Keywords

1 Introduction

In the Libyan semi-arid Mediterranean region, erosion-related soil degradation has a considerable impact on vulnerable soils, which already experience shallowness and have been subjected to organic matter decline. The region of Al-Jabal al Akhdar suffers from the most severe natural soil erosion in Libya. This phenomenon, according to FAO (1969) is because the relatively high rainfall promotes increased runoff. Some of the studies conducted in the region were very limited in terms of scale and time. However, since these studies applied direct measurements in the field, they were generally able to relate the problem of degradation to land use. Gebril (1995) observed that the agricultural activities by local people, such as clearance of vegetation, often lead to serious accelerated erosion, especially on the steeper slopes. There are several types of water erosion such as: splash, sheet, rill and gully erosion. The most serious types are sheet and rill erosion, which are responsible for most of the topsoil movement in the area. Gerbil’s (1995) measurements of soil erosion by means of erosion pins in the Al-Jabal Alkhdar region showed that considerable soil material was transported by surface wash processes.

The study carried out by Selkhoze Prom Export (1980) was exceptional because it covered the whole region with very detailed investigation. The study reported that during the second half of the twentieth century, intensive forest fires and human-induced woodland removal to establish field and horticulture crops had taken place widely in the area. Some parts of the deforested land were utilized successfully, while the rest was left fallow for various practical reasons, resulting in severe erosion in much of the area. The study estimated that about 79% of Al-Jabal Alkhdar north of the 200 mm rain isoline (i.e., about 14,000 km2) was to some degree affected by water erosion, mainly types of sheet erosion and linear forms of gulling. According to this study the main factors affecting the development of erosion processes are: relatively heavy downpours of rain on the uplands. absence of adequate vegetation cover, particularly during heavy rainfall events and periods of strong winds, slopes of different steepness and length, low permeability of the fine textured lower horizons of deeply developed soils, small humus content, absence of suitable soil tillage technology and regulated grazing, and low soil moisture holding capacity.

Classification of soil erosion was made by the above mentioned body from data collected during large-scale soil surveys (1: 50,000) and soil erosion maps were produced. The description of the degree of soil erosion was made by comparison of the thickness of the A1 and A1+B1 horizons of an eroded soil with the thickness of the A1 and A1+B1 horizons of standard non-eroded soils used as benchmarks, although the establishment of pedogenetic benchmarks was not rigorously applied. Moreover, since this comprehensive and detailed investigation was carried out forty years ago, noticeable changes could have taken place in the region as a result of continued intensive human activities. Consequently, the reliability of the Selkhoz Prom soil erosion maps for the present environment is questionable.

Other studies have applied predictive relations in order to assess soil erodibility and erosion-related degradation. Those predictive relations were developed under different environmental conditions and used in the region without validation. Therefore, the accuracy of their results could be in doubt. GEFLI (1975) stated that in the major part of the Al-Jabal Alkhdar area soil erosion is limited (GEFLI 1975), although, in some plots with shallow soils erosion has occurred, in which reddish runoff has been noted during the heavy rains, particularly when the first rains fall after the dry season. The study estimated soil loss based on the universal soil loss equation (USLE), where the rainfall erosivity factor (R) calculated from rainfall data and soil erodibility factor (K) were estimated by analogy with comparable soils studied in Tunisia. A weakness in this study was the estimation of soil erosion by applying a predictive equation USLE (Wischmeier et al. 1971) without validation and estimating K from soils in Tunisia instead of measuring the actual soil loss in the region. Field observations carried out in the GEFLI study noted few visible signs of soil erosion during the rainy season (1973/1974), while no traces of runoff were recorded within the study area and no rills were recorded. The study concluded that there was a very little, or no, danger of severe soil erosion for the deep soils. In contrast, slowly permeable, shallow and sloping soils can be affected and subjected to excessive soil losses.

OMU (2005) investigated soil erodibility for more than 50 sites across the Al-Jabal al Akhdar region using the K-USLE index. The results showed that most of the sites have medium erodibility. However, there were some highly erodible sites that were protected by good vegetation cover, which indicates the importance of the plant cover factor. In other words, severe erosion is likely to occur if these erodible soils lose the protection of the existing plant cover. As in the case of the GEFLI (1975) research, this study also applied the USLE predictive formula (Wischmeier et al. 1971) without validating the results against any direct measurements. It is argued here that a better understanding of soil degradation-related erodibility in this part of Libya is required in order to design appropriate soil conservation measures and that such measures must be better related to more sustainable land management practices.

There has been no previous attempt to develop indicators for soil erodibility estimation for the region of study in particular. This makes the investigation of soil erodibility and the causes of erosion-related soil degradation in this region a priority, with the aim of providing a formula to estimate soil erodibility in a spatial framework and relate this to the hazard of ecosystem degradation and to the local land use practices. This study aims to focus mainly on a better understanding of the physical processes of soil degradation, with particular emphasis on the assessment of soil erosion in relation to soil type and land use practices, and an evaluation of why conventional approaches to soil erosion prediction in Mediterranean environments often fail to fully reflect the degree of soil degradation observed in the field. The main objective is to develop a method of predicting soil erodibility based on easily measurable and appropriate soil properties.

2 Materials and Methods

2.1 Study Sites and Soil Description

The study was conducted on the northern and southern slopes of Al-Jabal Alkhdar (Fig. 1) using different criteria. The Mediterranean type climate has an annual rainfall range of 400–500 mm, falling mainly in the period from October to April. Parts of the area have been cleared of natural vegetation and are currently cultivated with rain-fed barley or are used for tomatoes and fruit production, with some fields left as fallow.

Fig. 1
figure 1

Sites of study at the northern and southern slopes of Al-jabal Alkhdar

Full size image

2.1.1 Field Erosion Plots Experiment, Northern Slopes

Erosion plots were established during the 1995–1997 seasons, at two locations in Al-Jabal Alkhdar; the first location, at Al-Hamama, lies on average slopes of 12%, and the second location, at Shahat, was located on average slopes of 9%. Each of the two locations was divided into three parts. The first part was covered with forest trees, the second was cultivated with barley after deforestation, and the third part was left fallow after forest removal. Three field plots (3 x 10 m each) were laid out randomly to evaluate soil loss by rain from a Mediterranean semi-arid Calcic, Typic, Lithic Rhodoxeralfs and Lithic Rendolls) soils (Fig. 2). Rain-gauge recorder was installed at each experiment location so that rain events data can be recorded.

Fig. 2
figure 2

Soil erosion plots for soil loss measurements at the study locations; (a) Alhamama, and (b) Shahat

Full size image

Soil loss amounts (A) from the field plots were measured for each rain event as well as the whole season, and rainfall erosivity factor (R) values were calculated using the rain-gauge recorder data. Hence, soil erodibility factor values (K) were calculated using the basic eq. A = R . K, where A is the annual soil loss (ton\ha), (Wischmeier et al. 1971).

2.1.2 Estimation of Soil Erodibility by the Rainfall Simulation Experiment, the Northern Slopes

This study was carried out during the 2009 season on soil samples from the Al-Jabal Alkhdar region, Lussaita – Al Hamama area which lies between longitude 21° 30′ to 21° 45′ E and latitude 32° 45′ to 33° 00’ N at elevations ranging from 20 to 400 metres above sea level. The soil samples were taken from the A horizon (0–15 cm) of cultivated soils of the four soil classes under investigation, Table 1. These soil classes are Typical Carbonate Red Ferrisiallitic Soils (Calcic Rhodoxeralfs), Typical Leached Red Ferrisiallitic Soils (Typic Rhodoxeralfs), Leached Red Ferrisiallitic Soils with a truncated profile (Lithic Rhodoxeralfs), and Red Carbonate Rendzinas (Lithic Rendolls). Four replicates of each soil type were collected, with total of 16 samples. The samples were dried and sieved through a 5 mm sieve. The method used for this experiment is based on the work of Meyer (1958) and was also used by Aknaro (1983) and Macharia et al. (1997). The experiment used trays of soil 25x25x20 cm in size (Fig. 3); the soil layer (5 cm thick) was packed above 5 cm layer of sand, and a cloth material was placed between the two layers to separate them. A gravel layer of 5 cm was added below the sand layer to facilitate the free drainage of water from the bottom of the tray which was perforated. The soil surface had a slope of 5° towards the collection container. An easily hand rotated rainfall simulator was used with one Floodjet (3/8 K SS45) spray nozzle, as used by Blanquies et al. (2003). The nozzle was placed at 3 m height on the top of metal rig to provide a uniform rainfall on the soil tray which was located beneath the nozzle to form the most suitable spray angle. The nozzle was connected to the pump via a plastic pipe to carry the pressurized water to the nozzle at 4–7 psi (approximately 0.3–0.5 bar). The pressure was monitored using a pressure gauge and several regulators were placed across the pipe to allow full control of the flowing water. The four sides of the rig were protected by a windshield. In order to keep steady water flow, two tanks were connected to provide a total capacity of 400 litres and supplied with constant water flow from the mains water supply. A drainage system was placed beneath the rig to carry the used water away.

Table 1 The measured physical and chemical properties of the investigated soil classes (Mean ± standard error)
Full size table
Fig. 3
figure 3

Investigating soil susceptibility to erosion using the simulated rainfall test; (a) simulated runoff, and (b) rainfall simulator

Full size image

2.1.3 Soil Erodibility Estimation Using Dynamic Laboratory Tests, the Southern Slopes

The study was carried out during the seasons (2015–2017) to estimate soil susceptibility to water erosion and to assess erosion risk on the southern slopes of Al-Jabal Alkhdar. Five areas were selected from west to east: Meseliba (Taknes); Marawa; Sirat Alia (Gandulah), Grehat (Gandulah) and Qasar Mestashi (Salantah), as shown in Fig. 1 and Table 2. These areas are within the most degraded lands in Al-Jabal Alkhdar, where grazing and rainfed cultivation represent the basic activity for many residents. Also, the indicators of desertification in this area are clear as a result of uncontrolled human activities, soil erosion and consequently decreasing the quantity and quality of natural vegetation. The dynamic laboratory tests included experiments that simulate the detachment of soil aggregates, the tests are listed below:

  1. (a)

    Water aggregate stability index: in this experiment, the wet sieving method was applied on surface soil samples. The method was reported by Morgan (1996) within soil erodibility indicators, and was used in the soil laboratory of Cranfield University in the United Kingdom on soil samples from Africa by Ekwue(1984), which was based on the work of Adams et al. (1958).

  2. (b)

    Instability index: the index = % silt + clay% / (% aggregats&>0.2 mm after wet sieving) - 0.9 (% coarse sand). The index was reported in Combeau and Monnier (1961), and listed in Morgan (1996) as an indicator of soil erodibility.

  3. (c)

    Detachability index: the index can be measured by dividing the residual weight on the sieve in experiment (a) on the shaking duration to obtain the rate of detachment in grams per minute. Higher residual weight in and lower detachment rate could indicate greater structure stability and resistance to erosion. The test was applied by Russell and Feng (1947), and pointed out by Kemper and others (1985).

Table 2 Some general characteristics of the southern slopes
Full size table

2.2 Erosion Risk Assessment of the Southern Slopes Using GIS

Various types of data and satellite imagery were used for the present study; the method involved integration of different thematic layers such as soil map and DEM in GIS environment (ArcGIS10.3.1 software). The produced erosion risk maps were prepared to illustrate and classify potential soil erosion on the southern slopes of Al-Jabal Alkhdar using the following criterion:

  1. 1.

    Soils with good content of clay, organic matter and iron oxides and with medium and deep depth (such as Alfisols) are highly aggregated soils. In this case the risk of erosion under similar climate will depend on the degree and length of the slope, and often the risk of erosion ranges between slight to medium.

  2. 2.

    Soils with good content of clay and medium organic matter content, such as the shallow calcareous Mollisols, are also aggregated soils, but these soils are shallow with undeveloped profiles due to its formation on slopes. The shallow Mollisols are less resistant to erosion compared to the deep Alfisols, and the risk of erosion ranges between slight to severe according to the slope characteristics.

  3. 3.

    Soils with fragile, shallow and weak soil structure due to the low content of organic matter, such as the soils of arid lands Aridisols, will have erosion risk ranges between medium and severe according to the slope characteristics.

    Note: the erosion hazards classification excludes climate and land use factors since all sites in this study are within the same climate and land use conditions.

3 Results and Discussion

3.1 Erosion Plots Experiment, Northern Slopes of Al-Jabal Alkhdar

The results showed that the largest soil and water losses occurred on the bare plots in both sites (Al-Hamama and Shahat) during the seasons 1995/1996 and 1996/1997. In Al-Hamama site, for example (Table 3), soil losses amounted to 1.62 and 4.14 tons/ha, organic matter 51.7 and 139.8 kg/ha, total nitrogen 3.7 and 10.1 kg/ha and available phosphorus as little as 3 and 4 g/ha in the first and second seasons, respectively. In comparison, the losses from the barley cultivated plots were less, with soil losses of 0.219 and 0.046 tons/ha, and organic matter losses of 0.7 and 1.4 kg/ha in the first and second seasons, respectively. Only traces of total nitrogen and available phosphorus were lost from Al-Hamama barely cultivated soil in both seasons. On the other hand, practically no losses occurred in the forest-covered plots, characterized by significantly higher organic matter, infiltration rates and aggregate stability. Losses from Shahat site were much lower than those from Al-Hamama site in both seasons, which were due to significantly higher soil depth and infiltration rates.

Table 3 The annual losses of soil components and plant nutrients during the tow seasons of study at Al-Hamama site
Full size table

3.2 Soil Erodibility Assessment

Table 4 shows values of soil erodibility factor K which were measured directly using the field erosion plots data and applying the basic eq. (A = R.K), where A is the annual soils losses ton\ha and R is the rainfall erosivity factor calculated using the rain-gauge rainfall data. The Measured K values were compared to the estimated K values which were assessed using the nomograph suggested by Wischmeier et al. (1971) which is based on a statistical relation between specific soil properties. It appears clearly that the nomograph overestimates values of the soil erodibility factor, and the reason may be that it was not originally designed for conditions similar to the current study area. The question therefore arises whether soil erodibility can be predicted by using formulae based on selected soil properties that affect soil structure, particularly, organic matter, clay content, carbonate content and Fe and Al oxides (Diaz-Fierros and Benito 1996). According to Abu Hammad (2005) the factor is dynamic and affected significantly by dynamic soil properties that were in turn influenced by environmental factors. This conclusion can explain why erodibility overestimation was repeatedly reported when K-USLE index was applied in the Mediterranean.

Table 4 Measured and estimated soil erodibility factor (K) values
Full size table

The results shown in Table 5 could confirm the contribution of certain soil properties to soil resistance to erosion (or soil susceptibility to water erosion). Organic matter, soil particle size distribution, aggregate stability and infiltration rate were all correlated to soil resistance. Soil susceptibility to erosion (soil erodibility) varied between soils according to these properties, and this can to some extent explain the significant differences in soil losses between the plots within the same site or between the sites. High clay content, moderate organic matter content, low content of silt, and very fine sand all contributed to a stable soil structure. The moderate aggregate stability of Typic and Calcic Rhodoxeralfs significantly improved their resistance to rain drop impact. These characteristics were reflected in moderate and stable infiltration rate, which is predicted to result in less water runoff and soil loss. On the other hand, relatively lower content of organic matter and higher content of silt and very fine sand have resulted lower aggregate stability and saturated hydraulic conductivity among the Lithic Rhodoxeralfs. Greater runoff and soil loss is expected on these soils, as they are more erodible and less resistant to the processes of erosion.

Table 5 The correlation (r) between measured soil erodibility factor (K) values with some physical soil properties
Full size table

3.3 The Simulated Rainfall Test

As a dynamic property, soil erodibility values may vary with different rainfall characteristics. Table 6 shows soil losses from the four soil types under investigation under several levels of rainfall intensity and duration. Low values of soil loss were generally recorded in the experiment. This is consistent with the field observations and the previous field plot experiment data, all of which showed a low to moderate level of soil erosion. The Rendolls were unable to maintain their stability under simulated rainfall conditions, resulting in a significant decrease in hydraulic conductivity and increase in runoff and soil loss. These results for Lithic Rendolls agree with findings by other researchers that, due to the changeability in some erodibility-related properties such as wetting, sealing and crusting, the resistance of the soil surface to interrill erosion can vary during rainfall events (Kuhn and Bryan 2004).

Table 6 Mean and standard deviation of measured soil loss values (kg.m−2.hr.−1) for the rainfall simulation test on four soil types under different rainfall intensity and duration
Full size table

The results showed soil erodibility values varied from 0.00 to 0.195 (t.MJ−1) for soil losses in the range 0.00–4.04 t.ha−1. The findings indicate that soil classes were differentiated relatively by the soil erodibility factor (Table 7, Fig. 4). Although all erodibility values fall within the low to low-moderate ranges, the statistical analysis indicates that soils have different levels of susceptibility. Calcic Rhodoxeralfs and Typic Rhodoxeralfs have the lowest soil erodibility while Lithic Rhodoxeralfs and Lithic Rendolls have higher erodibility. The relative erodibility of the four soils follows this order: Lithic Rendolls and Lithic Rhodoxeralfs greater than Calcic Rhodoxeralfs and Typic Rhodoxeralfs.

Table 7 Mean values of the measured soil erodibility factor of the four soil classes
Full size table
Fig. 4
figure 4

The trend of erodibility of different soil types with the increase of rainfall intensity for 50 minutes rainfall duration

Full size image

The lower soil loss and erodibility values associated with Typic Rhodoxeralfs and Calcic Rhodoxeralfs reflect the effect of higher hydraulic conductivity on these soils compared to Lithic Rendolls and Lithic Rhodoxeralfs soils. Furthermore, the moderate aggregate stability of these soils significantly improves their resistance to rain drop impact. This may also reflect the associated effects of higher clay content with moderate organic matter content and low content of silt + very fine sand (Table 1). Similar results were reported for Spain (Sanroque et al. 1990) and Ethiopia (Fufa et al. 2002; Adekalu et al. 2006) showing that structural stability and organic matter content are the properties most related to erodibility. Fine texture and strong, medium to coarse, angular and sub-angular blocky or prismatic structure were the factors behind the low erodibility of Red Mediterranean soils in Greece (Yassoglou et al. 1997).

In Spain, Diaz-Fierros and Benito (1996) stated that calcareous soils and Cambisols associated with Rendzinas have moderate to good stability and that organic matter was the main agent stabilizing soil structure. But in the present study this was not the case, as the field observations and the rainfall simulation tests showed them to be more easily eroded relative to Typic and Calcic Rhodoxeralfs, which have similar levels of clay, organic matter and aggregate stability. The relationship between aggregate breakdown under raindrop impact and clay type, crusting, seal formation and decreasing infiltration may help to explain the change in erodibility and soil loss in Lithic Rendolls during rainfall events. Selkhoz Prom Export (1980) reported a relatively higher content of swelling clays in Lithic Rendolls than in Rhodoxeralfs. In Al-Jabal al Akhdar region of Libya, Rhodoxeralfs are rich in kaolinite and very poor in their content of smectite (Bin Mahmoud 1995). It seems that it is difficult to predict the behaviour of soils rich in swelling clays in relation to erodibility. The effects of drying and wetting can lead these minerals to greater swelling and shrinkage. Consequently, these soils cannot maintain their pore spaces during wetting and swelling resulting in progressively slower infiltration capacity (Morgan 1996).

In spite of the fact that soil erodibility is a dynamic property affected by rainfall characteristics, it can be suggested that relative soil erodibility will be affected mainly by the inherent soil properties. This hypothesis is confirmed by the fact that the order of average erodibility values of all soil types was the same in this experiment and maintained the same trend under different intensities of simulated rainfall (Fig. 4). Assouline and Ben-Hur (2006) reported that with gentle slopes; interrill soil erodibility values were constant and independent of the effect of rainfall intensity soon after the beginning of runoff. In the same way, Kinnell (2000) suggested that soil erodibility will be constant as soon as steady state occurs and sediment concentration stabilizes.

Table 8 shows the correlation coefficient between soil erodibility and various soil properties. The erodibility was significantly correlated with only two parameters, water stable aggregates (WSA% >2 mm) and saturated hydraulic conductivity (Ks cm.min−1), and both were negatively correlated. Although neither organic matter nor soil particle size (clay, silt + very fine sand) showed any significant correlation, the results suggest a combined positive effect of organic matter and clay content on the stability of soil structure, in addition to maintaining moderate hydraulic conductivity within the soil layers. Assouline and Ben-Hur (2006) found that erodibility values can vary due to surface sealing which causes noticeable changes in upper soil structure and permeability.

Table 8 The correlation coefficient (r) of measured soil erodibility with various soil parameters
Full size table

A regression analysis (Table 9) was used to find out the most relevant properties to soil erodibility assessment, using the most correlated soil properties to erodibility (Table 8). The regression analysis shows a significant relationship between a combination of water aggregate stability index (WSA% >2 mm) and saturated hydraulic conductivity (Ks cm.min−1) with soil erodibility. Similarly, Le Bissonnais (1996) suggested that water stable aggregates can efficiently estimate soil erodibility if an improved method is used. Ben-Hur and Agassi (1997) found a significant correlation between final infiltration rate (FIR) and interrill erodibility (Ki), which suggests that the soil properties can be a useful estimate of erodibility.

Table 9 Regression equation of the relations between measured erodibility (K) and measured soil properties
Full size table

3.4 The Dynamic Laboratory Tests of Water Aggregates Stability (Wet Sieving)

Statistical comparison was made using the correlation analysis to test the association between the dynamic soil erodibility indicators under investigation with some field parameters such as soil depth, soil resistance to penetration, final filtration rate and bulk density (Table 10). The dynamic laboratory indicators were significantly correlated with the field parameters of the southern slopes of Al-Jabal Alkhdar. The success of the dynamic indicators in the expression of erosion-related land degradation was reported in the literature, which considered that wet sieving-soil stability indices were capable of predicting soil erodibility and degradation due to their direct relationship to climate, soil properties, parent materials, vegetation and land use. Moreover, due to the proximity of simulations to the natural conditions of the field, these indices could play a very helpful tool for erosion risk assessment (Cerda 2000; Barthes et al. 2000; Rosse 2002). In particular, the dynamic stability indicators will be useful, low-cost and easy-to-apply tools to describe the risk of erosion when comparing different types of soils or land uses.

Table 10 Correlation matrix between soil erodibility indicators and some field parameters
Full size table

The multiple regression tests were applied to select the dynamic stability indicators that are most related to the field state. In Table 11 some of the field parameters were employed in set of regression equations to relate them to the dynamic indicators, and to test the indicators validity to estimate erodibility of the study area soils, and to determine that soil erodibility is a measurable and useful index for estimating actual soil erosion.

Table 11 The multiple regression equations of the soil erodibility indicators and field parameters
Full size table

Table 11 showed that the following dynamic stability indicators, AgSt>1 mm index and Detachability index 1 mm, have the highest (R2) values when using the most relevant field parameters set together in the multiple regression equation. Therefore, the findings indicate a relative capability for the predictive formulae to broadly indicate erodibility in the field in the southern slopes of Al-Jabal Alkhdar. However, reflecting the reality of erosion-related soil degradation in the field is still questionable and needs further investigation. Le Bissonnais (1996) suggested that water stable aggregates can efficiently estimate soil erodibility if an improved method is used. Aggregates stability in the present study was measured using wet sieving, which may not have the same effect of rain drops on soil aggregates. Rejman et al. (1994) suggested that considerable improvement can be made in soil erodibility measurement if the indicator is measured by using a rainfall simulator rather than wet sieving. However, due to the large areas under investigation and the lack of resources, the wet sieving method was the most appropriate means to achieve some of the objectives of this study such as mapping soil erosion (erosion risk assessment).

3.5 Erosion Risk Assessment

The areas that need appropriate management to limit the phenomenon of erosion-related soil degradation and land desertification can be recognised from the potential erosion risk maps produced in this study (Figs. 5, 6, 7, 8 and 9). The maps classified the lands according to the severity of erosion and indicated the slopes that require extra measures to conserve soil and resist water and wind erosion on the slopes. The required measures would control surface runoff using appropriate barriers that help soil preservation, water harvesting and moisture retention. Consequently, plant environment, such as soil characteristics, fertility and productivity, and its diversity would be improved. Figure 10 shows the role of some soil conservation measures in improving the natural environment on the southern slopes of Al-Jabal Alkhdar.

Fig. 5
figure 5

Potential erosion risk classes, Meseliba area

Full size image
Fig. 6
figure 6

Potential erosion risk classes, Marawa area

Full size image
Fig. 7
figure 7

Potential erosion risk classes, Grehat area

Full size image
Fig. 8
figure 8

Potential erosion risk classes, Sirt Alia area

Full size image
Fig. 9
figure 9

Potential erosion risk classes, Mestashi area

Full size image
Fig. 10
figure 10

Effectiveness of soil conservation measures on the semi-arid southern slopes in improving soil quality and productivity in Marawa area; (a) tide-ridges, and (b) semi-circular bunds

Full size image

From the erosion risk maps we notice that the areas with the highest erosion risks are found mainly within the shallow arid soils (Aridisols soil type in particular), however, Mollisols soil type (shallow Rendezina) were also vulnerable to noticeable degradation. But in both cases, it is often (the soil factor) linked to the steeper lands (slope factor) that were nearby active water ways that had the highest risk of erosion (Figs. 5, 6, 7, 8 and 9). As previously emphasized, this behaviour is expected within Aridisols soils because of their weak cohesion and resistance to erosion factors. As for the Rendezina soils, their shallow depth will always be a major weakness in their ability to resist erosion factors despite their relative cohesion and stability. However, as is the case in this example, most of the lands, especially the red Mediterranean soils, Alfisols, and a large proportion of Mollisols (Rendezina soils), fall within the degrees of erosion risk that can be controlled by applying some known agricultural and engineering soil conservation procedures. In this low-scale study, the assessment of soil erosion risk using GIS technology focused on the most accessible sites, however, the assessment can be expanded in future to most affected areas. The produced maps of potential erosion hazards will definitely help in highlighting the fragile areas that should maintain the natural plant vegetation, the areas that allowed for restricted and limited agricultural uses, and the areas that are flexible for a wide range of crops.

4 Conclusions

The validation of estimated K-USLE erodibility values against field observations and field plot data in the present research showed that the estimating formula overestimated soil erodibility under semi-arid Mediterranean Libyan conditions. In the present study, simulated rainfall tests found that aggregate stability and saturated hydraulic conductivity together explained most variations in measured soil erodibility of the soil types under investigation. Although soil properties play the main role in soil resistance to erosion, the rainfall simulation test showed that rainfall characteristics, intensity and duration, can contribute to the changeability of erodibility. It is very important to determine the dynamism of the erodibility factor in relation to the climate effect, which can greatly affect the accuracy of erodibility predication. The dynamic nature of erodibility can also explain the unexpected lower resistance of Lithic Rendolls soils and it is concluded that change over time in these soils is probably due to the physical effects of swelling clays. The finding of this study suggest that the test for water stable aggregates (WSA) is a fairly reliable predictor of soil erodibility and relate to the fact that it is a parameter that combines the effects of several soil particle binding agents. It achieves this by directly measuring the cohesive forces that determine soil stability against water erosion by simulating the destructive action of rainfall and water action through wet sieving. One of the most important aspects of this research was estimating soil erodibility on a spatial basis by using GIS technology with information on soil properties and distribution patterns derived from existing national soil survey maps and reports, and from the soil classification systems employed to describe soil map units. Mapping soil erodibility and erosion risks based on soil survey data can undoubtedly be a useful tool in soil conservation-related policies and land use planning in Libya. The present study has highlighted the importance of WSA% and Ks as parameters for estimating soil erodibility, however, considering their often limited availability, further investigation to find other more easily measured, or readily available, parameters should be encouraged, so that a wider-scale spatial prediction of soil erodibility can be achieved in Mediterranean soil environments.