorizons. Subsequently a field knife was used to determine the thickness, whether it was fragmentary or continuous, the fraction of roots in the crust and the fraction of rocks. Biocrusts were only mentioned when they were present (See Appendix… for description form).

Samples of the soil crusts were taken at the same sites where the crusts were classified along the five transects. Only the crust-layer was sampled which was around 0.5 cm thick. From these 40 crusts some material has been taken which was put into little boxes with some cotton wool to help it stay intact during transport. In addition, 100 gram of the 2 mm fraction of the crusts was collected using a sieve.Of these samples the soil organic matter content, soil organic carbon content, inorganic carbon content and total carbon cocntent was determined back in the laboratory.

In the field, the Herrick test was performed on some obtained surface aggregates (pieces of crust) which is an aggregate stability field test. It consists of 8 small sieves on which a small sample of the crust can be put. The 8 sieves with the crust samples are put in demineralized water and after five minutes, all 8 sieves are moved up and down in the water five times after which the amount of the crust that is left will be determined as a measure of its stability (Herrick et al., 2001).

(9)

Laboratory methods

Aggregate stability

On the crust samples that were taken to the lab a dripping test has been performed to determine the aggregate stability of the crusts (Imeson and Vis, 1984). First, the aggregates were saturated with distilled water for 24 hours. Then the stability of the aggregates was tested by counting the number of drops needed for the aggregate to pass through a 2.8 mm sieve (CND). The more droplets that are needed for the aggregate to pass through the sieve, the higher its stability is. This test was performed on 20 subsamples for each crust sample.

Soil organic matter

The soil organic matter was determined by using the Loss on Ignition (LOI) method of Dean (1974). First, the soil moisture is removed by drying 4 – 5 grams of the <2 mm fraction at 105 °C for 24 hours. With the difference in weight before and after the drying the soil moisture content has been calculated as can be seen in equation 1.

[1] 𝑆𝑀𝐶 [%] = 𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 [𝑚𝑔] − 𝐷𝑟𝑖𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 [𝑚𝑔]_{𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 [𝑚𝑔]} · 100%

Subsequently, the soil organic matter content was determined by igniting the samples in an oven at 500 °C overnight. The difference in weight with between the dried samples and the ignited samples is used to calculate the soil organic matter content as can be seen in equation 2.

[2] 𝑆𝑂𝑀 [%] = 𝐷𝑟𝑖𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 [𝑚𝑔] − 𝐼𝑔𝑛𝑖𝑡𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 [𝑚𝑔]

𝐷𝑟𝑖𝑒𝑑 𝑠𝑎𝑚𝑝𝑙𝑒 𝑤𝑒𝑖𝑔ℎ𝑡 [𝑚𝑔] · 100%

5.2.3 Inorganic carbon content

The inorganic carbon content was determined with the <2 mm fraction that had been milled in the lab. For determining inorganic carbon the Van Wesemael method was used (van Wesemael, 1955). The gravimetric loss of CaCO3, which is caused by adding a reaction with HCL is used to determine the carbonate content in the soil (see equation 3).

[3] 𝐶𝑎𝐶𝑂3+ 𝐻+→ 𝐶𝑎2++ 𝐶𝑂2+ 𝐻2𝑂

The entire procedure of this experiment can be read in the paper by van Wesemael (1955). Due to time restrictions only 18 samples were used to determine inorganic carbon. Inorganic carbon of other samples were calculated using a linear regression between organic matter content and organic carbon content (Howard and Howard, 1990) (see Appendix A.3.1). To calculate IC the evaporated CO2 content was calculated using equation 4. The outcome of equation 4 was used in equation 5 to calculate the inorganic carbon content.

[4] 𝐶𝑂₂[%] = 𝑃 𝑥 𝑔𝑟𝑎𝑚𝑠 𝐶𝑎𝐶𝑂3 𝑥 44

𝑄 𝑥 100 𝑥 𝑅 · 100%

P: gravimetric loss of the sample in grams Q: gravimetric loss of CaCO3 in grams R: grams of air dry sample

[5] 𝐶𝑎𝐶𝑂3[%] = 𝐶𝑂2[%] ·

100 44

(10)

Total carbon and organic carbon

The soil total carbon content was determined using the Vario El Cube Elemental Analyser by Elementar Analyse Systems, also referred to as CNS analysis. Milled soil samples and measuring standards consisting of sulfanilic acid were folded into alluminium foil cups. The samples and standards were put in the Elemental Analyser that measures element levels through flash combustion at 1800°C with an excess of oxygen (Elementar Germany, 2017). After passing through several traps N2, CO2, H2O and SO2 are the only particles left. The masses of these particles were used to calculate the total carbon content. By subtracting the inorganic carbon content from the total carbon content the organic carbon content can be determined.

Crust description with microscope

The sampled crusts were described and photographed by using a microscope. This was done to gain more insight in the variation of the crusts in terms of root content, amount of pores and cracks, surface smoothness and the ability to distinguish the layers. In addition, photos have been taken from both the top view and the cross section of the crusts.

Statistical analysis

In order to test whether certain groups differ significantly from each other, several Kruskal-Wallis tests have been used. Kruskal-Kruskal-Wallis tests have been chosen as the data is not normally distributed. These tests have been used to compare the following groups: canopy border and bare patch, hillslope and alluvial fan and Macrochloa tenacissima and Anthyllis cytisoides. In addition, a linear regression analysis will be performed to test whether there is any correlation between the organic carbon content of the crusts and their aggregate stability.

Furthermore, all descriptions of the crusts and the photos taken with the microscope were used to categorize all crust as either continuous or fragmentary, containing much coarse material or not much coarse material, containing much plant material or not much plant material, containing much pores and cracks or not much pores and cracks and thin or thick. Whether a crust sample falls into one or the other category was determined by using predominantly, the crust descriptions that were made in the field.

Whether crusts were fragmentary or continuous was already included in these descriptions and were categorized accordingly. Crusts with both on the surface as in the crust itself containing 30% or more material of the two biggest categories (See Appendix A.1.3. for description form) were categorized as containing much coarse material. The rest of the samples as containing not much coarse material. Crusts with more than 5 plants growing on approximately 20 cm2_{were categorized as containing much plant material. The rest of the} samples as containing not much plant material. Crust with a thickness of 0.4 cm or more were categorized as thick. The rest of the samples as thin. To categorize the samples as containing much pores or not the descriptions made in the laboratory were used. If the crusts were described as containing much pores they were categorized accordingly. The rest of the samples were categorized as containing not much pores (see Appendix A.2.3).

A chi-squared test was used to test whether these categorized properties are independent of the location on the catena, the plant species or the orientation from the plant.

(11)

Results

Boxplots

First, some boxplots are provided to give a quick overview of the data gained by performing the laboratory experiments. The data of the organic carbon content and the median of the CND have been visualized in these boxplots. A distinction has been made between the samples at the Macrochloa tenacissima plants and at the Anthyllis cytisoides plants (referred to as Vegetation), the samples on the Hill and on the Fan (referred to as Catena) and the samples at the bare patch and the canopy border (referred to as Orientation). More soil properties were determined through performing the laboratory experiments. However, only these two properties have been visualized in boxplots.

Figure 6:Difference in soil organic carbon content between samples on the fan and samples on the hill and samples at both plant species.

Figure 7:Difference in soil organic carbon content between the samples on the fan and on the hill for both plant species.

Figure 8:Difference in median CND between samples on the fan and samples on the hill and samples at both plant species.

Figure 9:Difference in median CND between samples on the fan and on the hill for both plant species.

(12)

The central mark in the boxplots indicates the median. The bottom and top edges of the box indicate the 25th and 75th percentiles. The whiskers extend to the most extreme data points not considered outliers, and the outliers are plotted individually using the '+' symbol. By using these boxplots it can be seen whether there are any differences in the median of two groups.

Figure 10: Difference in soil organic carbon content between bare and canopy for both species.

Figure 11: Difference in soil organic carbon content between the two plant species for both bare and canopy border.

Figure 12: Difference in median CND between bare and canopy border for both plant species.

Figure 13: Difference in median CND between the two plant species for both bare and canopy border.

(13)

Kruskal Wallis

The data visualized in the figure 6 shows that there is a difference in the soil organic carbon content between the Macrochloa tenacissima and the Anthyllis cytisoides as the mean is higher at the Anthyllis cytisoides. As can be seen in table 1 no significant difference was found in the organic carbon content between the two plant species. Furthermore, figure 7 shows that the mean organic carbon content is higher for the Anthyllis cytisoides on the fan compared to the hill. This figure also shows that the mean organic carbon content is higher for the Macrochloa tenacissima on the hill compared to the fan. However, as can be seen in table 1 no significant difference was found in both cases.

Figure 8 shows that the median CND is higher on the fan compared to the fan and higher at the Anthyllis cytisoides compared to the Macrochloa tenacissima. However, as can be seen in table 1 there were no significant differences found in the median CND between the two plants species and the two geomorphological units.

Figure 10 shows there is a difference in the mean organic carbon content between the bare patch and the canopy border at the Anthyllis cytisoides as the mean is higher for the canopy border of the Anthyllis cytisoides. As can be seen in table 1 this difference is significant. No boxplots of the organic matter content have been visualized. However, as can be seen in table 1 there also is a significant difference in the organic matter content between the bare patch of the Anthyllis cytisoides and the canopy border of the Anthyllis cytisoides. Figure 11 shows the mean organic carbon content is higher at the canopy border of the Anthyllis cytisoides compared to the canopy border of the Macrochloa tenacissima. However, no significant difference was found as can be seen in table 1.

In addition, in figure 13 a difference can be seen in the median CND between the two plant species for both the bare samples and the canopy border samples. In both cases the mean of the Anthyllis cytisoides is higher. Table 1 shows this difference is not significant.

Finally, a significant difference has been found between the total carbon content of the Macrochloa tenacissima and the Anthyllis cytisoides (see table 1). No boxplots of this data have been visualized.

Tabel 1:Resulted p-values from Kruskal Wallis tests. The groups that have been compared with each other are listed in the first two columns. Behind each crust propert has been stated whether the difference between the mentioned groups is significant or not. The abbreviations stand for; M= Macrochloa Tenacissima, A= Anthyllis Cytisoides, H= Hillslope, F= Fan, B= Bare patch, C= Canopy border.

Goup 1 Group 2 SOM Significant OC Significant2 TC Significant3 CND Significant4

MB AB 0.8798 no 0.9397 no 0.2568 no 0.4818 no MC AC 0.4015 no 0.1711 no 0.0851 no 0.2405 no MB MC 0.3643 no 0.8206 no 0.7624 no 0.4279 no AB AC 0.0152 yes 0.0243 yes 0.0703 no 0.7107 no MH MF 0.1124 no 0.3643 no 0.6272 no 0.5774 no AH AF 0.6911 no 0.8946 no 0.6501 no 0.8429 no M A 0.3891 no 0.1484 no 0.0455 yes 0.202 no H F 0.4221 no 0.5495 no 0.6494 no 0.595 no

(14)

Figure 14, 15 and 16 contain the outputs of the Kruskal Wallis for the significant differences. Figure 14 shows that the soil organic matter content is higher for the canopy border of the Anthyllis cytisoides compared to the bare patch of the Anthyllis cytisoides. Figure 15 shows that the soil organic carbon content is higher for the canopy border of the Anthyllis cytisoides compared to the bare patch of the Anthyllis cytisoides (this can also be seen in figure 10). Figure 16 shows that the total carbon content is higher for the Anthyllis cytisoides compared to the Macrochloa tenacissima.

Figure 14: Kruskal Wallis test output S.O.M. Anthyllis, bare vs. canopy border.

Figure 15: Kruskal Wallis test output S.O.C. Anthyllis, bare vs. canopy border.

Figure 16: Kruskal Wallis test output Total Carbon, Macrochloa vs. Anthyllis.

(15)

Regression SOM and median CND

A linear regression line has been fitted through all the data of the soil organic matter content and the median of the CND. As can be seen in figure 17 the fitted line shows a negative correlation. The R2_{of the regression line is 0.0125 which indicates an explained variance of} approximately 1%. This means that the model does not explain the variance between the variables well. The p-value is 0.497 which indicates that the slope of the linear model does not differ significantly from zero, when using α = 0.05. It can also be seen that the data is very scattered and that there are many outliers. This indicates that it might not be meaningful to fit a regression line through this data. As the results of the Herrick test (see Appendix A.1.2) do not show much difference these results have not been used for another regression analysis.

(16)

Crust photos

Figure 18: 3FAB1 (3rd observation on Fan, at Anthyllis on

the Bare patch) top view. Figure 19: 1HMC1 (1st observation on the Hill, at _{Macrochloa at the Canopy border), top view.}

Figure 20: 2HAC1 (2nd observation on the Hill, at Anthyllis at the Canopy border), top view.

Figure 21: 5FAC1 (5th observation on the Fan, at Anthyllis at the Canopy border), top view.

Figure 22: 1HAB1 (1st observation on the Hill, at Anthyllis on the Bare patch), top view.

Figure 23: 5FMC1 (5th observation on the Fan, at Macrochloa at the Canopy border), top view.

(17)

In the figures 18 till 23 some top views from different crusts made with a microscope can be seen. These photos are some examples from what has been encountered in the field. From these photos it becomes clear that there is much variety in the appearance of the crusts. 1HMC1 and 5FAC1 contain much coarse material, plant material and pores in comparison to the other crusts. 1HAB1 also contains much coarse material , but does not contain much plant material and holes. In this crust it can be seen that the coarse material is interbedded in the finer material as can also be seen on the photo of 3FAB1. In addition, some dark material can be distinguished in the photo of 3FAB1. This may indicate some biocrust is present in this crust sample. 2HAC1 and 5FMC1 are much more smoother in comparison to the rest of the crusts. They do not contain much coarse material, plant material and pores. Further on the differences that can be seen will be discussed in more detail.

In figure 24 and 25 the cross sections of two different crusts can be seen. From these photos it becomes clear that, just as the top view the appearance of the cross section of different sieving crusts may vary much. In figure 24 it can be seen that the upper layer of the crust consists of fine material with coarser material and more pores in the lowest layer. The layers of the crusts are clearly visible. However, this distinction in layers is not so clearly visible in figure 25. These photos are added to give some insight in the way sieving crusts are built up and to point out that this may not always be the same.

Figure 24: 4FAB1 (4th observation on the Fan, at Anthyllis on the Bare patch), cross section.

Figure 25: 2FMB1 (2nd observation on the Fan, at Macrochloa on the bare patch), cross section

(18)

Chi squared test of independence

In table 2 the resulted p-values of the chi squared test of independence are provided. The location on the catena and the amount of plant material are not independent. The contingency table (see Appendix A.3.2) shows that much plant material occurs more on the Fan than on the Hill. Furthermore, the location on the catena and the amount of pores are not independent. The contingency table of these two properties shows that much pores occur more on the Hill than on the Fan. In addition, the location on the catena and the thickness of the crust are not independent. The contingency table of these two properties shows that thin crusts occur more on the Fan while thick crusts occur more on the Hill. Finally, the type of vegetation and the amount of plant material are not independent. The contingency table of these two properties show that much plant material occurs more near Anthyllis Cytisoides than near Macrochloa Tenacissima.

Table 2: p-values chi square test categorical data.

Property 1 Property 2 p-value Significant

Catena Covering 0.2424 no

Catena Coarse material 0.1404 no Catena Plant material 0.0062 yes

Catena Holes 0.0442 yes

Catena Thickness 0.0154 yes

Vegetation Covering 0.2424 no Vegetation Coarse material 0.8554 no Vegetation Plant material 0.0368 yes

Vegetation Holes 0.2391 no

Vegetation Thickness 0.0751 no Orientation Covering 0.2424 no Orientation Coarse material 0.267 no Orientation Plant material 0.8937 no

Orientation Holes 0.6225 no

(19)

Discussion:

During this research the distribution and development of soil surface crusts at Macrochloa Tenacissima and Anthyllis Cytisoides on two geomorphological units in the Rambla Honda in SE Spain is studied. The results will be discussed by using the formulated sub-questions beginning with the first sub-question: ‘Which types of soil surface crusts occur on the hillslope and the alluvial fan in the Rambla Honda catchment?’. The type of soil crusts that have been found in the Rambla Honda are all sieving crusts. Some were more developed than others in terms of being more continuous, thicker and harder to break apart in the field. At some sampling locations the crust covering was not abundant. It was expected that soil surface crusts would be abundant as Contreras et al., (2008) and Valentin and Bresson (1992) stated that soil surface crusts are likely to develop on the bare components of patchy vegetation patterns. It was also expected that sieving crusts would be abundant as Contreras et al., (2008) stated that these type of crusts are very common in the Rambla Honda. No clear biocrusts were found on the sampling locations on the hillslope and the alluvial fan. However, further down the alluvial fan some biocrusts were observed. Furthermore, on some photos that were made in the laboratory possibly some algal crust was observed. This research has not focussed on these biocrusts. However, as biocrusts are important ecosystem parameters that influence infiltration and runoff and plant growth, further research should focus on their distribution and development in the Rambla Honda (Belnap, 2006; Escudero et al., 2007; Lazaro et al., 2008; Maestre et al., 2011; Rodríguez-Caballero et al., 2017). The inorganic carbon content is extremely low in all samples which is characteristic for the Rambla Honda. It did not show any significant differences between categories.

The second sub-question was formulated as ‘How is the spatial development of the soil surface crusts related to the presence of Macrochloa Tenacissima and Anthyllis Cytisoides?’. It was expected that the presence of vegetation would influence crust formation and development through its influence on soil organic carbon content, aggregate stability and soil compaction. Vegetation cover contributes to the organic carbon content of the soil through the input of litter. Furthermore, it increases rainwater infiltration and causes a favorable microclimate for fauna and flora. This generally results in stronger aggregates (Oades, 1993; Cerdà, 1998, King et al., 2014). Consequently, soils beneath vegetation cover are commonly characterized by higher soil organic carbon content, total porosity, soil aggregate stability and lower surface compaction than soils in the bare patches (Bautista et al., 2007; Rey et al., 2011). Following this, it was expected that both the organic carbon content and the aggregate stability of the soil surface crusts would be higher near the canopy border of the plants in comparison to the bare patches. Results obtained with the Kruskal-Wallis tests show that for both the soil organic matter content and the soil organic carbon content a significant difference between the bare patch and the canopy border has only been found at the Anthyllis cytisoides. A possible explanation for the fact that this difference has not been found between the bare patch and the canopy border of the Macrochloa tenacissima is that Anthyllis cytisoides may have more undergrowth of different smaller species although it generally does not have a thick litter layer (Bochet et al., 1999; Escós et al., 1997). This would increase the organic carbon content at the canopy border of Anthyllis cytisoides which might not be the case at the densely canopied Macrochloa tenacissima (Bochet et al., 1999).

The results of the Kruskal Wallis tests also show that no significant difference in the aggregate stability has been found between the bare patches and the canopy borders of both plants. A possible explanation for this could be that that the aggregates would break apart very easily. Research from Cantón et al. (2009) already found poor aggregate stability in the Rambla Honda with a mean of only 26 drop impacts necessary to break down the aggregates.

(20)

In this research the median CND ranged from 2 to only 7. This is even lower than Cantón et al. (2009) found. The season in which the samples have been taken might be different in the research by Cantón et al. (2008) as the data of sampling is not mentioned. Not much is known about the influence of the season in which the samples have been taken on the aggregate stability. The lower aggregate stability can also be caused by the fact that not in all cases 20 aggregates could be obtained for the test and there was not much choice in aggregates. If the aggregates for the dripping test would have been collected in the field, this would possibly have resulted in bigger, more and stronger aggregates. Moreover, soil organisms and clay content can influence aggregate stability (Bronick and Lal, 2005). However, these properties have not been researched in this research. The extremely low aggregate stability values may also explain why no positive correlation has been found between the soil organic matter content and the CND which was expected (Bautista et al., 2007, Rey et al., 2011). Moreover, the amount of soil organic matter is relatively high compared to the results from Cantón et al., (2008).

In addition, it was expected that there would be more continuous crusts on the bare patches as there is less soil compaction near vegetation (Bautista et al., 2007; Rey et al., 2011). Moreover, the textural differentiation process which forms the sieving crust is slower or even disturbed below plant canopies due to the influence of roots and soil biota (Puigdefábregas et al., 1999). Therefore, it was expected that the crusts near the plant canopy is less developed than the crusts on the bare patches. This would mean that the thickness of the crust is smaller, it is more difficult to obtain samples and it is more mixed with coarse material (Puigdefábregas et al., 1999). However, the results of the chi square test of independence show no significant differences in amount of continuous crusts, thick crust and crusts containing much coarse material between the bare patches and the canopy borders for both plant species. This may be caused by the method of describing the crust and measuring its thickness. For the these properties not all crusts have been properly compared to each other. The thickness has only been measured of a small piece of crusts while the thickness of the crust may vary due to the initial surface roughness (Fox et al., 1998). Describing the ratios of coarse and smaller material (See Appendix… for crust descriptions) may also have been done subjectively which may have influenced the results of the chi squared test.

Furthermore it was expected that there would be differences in crust properties between the crusts near Macrochloa tenacissima and the crusts near Anthyllis cytisoides. Research of Bochet et al., (1999) found that soils near the canopy cover of Anthyllis cytisoides contained less organic carbon, were more compacted, had a lower porosity and a higher fragment of rock fragments compared to the soils near the canopy cover of Macrochloa tenacissima. This may be explained by the difference in plant canopy cover between Macrochloa tenacissima and Anthyllis cytisoide. Macrochloa tenacissima is characterized by a dense umbrella-shaped canopy cover while Anthyllis cytisoide generally has a more open deciduous canopy cover. This causes the litter layer near the Anthyllis cytisoide to be less abundant. Consequently, the organic carbon content may be lower which also influences soil compaction and porosity (Bochet et al., 2006). The results from the Kruskal-Wallis test showed that there is a significant difference in total carbon content between both plant species. However, it seems that the total carbon content is significantly higher at the Anthyllis cytisoides which was not expected. The p-value is not very small so, the difference is only just significant.

A possible explanation might be that the more undergrowth of different smaller plants enhances the soil total carbon content although Anthyllis cytisoides itself does not produce much litter (Bochet et al., 1999; Escós et al., 1997). This can also be linked to the results of the chi squared test that showed that there is more plant material at the Anthyllis cytisoides compared to the Macrochloa tenacissima.

(21)

The third sub-question was formulated as: ‘What is the difference between soil crust formation on the hillslope and on the alluvial fan?’. The amount of vegetation increases in a downslope direction in the Rambla Honda (Puigdefábregas et al., 1999). Consequently, the crusts present on the fan were expected to have both a higher organic carbon content and aggregate stability.

Furthermore, the presence of vegetation reduces soil compaction and the textural differentiation process which forms the sieving crust (Bautista et al., 2007; Puigdegábregas et al., 1999; Rey et al., 2011). Therefore, it was expected that more vegetation would result in less developed crusts with a smaller thickness and much coarse material. The chi squared test of independence showed that thin crusts and crusts with much coarse material occur more on the fan in comparison with the hill as was expected.

However, the results from the Kruskal Wallis tests show that no significant difference has been found in organic carbon content and aggregate stability between the fan and the hill. As mentioned before, the aggregate stability is extremely low which may cause the fact that no significant difference could be found between the fan and the hill. Furthermore, it might be the case that the different plant species are more important for the carbon content than the two geomorphological units which may have influenced the results.

Future research might be improved by also analyzing soil organisms and biocrusts to gain more insight in how erosion and land degradation interacts with semi- arid ecosystems. Furthermore, improved crust description and sampling methods could improve the accuracy of the results. Moreover, more time in the laboratory would have resulted in a better organic carbon determination as the values then would not have to be predicted as was done for the majority of the samples in this research. Finally, taking more samples in the field increased the chance that significant will be found when present in the field.

(22)

Conclusion

In this research the the distribution and development of soil surface crusts at Macrochloa tenacissima and Anthyllis cytisoides on two geomorphological units in the Rambla Honda in SE Spain have been studied. Sieving crusts have been found at all sampling locations, variating in abundancy, amount of coarse material, amount of plant material, amount of holes and thickness. It was found that the organic matter content and the organic carbon content is significantly lower on the bare patch of the Anthyllis cytisoides compared to its canopy border. This results was not found for the Macrochloa tenacissima. Furthermore, it was found that the total carbon content is higher at the Anthyllis cytisoides, although the difference was only just significant. No significant differences were found for the CND between the hill and the fan, the Anthyllis cytisoides and the Macrochloa tenacissima and the bare patch and the canopy border which can be attributed to the extremely low CND values. It can then be concluded that sieving crusts are abundant in the Rambla Honda and that their organic matter content and organic carbon content is significantly higher at the canopy border of the Anthyllis cytisoides compared to its bare patch. The total carbon was also found the be higher at the Anthyllis cytisoides compared to the Macrochloa tenacissima as is the amount of plant material on the crust. Thin crusts and crusts with much coarse material were found to occur more on the alluvial fan in comparison to the hill. These conclusions should be compared with results on infiltration and runoff in the Rambla Honda. Then, the influence of the distribution and development of sieving crusts on the potential erosion in the Rambla Honda can be assessed. Further research should focus on the distribution and development of biocrusts and crust field description methods which may result in more accurate and complete conclusions.

(23)

Acknowledgements

I would like to thank dr. L.H. Cammeraat for sharing his knowledge and experience and his supervision both in the field and during the rest of the project. Furthermore, I would like to thank J. Zethof MSc for his supervision and support during the fieldwork and his valuable feedback during the rest of the project. I would like to thank R. L. van Hall MSc for his supervision and advise during the laboratory experiments. Lastly, I would like to thank A.B. Dekkers, E.V. de Jong, M. Wadman and N. Verweij for the successful collaboration in the field, their support during the entire project and their valuable input.

(24)

References

Assouline, S., Thompson, S., Chen, L., Svoray, T., Sela, S., & Katul, G. (2015). The dual role of soil crusts in desertification. Journal of Geophysical Research: Biogeosciences, 120(10), 2108-2119. Bautista, S., Mayor, A. G., Bourakhouadar, J., & Bellot, J. (2007). Plant spatial pattern predicts hillslope runoff and erosion in a semiarid mediterranean landscape. Ecosystems, 10(6), 987-998.

Belnap, J. (2006). The potential roles of biological soil crusts in dryland hydrologic cycles. Hydrological

Processes, 20(15), 3159-3178.

Bochet, E., Poesen, J., & Rubio, J. (2006). Runoff and soil loss under individual plants of a semi‐arid mediterranean shrubland: Influence of plant morphology and rainfall intensity. Earth Surface Processes

and Landforms, 31(5), 536-549.

Bochet, E., Rubio, J., & Poesen, J. (1999). Modified topsoil islands within patchy mediterranean vegetation in SE spain. Catena, 38(1), 23-44.

Bronick, C. J., & Lal, R. (2005). Soil structure and management: A review. Geoderma, 124(1-2), 3-22. Calvo-Cases, A., Harvey, A., Paya-Serrano, J., & Alexander, R. (1991). Response of badland surfaces in south east spain to simulated rainfall. Cuaternario y Geomorfología, 5, 3-14.

Cammeraat, L., Willott, S., Compton, S., & Incoll, L. (2002). The effects of ants' nests on the physical, chemical and hydrological properties of a rangeland soil in semi-arid spain. Geoderma, 105(1-2), 1-20. Cantón, Y., Solé-Benet, A., Asensio, C., Chamizo, S., & Puigdefábregas, J. (2009). Aggregate stability in range sandy loam soils relationships with runoff and erosion. Catena, 77(3), 192-199.

Cantón, Y., Solé-Benet, A., & Domingo, F. (2004). Temporal and spatial patterns of soil moisture in semiarid badlands of SE spain. Journal of Hydrology, 285(1-4), 199-214.

Casenave, A., & Valentin, C. (1989). Les états de surface de la zone sahélienne: Influence sur

l'infiltration

Casenave, A., & Valentin, C. (1992). A runoff capability classification system based on surface features criteria in semi-arid areas of west africa. Journal of Hydrology, 130(1-4), 231-249.

Cerdà, A. (1997). The effect of patchy distribution of Stipa tenacissimaL. on runoff and erosion. Journal

of Arid Environments, 36(1), 37-51.

Cerdà, A. (1998). Soil aggregate stability under different mediterranean vegetation types. Catena, 32(2), 73-86.

Contreras, S., Cantón, Y., & Solé-Benet, A. (2008). Sieving crusts and macrofaunal activity control soil water repellency in semiarid environments: Evidences from SE spain. Geoderma, 145(3-4), 252-258. Dean, W. E. (1974) ‘Determination of Carbonate and Organic Matter in Calcareous Sediments and Sedimentary Rocks by Loss on Ignition: Comparison With Other Methods’, SEPM Journal of Sedimentary

Research. GeoScienceWorld, Vol. 44(1), pp. 242–248. doi:

10.1306/74D729D2-2B21-11D7-8648000102C1865D.

Dunne, T., Zhang, W., & Aubry, B. F. (1991). Effects of rainfall, vegetation, and microtopography on infiltration and runoff. Water Resources Research, 27(9), 2271-2285.

Eldridge, D., & Greene, R. (1994). Microbiotic soil crusts-a review of their roles in soil and ecological processes in the rangelands of australia. Soil Research, 32(3), 389-415.

Elementar Germany (2017) ‘vario EL Cube Elemental Analyser’. Available at:

https://www.elementar.de/fileadmin/user_upload/Elementar_Website/Downloads/Flyer/Flyer-vario-EL-cube-EN.pdf (Accessed: 7 June 2018).

Escós, J., Alados, C., & Emlen, J. (1997). The impact of grazing on plant fractal architecture and fitness of a mediterranean shrub anthyllis cytisoides L. Functional Ecology, 11(1), 66-78.

(25)

Escudero, A., Martínez, I., De la Cruz, A., Otálora, M., & Maestre, F. (2007). Soil lichens have species-specific effects on the seedling emergence of three gypsophile plant species. Journal of Arid

Environments, 70(1), 18-28.

Fox, D. M., Le Bissonnais, Y., & Quetin, P. (1998). The implications of spatial variability in surface seal hydraulic resistance for infiltration in a mound and depression microtopography. Catena, 32(2), 101-114. Harper, K. T., & Belnap, J. (2001). The influence of biological soil crusts on mineral uptake by associated vascular plants. Journal of Arid Environments, 47(3), 347-357.

Harper, K. T., & Pendleton, R. L. (1993). Cyanobacteria and cyanolichens: Can they enhance availability of essential minerals for higher plants? The Great Basin Naturalist, , 59-72.

Howard, P., & Howard, D. (1990). Use of organic carbon and loss-on-ignition to estimate soil organic matter in different soil types and horizons. Biology and Fertility of Soils, 9(4), 306-310.

Huang, J., Yu, H., Guan, X., Wang, G., & Guo, R. (2016). Accelerated dryland expansion under climate change. Nature Climate Change, 6(2), 166.

Imeson, A., & Vis, M. (1984). Assessing soil aggregate stability by water-drop impact and ultrasonic dispersion. Geoderma, 34(3-4), 185-200.

Kröpfl, A., Cecchi, G., Villasuso, N., & Distel, R. A. (2013). Degradation and recovery processes in semi‐ arid patchy rangelands of northern patagonia, argentina. Land Degradation & Development, 24(4), 393-399.

Lal, R. (1994). Soil erosion research methods CRC Press.

Lázaro, R., Cantón, Y., Solé-Benet, A., Bevan, J., Alexander, R., Sancho, L., et al. (2008). The influence of competition between lichen colonization and erosion on the evolution of soil surfaces in the tabernas badlands (SE spain) and its landscape effects. Geomorphology, 102(2), 252-266.

Ludwig, J. A., Tongway, D. J., & Marsden, S. G. (1994). A flow-filter model for simulating the

conservation of limited resources in spatially heterogeneous, semi-arid landscapes. Pacific Conservation

Biology, 1(3), 209-213.

Maestre, F. T., Bowker, M. A., Cantón, Y., Castillo-Monroy, A. P., Cortina, J., Escolar, C., et al. (2011). Ecology and functional roles of biological soil crusts in semi-arid ecosystems of spain. Journal of Arid

Environments, 75(12), 1282-1291.

Oades, J. (1993). The role of biology in the formation, stabilization and degradation of soil structure. Soil

Structure/Soil biota interrelationships (pp. 377-400) Elsevier.

Puigdefábregas, J., Sole, A., Gutierrez, L., Del Barrio, G., & Boer, M. (1999). Scales and processes of water and sediment redistribution in drylands: Results from the rambla honda field site in southeast spain. Earth-Science Reviews, 48(1-2), 39-70.

Rey, A., Pegoraro, E., Oyonarte, C., Were, A., Escribano, P., & Raimundo, J. (2011). Impact of land degradation on soil respiration in a steppe (stipa tenacissima L.) semi-arid ecosystem in the SE of spain. Soil Biology and Biochemistry, 43(2), 393-403.

Rodríguez-Caballero, E., Castro, A. J., Chamizo, S., Quintas-Soriano, C., Garcia-Llorente, M., Cantón, Y., et al. (2017). Ecosystem services provided by biocrusts: From ecosystem functions to social

values. Journal of Arid Environments,

Thompson, S., Harman, C., Heine, P., & Katul, G. (2010). Vegetation‐infiltration relationships across climatic and soil type gradients. Journal of Geophysical Research: Biogeosciences, 115(G2)

UNCOD [United Nation Convention on Desertification], 1977. Desertification, its Causes and Consequences. UNEP-Pergamon Press, New York.

Valentin, C., & Bresson, L. (1992). Morphology, genesis and classification of surface crusts in loamy and sandy soils. Geoderma, 55(3-4), 225-245.

Valentin, C., & d'Herbès, J. (1999). Niger tiger bush as a natural water harvesting system. Catena, 37(1-2), 231-256.

(26)

Van Wesemael, J. C. (1955). De bepaling van het calciumcarbonaatgehalte van gronden. Chemisch

Weekblad, 51, 35-36.

Verheijen, F., & Cammeraat, L. (2007). The association between three dominant shrub species and water repellent soils along a range of soil moisture contents in semi‐arid spain. Hydrological Processes, 21(17), 2310-2316.

(27)

Appendix

A.1. Fieldwork data

A.1.1. Sample codes, vegeation, orientation and catena

Table 3: Observation codes, vegeation, orientation and catena.

Observations Soil

layer Vegetation Orientation Catena

1FAB1 1 Anthyllis Bare Fan

1FAC1 1 Anthyllis

Canopy

border Fan

1FMB1 1 Stipa Bare Fan

1FMC1 1 Stipa

Canopy

border Fan

1HAB1 1 Anthyllis Bare Hill

1HAC1 1 Anthyllis

Canopy

border Hill

1HMB1 1 Stipa Bare Hill

1HMC1 1 Stipa

Canopy

border Hill

2FMC1 1 Stipa

Canopy

border Fan

2HAC1 1 Anthyllis

Canopy

border Hill

2HMC1 1 Stipa

Canopy

border Hill

3FAC1 1 Anthyllis

Canopy

border Fan

3FMC1 1 Stipa

Canopy

border Fan

3HAC1 1 Anthyllis

Canopy

border Hill

3HMC1 1 Stipa

Canopy

border Hill

4FAC1 1 Anthyllis

Canopy

border Fan

4FMC1 1 Stipa

Canopy

(28)

4HAB1 1 Anthyllis Bare Hill 4HAC1 1 Anthyllis

Canopy

border Hill

4HMC1 1 Stipa

Canopy

border Hill

5FAC1 1 Anthyllis

Canopy

border Fan

5FMC1 1 Stipa

Canopy

border Fan

5HAC1 1 Anthyllis

Canopy

border Hill

5HMC1 1 Stipa

Canopy

(29)

A.1.2. Herrick test

Table 4: Stability results from Herrick test. HS1 till HS8 represent all 8 replicates.

Observations HS1 HS2 HS3 HS4 HS5 HS6 HS7 HS8 1FAB1 3 3 3 3 3 3 3 3 1FAC1 3 3 3 3 3 5 5 4 1FMB1 5 3 5 3 4 4 4 3 1FMC1 3 4 5 5 3 4 4 5 1HAB1 5 6 6 5 6 5 6 6 1HAC1 6 6 6 5 5 6 6 6 1HMB1 6 6 6 6 5 6 6 6 1HMC1 6 6 6 6 6 6 6 6 2FAB1 5 4 4 6 6 5 6 6 2FMB1 6 5 6 5 6 6 6 6 2FMC1 5 6 6 5 6 6 6 6 2HAB1 6 6 6 6 6 6 6 6 2HAC1 6 6 6 6 6 6 6 6 2HMB1 5 6 6 6 6 6 6 6 2HMC1 4 5 5 6 6 3 5 6 3FAB1 6 6 6 6 6 6 6 6 3FAC1 4 5 6 6 4 5 5 6 3FMB1 5 4 6 6 5 6 6 6 3FMC1 6 6 6 5 5 6 6 6 3HAB1 6 6 4 5 6 6 6 6 3HAC1 6 6 6 6 5 5 6 6 3HMB1 6 6 6 6 6 6 6 6 3HMC1 5 6 6 6 6 6 5 6 4FAB1 6 6 6 6 6 5 6 6 4FAC1 6 6 6 6 6 6 6 6 4FMB1 4 4 5 3 5 3 6 6 4FMC1 6 6 6 6 6 6 6 6 4HAB1 5 6 6 4 6 5 6 6 4HAC1 6 6 6 6 5 5 6 6 4HMB1 6 6 5 5 6 6 6 6 4HMC1 6 6 6 6 6 6 6 6 5FAB1 3 3 4 3 3 3 5 3 5FAC1 6 6 6 6 6 6 6 5 5FMB1 6 6 5 5 5 5 6 4 5FMC1 6 6 6 6 6 6 6 6 5HAB1 6 6 5 6 6 6 5 6 5HAC1 5 6 6 6 6 6 6 6 5HMB1 6 6 6 6 6 6 5 6 5HMC1 6 6 5 6 6 6 6 6

(30)

A.1.3. Field crust description form

Figure 26: Example field description form from crust the 2nd observation on the Hill at the Macrochloa on the Bare patch.

(31)

A.2. Laboratory data

A.2.1. SMC, SOM, TotC, IC and OC

Table 5: SMC, SOM, TotC, IC and OC measured and fit.

SMC_mean [%] SOM [%] TotC [%] IC [%] OC [%] OC Determination

7,05 4,00 1,71 0,00 1,71 Fit 8,14 4,87 2,83 0,48 2,35 Fit 10,01 4,19 2,89 0,90 1,99 Fit 9,48 3,69 1,64 0,05 1,59 Fit 0,68 3,74 2,73 0,04 2,69 Measured 1,05 6,40 3,37 0,20 3,17 Fit 0,57 3,18 1,48 0,04 1,44 Measured 1,07 5,26 2,69 0,12 2,56 Fit 7,31 7,22 3,84 0,22 3,62 Measured 3,70 2,40 0,92 0,00 0,92 Fit 4,19 1,86 0,86 0,01 0,85 Fit 0,65 2,74 1,65 0,44 1,21 Fit 0,77 4,37 1,83 0,00 1,83 Fit 0,51 3,13 1,50 0,01 1,49 Fit 0,81 5,38 2,47 0,00 2,47 Measured 0,28 2,06 0,95 0,00 0,95 Fit 0,87 4,37 1,78 0,00 1,78 Fit 1,11 3,02 1,89 0,53 1,36 Fit 1,61 1,90 0,99 0,23 0,76 Fit 1,27 2,50 1,49 0,40 1,08 Fit 0,92 2,70 1,29 0,01 1,28 Fit 0,82 1,89 1,14 0,39 0,75 Fit 0,89 3,57 0,82 0,02 0,80 Measured 0,46 2,94 1,60 0,28 1,32 Fit 2,24 4,34 1,96 0,00 1,96 Fit 0,33 2,86 1,17 0,04 1,12 Fit 0,49 1,82 1,13 0,41 0,72 Measured 0,40 2,35 2,61 1,61 1,00 Fit 0,55 2,79 1,74 0,50 1,23 Fit 0,44 3,10 1,47 0,07 1,40 Fit 0,48 3,14 1,30 0,12 1,18 Fit 0,40 2,11 1,12 0,24 0,87 Fit 0,96 4,57 2,32 0,10 2,22 Fit 0,61 3,42 1,95 0,38 1,57 Fit 1,04 3,26 1,71 0,22 1,49 Fit 0,52 4,60 1,92 0,00 1,92 Fit 0,94 6,97 3,55 0,02 3,52 Fit 0,27 3,30 1,44 0,00 1,44 Fit 1,53 7,46 2,51 0,00 2,51 Fit

(32)

A.2.2. Aggregate stability

Table 6: mean CND and median CND.

Kolom1 Kolom2 Kolom3

Observations CND_mean CND_median

1FAB1 3 2,5 1FAC1 5,5 4 1FMB1 2,4 2 1FMC1 3,75 3 1HAB1 7,1 7 1HAC1 4 3 1HMB1 4,5 4 1HMC1 4 3 2FAB1 1,7 2 2FMB1 2,7 2 2FMC1 4,1 3,5 2HAB1 1,9 2 2HAC1 7,2 6 2HMB1 3,3 3 2HMC1 2,9 2,5 3FAB1 7,15 6,5 3FAC1 5,55 4 3FMB1 3,35 3 3FMC1 5,4 4 3HAB1 4,2 4 3HAC1 3,95 3 3HMB1 3,75 3 3HMC1 4,4 4 4FAB1 5,45 4 4FAC1 13,6 7 4FMB1 4,55 4 4FMC1 4,05 4 4HAB1 1,95 2 4HAC1 4,05 4 4HMB1 3,85 3,5 4HMC1 3,75 3 5FAB1 4,3 4 5FAC1 2,95 3 5FMB1 2,9 3 5FMC1 5,45 5 5HAB1 4,35 4 5HAC1 3,8 4 5HMB1 4,5 4 5HMC1 3,35 3

(33)

A.2.3. Crust description

(34)

A.3. Statistical analysis

A.3.1. Regression SOM and OC

A.3.2. Categorical data for chi square test

Table 7: Categories for chi square test.

Covering Coarse material Plant material Holes Thickness

fragmentary not much not much

not

much thin fragmentary not much not much

not

much thin fragmentary not much not much much thick fragmentary not much not much

not

much thin continuous not much not much

not

much thick

fragmentary much much much thin

continuous much not much much thick

fragmentary not much not much much thick

fragmentary much much

not

much thin fragmentary much not much

not

much thick continuous not much not much

not

much thin

(35)

continuous not much not much

not

much thick fragmentary much not much

not

much thick continuous not much much

not

much thick

not

much thin

not

much thick fragmentary not much not much

not

much thin fragmentary much not much much thin continuous not much not much

not

much thick continuous much not much

not

much thick fragmentary not much much

not

much thin

not

much thin continuous not much much

not

much thick

not

much thin

continuous much not much much thick

continuous not much not much

not

much thick fragmentary not much much

not

much thin fragmentary not much much

not

much thin continuous not much much

not

much thin

not

much thin

not

much thick fragmentary much not much much thick

(36)

A.3.2. Contingency tables from chi squared test

Not much Much

Fan 7 12

Hill 16 4

Table 8: Contingency table for amount of coarse material on fan and hill.

Fan 18 1

Hill 14 6

Table 9: Contingency table for amount of pores on fan and hill.

Thin Thick

Fan 14 5

Hill 7 13

Table 10: Contingency table for thickness on fan and hill.

Anthyllis 8 11

Macrochloa 15 5

Soil crust distribution and development in a semi-arid landscape dominated by Macrochloa tenacissima and Anthyllis cytisoides A field work research in the Rambla Honda Basin, Almería Province, SE Spain