This study was conducted at the Halashi catchment, located about 30 km southeast of Kermanshah, Iran (Figure 1) where is a part of Zagros mountain chains, at the west of Iran. The study area (7357 ha) included forest (927 ha), rangeland (2229 ha) and agriculture (4200 ha). It lies between 34˚00'38"N to 34˚09'20"N and 47˚04'20"E to 47˚21'15"E (UTM 1984; zone 38 N). The minimum and maximum elevations above sea level are 1440 and 2720 m, respectively that means annual precipitation and temperature are 470 mm and 18˚C, respectively indicating a semi-arid region.

Geological layers contain limestone, sandstone, shale and marls deposits of Cretaceous, Pleistocene and Miocene forming a soil with moderate to good conditions in terms of CEC, pH, nutrients and soil depth for vegetation and agricultural activities. However, this catchment is currently suffering from anthropogenic environment degradation mainly deforestation, heavy grazing and improper agricultural activities representing main land use change in the Zagros regions of Iran. About 37% of the study area is occupied by rangeland and forest,

which are more sensitive to soil erosion due to topography and geological properties.

The maps of land use, geology, erosion form and vegetation for this study were prepared by using satellite images and topographic maps (1:50,000). The geology map was prepared based on geology quadrangle map of Kermanshah (on a scale of 1:100,000), using data obtained from the Geological Survey of Iran (GSI) [19] .

The map of forest and rangeland was prepared based on physiognomy method and nominated by two to three dominant plant species. Finally the border of each map was checked in the field. Vegetation type of rangeland was nominated by three dominants plant species based on their ground surface coverage (%). The above ground covers refer to the area covered by the vertical projection of the plant crown, stone and litter onto the soil surface. There were estimated on field using a quadrate plot (1 m²). The forest is dominated by oak (Quercus persica).

Thirty eight soil samples were collected from 0 - 20 cm depth within each land use followed by stratified random soil sampling and their coordinates were determined by global positioning system (GPS). The dried soil samples were sieved through 2 mm mesh sieve. Soil physicochemical characteristics were determined in the laboratory. The particle size distribution and soil texture were determined by hydrometric method [20] . The pH of the saturated soil paste was measured by a pH meter. Soil organic carbon was determined by the Walkley and Black method [21] . Total nitrogen was determined by the Kjeldahl method [22] . Available phosphorus was measured by the method of Olsen [23] . Exchangeable potassium was extracted by NH4OAc buffered at pH 7.0 and K in the solution was determined by flame photometer [24] . The sample was centrifuged and extracted solution run by atomic absorption spectrophotometer (Perkin-ELMER).

The data were statistically analyzed using SAS version 6.12 for variance of each soil variable including one way ANOVA (General Linear Models (GLM) procedure and Student Newman Keuls (SNK) test due to unequaled samples size) at 0.05 probability level.

The most parts of the studied area are occupied by fine grained materials of marl deposits including younger terraces, older terraces, Kashkan (marl) and Shabazan formations as shown in Figure 2. An outcrop of Kashkan Formation shows the presence of reddish claystone, siltstone and sandstone, inter-layered with conglomerate

and limestone. The topographic landscape of Kashkan includes hilly areas which are susceptible to piping and landslide incidence.

Shahbazan includes limestone with high holes and pores on exposed outcrops. The main areas of this catchment are occupied by Quaternary deposits comprising older and younger comprising marl, argillaceous, sand and silt deposits with middle layers of gravel, and limestone which originated from adjacent geological formations in upper parts (mainly Kashkan formation).

Gully and river bank are dominants erosion forms in the agricultural areas. In these areas, gullies occur and develop where geologic structures are susceptible to gully forming due to fine-grained materials such as clay and silt. It is human- induced by agricultural activities, vegetation clearance and farming practices as well, it can be developed by concentrated runoff that leads to fluvial incision [25] [26] .

Inter-rill (sheet erosion) with rill erosion was found as the dominant erosion feature in the rangeland and some parts of rain-fed areas where hillslope is 10% - 20%. Tillage practice and overgrazing contribute to accelerating this erosion. Rill erosion is initiated where overland flow becomes channeled and sheer stress (by overland flows) is more than shear strength of the soils [27] , causing severs sedimentation and SOC lost [28] .

Land slide in the studied site potentially refers to geological properties, while it occurs mainly through human-induced deforestation. The effective factors on landslide include slope, aspect, topography, climate, lithology, tectonic, vegetation cover, physicochemical and mechanical characteristics of sliding layer and human activities. This is considerable in forest area where Kashkan is dominant geological formation.

The smectite is a major mineral in both parent material and the soil of Kashkan formation [8] . As such, soil erosion was high especially in the areas with soils containing smectite (with high potential to absorb water) resulting in swell- crack occurrence and damaging the engineering structures in the study area. In the forest, soil erosion, sedimentation, piping, gulling and landslides often take place.

In some areas where marl layers are overlain by limestone, swelling smectite causes sliding of marl layers and shrinks upon dehydration, leading to cracking of clay rich materials [29] . Landslide occurrence is also related to the absorption of water by smectite during the rainy season. It can absorb water and can expand up to 30% of its original volume [30] . Field observation demonstrated that during the wet season the soil in lands with gully and landslide (where smectite is dominant mineral) absorbs water and expands, increasing cracking and piping which in turn, induce gully erosion and landslide occurrence (Table 1, Figure 3).

Ce = Cenozoic, Me = Mesozoic, Qtc = Quaternary (cultivate area), Ter = Tertiary, Qtf = Quaternary (gravel fan), Ka = Kashkan, Sh = Shahbazan, P = Plain, N = North.

Rangeland of study area includes one vegetation type namely Daphenea. Sp, Astragalus. sp, and Bromus sp. The above ground cover including plant canopy, bare soil, plant litter and stoniness in this type was estimated 41.7, 25.3, 3 and 30%, respectively. Furthermore, these respective values in understory of forest were 57, 18, 9% and 18% (Table 2), which in turn, both plant coverage and litter in the forest were more than rangelands.

The inventory of the RL and FL showed that Astragalus sp, Amygdalus sp., Festuca ovina, Quercus persica and Daphnae sp. with annual species are dominant vegetations in the study site. Although the forest is preserved, nomads allow their herd to graze from early spring until Mid-summer sharing deforestation and soil degradation. There is more considerable where forest canopy cover is less than 50% [31] .

Descriptive statistical analysis of soil variables and their ANOVA are presented in Table 3 and Table 4. The average amount of sand in the AL, RL and FL are 21.23%, 30.1% and 21.75%, respectively that was significantly higher (p ≤ 0.05) in the RL than those other sites. Mean soil clay in that orders were found 40.21%, 41.63% and 37.37% and no significant differences among them. However, the respective levels of soil clay in the AL, RL and FL are 38.03%, 31.33% and 40.52% revealing lower in the RL. Overall soil in the study area was mainly clayey and silty in nature due to considerable amounts of silt and clay, thereby susceptible to erosion hazards. High clay content would tend to increase water absorption which may promote piping and sliding through increasing dissolve phenomenon. Furthermore, higher slope in FL and RL (more than 15%) cause transport of clay particles.

The mean soil pH value of the agriculture, rangeland and forest are 7.72, 7.79 and 7.77, respectively indicating moderate alkalinity. The soils are moderately alkaline with pH value of 7.4 - 8.4. The pH of the soils in the study area was not expected to have a significant relationship with the soil degradation (Table 3 and Table 4). The average value of EC was found 0.85, 0.66 and 0.71 dS/m in the AL, RL and FL respectively and no significant difference (p ≤ 0.05) among them showing not effects for soil dispersion potential.

*based on dominant plant species; **understory cover.

*Mean with the same letters are not significantly different at p ≤ 0.05%, **EC = electrical conductivity.

SOC: The mean SOC level in the AL, RL and FL was 1.14%, 1.45% and 2.13%, respectively and showing significantly (p ≤ 0.05) different from each other among them (Table 5). It can be seen that the CV and the difference between maximum and minimum in the FL was lower than that of any other lands. The higher level of SOC in the FR and RL are related to relatively higher biomass production compared with AL, enhancing soil. The lower value of SOC in the AL as compared to other lands is associated with conventional tillage practice and burning of crop residues by the local farmers, especially in the hilly lands, increasing CO₂ emission.

Field observations showed that crop residues are burnt early autumn into the atmosphere. Improper tillage practices strongly accelerate soil erosion processes by destroying soil organic matter (SOM) and soil structure [14] . Relatively lower level of SOC in the RL is affected by overgrazing. This cause imbalance SOC through defoliation and thereby reducing plant canopy, forage production and litter accumulation as well as different slope gradient [9] [32] . Furthermore, heavy sheep grazing decreased about 16.5% of SOC [33] . In this area accelerated gully erosion and landslide incidence attribute to 60% SOC lost. In these areas of Zagros, highest loss in SOC reaches to 306 kg ha^-1 yr^-1 by soil erosion in that area [28] .

Total N: As shown in Table 5, total N in the AL, RL and FL is 0.14, 0.14 and 0.18 mg/kg, respectively. Significant lower of total N in the AL and RL related to crop uptake (without fallow and rotation practices), residue burning and tillage practices as well as improper livestock grazing. This is consistent with other areas, where N₂O is emitted into the atmosphere, causing global warming although only to a limited extent [3] [34] .

Available P: The ANOVA showed that P in the AL was significantly higher than that of the other sites (Table 5). The higher level of P in the AL may due to over application of P₂O_5, while, overgrazing and deforestation factors cause P deficit in the soil [3] [4] .

Over P input into soil can remain in the soils profile for a longer time compared to N, and consequently can be accumulated in lower drainage systems, rivers and lakes causing eutrophication. This process can be accelerated by conventional tillage practices and erosion compared with no-tillage [35] . This also

*Means with the same letters within rows are not significantly different (at p ≤ 0.05%).

related to high soil clay content in the study area. The P level in the run-off sediments increases with decreasing soil particle size [36] .

Exchangeable K: The mean exch. K was 342.5, 375.0 and 374.5 mg/kg in AL, RL and FL, respectively. As revealed by ANOVA, soil K was not significant different from each other among three lands use. There is related to dominants clay minerals, particularly smectie and alkaline pH (7 - 8). The main source of K in the soils is weathering of mica, vermiculite and smectite [37] . The high range between maximum and minimum levels of exchangeable K in AL (580 and 99 mg/kg) is due to different level of K input and adversely different crops uptake. Overall, sever depletion of SOC and soil NPK are occurred via converting rangeland and forest to rain-fed lands in the Zagros. High CV of K in rangeland (Table 4) is due to slope and relatively higher fragmented vegetation causing its spatial variation due to runoff compared with other sites. The removal of soil K by runoff is more than that of phosphorus [38] . However, field verification indicated that there are many sources of organic fertilizer in the study area including livestock manure, urban and rural wastewater that can be considered for improving NPK and OC content. It is an ecological sanitation approach to promote closed-loop flows of resources and nutrients from sanitation to agriculture.