ارزیابی برخی مدل‌های نفوذ آب به خاک در کاربری‌های اراضی مختلف

Water infiltration into the soil is one of the most important components of the hydrological cycle. Land use cover changes are affected by human-induced activities and growth socio-economic factors expansion of the forests, grazing, agricultural activities government policies and environmental factors such as drought. In the rural region, people encroach on forest areas for conversion to other uses, mainly cropland and rangeland though the land is not suitable for these purposes. The conversion of forests to other land uses may cause increase of water erosion, mass movements, and soil compaction by trampling and alteration of the hydrology cycle, among others. The aim of this study was to analyze the effect of some land uses on the amount of water infiltration and parameters of some infiltration models in forests, rangeland and agricultural land uses. This research was conducted at Fandoghlou sylvan region, located in Ardabil Province, that districted by Namin, Astara and Hir counties (38?, 10? to 38?, 30? East and 48?, 30? to 48?, 45? North). The acreage of the study area is about 16000 hectares. The Fandoghlou forest region is a continuance of Gilan semi tropical forest. The elevation in the catchment ranges between 1500 and 1850 m, and average of annual precipitation was recorded 290.4 mm by Ardabil synoptic station. Water infiltration experiments were conducted in the Fandoghlou region of Ardabil using double ring infiltrometer in forest, pasture and agricultural land uses in three replications. Totally 9 experiments were carried out in saturated state. Experiments were continued until at least 3 consecutive head fall data of double ring infiltrometer have been constant. Minimum time of infiltration experiments were 120 min. Initial water content was measured before water infiltration tests beginning, that amount of water content in forest land use was more than that measured in pasture and agricultural land use. A disturbed soil sample was collected from each different land use to measure soil physical and chemical properties. The organic carbon percentage was measured by the organic carbon oxidation method. Bulk density and saturated hydraulic conductivity were determined by the core method (Undisturbed soil samples were taken with steel core samples of 98.125 cm3). Soil pH and Electrical conductivity were measured by a pH and a conductivity meter. In this study, water infiltration rate was estimated using the five infiltration models including Philip, Kostiakov, Kostiakov-Lewis, Green-Ampt and Horton. To determine the best model among mentioned models in each land use, three statistical parameters RE, RMSE and R² were used. Results showed that for three land uses, including forest, pasture and agriculture, the Kostiakov-Lewis (RE, RMSE and R² were equal to 1.614, 0.023 and 0.9119, respectively), Green-Ampt (RE, RMSE and R² were equal to 2.824, 0.056 and 0.9196, respectively) and Philip (RE, RMSE and R² were equal to 2.901, 0.046 and 0.8408, respectively) models with less statistical error are the best models for infiltration rate simulation, respectively. Minimum and maximum amount of A parameter of Philip infiltration model in pasture (0.179) and forest (0.2) land uses, S parameter of Philip infiltration model in forest (0.472) and pasture (0.1.486) land uses, a parameter of Kostiakov infiltration model in pasture (0.682) and forest (0.821) land uses, k parameter of Kostiakov infiltration model in forest (0.547) and pasture (1.347) land uses, a parameter of Kostiakov-Lewis infiltration model in pasture (0.217) and agricultural (0.379) land uses, b parameter of Kostiakov-Lewis infiltration model in forest (0.011) and pasture (0.044) land uses, c parameter of Kostiakov-Lewis infiltration model in pasture and forest (0.23) and agricultural (0. 21) land uses, A^' parameter of Green-Ampt infiltration model in forest (8.469) and pasture (66.966) land uses, B^' parameter of Green-Ampt infiltration model in pasture (13.188) and forest (14.516) land uses, k parameter of Horton infiltration model in pasture (4.514) and forest (11.23) land uses, were calculated.
Philip and Green-Ampt infiltration models have best closeness than other considered models in infiltration rate simulation, because both of them have physical base. Horton infiltration model at pasture and agricultural land uses had maximum statistical error and have an over-estimating on infiltration rate. In forest and agricultural land uses, Kostiakov-Lewis infiltration model has overestimated the infiltration rate and underestimated in pasture land use. In agricultural land use, Kostiakov infiltration model underestimated the infiltration rate and overestimated in pasture and forest land uses. Minimum sorptivity coefficient of the Philip infiltration model has been extracted in forest land use, because matric potential of this land use was less than pasture and agricultural land uses. The initial water content of forest land use was more than pasture and agricultural land uses. Maximum empirical (B) coefficient of the Philip infiltration model, also has been extracted in forest land use, that indicate in this land use soil structure was better than that of other considered land uses, because in forest land use the amount of organic matter was more than other studied land uses.