نوع مقاله : مقاله پژوهشی
نویسندگان
1 مؤسسه تحقیقات خاک و آب، سازمان تحقیقات، آموزش و ترویج کشاورزی، کرج، ایران.
2 بخش تحقیقات خاک و آب، مرکز تحقیقات و آموزش کشاورزی و منابع طبیعی استان اصفهان، سازمان تحقیقات، آموزش و ترویج کشاورزی، اصفهان، ایران.
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Background and Objectives: Estimating soil hydraulic properties using pedotransfer functions is both feasible and commonly practiced, leveraging readily measurable soil properties as input data in soil and water research. These estimated properties are essential for evaluating water and solute movement, as well as for managing water and salinity at various scales. The accurate description and estimation of soil water retention curves have been a primary focus of research for over six decades. However, estimating coefficients in empirical models remains challenging and time-consuming. Consequently, pedotransfer functions (PTFs), including neural network analysis and texture-based tables, have been employed to predict SWRC coefficients. The ROSETTA software package estimates soil hydraulic properties such as van Genuchten model parameters using a combination of bootstrap and artificial neural networks. Despite the use of pedotransfer functions for predicting soil hydraulic parameters in Iran, no study has specifically examined the application of these functions in estimating hydraulic parameters across different soil profile layers, particularly in salt-affected soils. The objective of this study was to evaluate the performance of the ROSETTA pedotransfer functions in estimating the hydraulic parameters of the Van Genuchten model across different layers of soil profiles in the salt-affected soils of the Sistan Plain, assessing these as preliminary input data for water management and salinity evaluation.
Materials and Methods: A total of 312 profiles was excavated in the Sistan Plain. Based on field observations and morphological characteristics (including color, texture, structure, presence of carbonates, gypsum, and roots), pedogenic (genetic) horizons were identified in each profile (ranging from 3 to 5 horizons per 100 cm profile). For statistical analysis and modeling purposes, the identified horizons were grouped into three pedogenic layers: surface (first), middle (second), and subsurface (third). Physical soil characteristics, including the sand, clay, and silt content, as well as bulk density and saturated hydraulic conductivity, were measured for each profile layer. Soil hydraulic properties, including Ks and the parameters of the van Genuchten model (residual (θr) and saturated (θs) soil water content, inverse of air entry pressure (α) and soil pore distribution (n)) were estimated by the ROSETTA software package. In order to evaluate the results, statistical indices of root mean square error (RMSE) and Nash-Sutcliffe coefficient (NSE) were used to compare the measured saturated hydraulic conductivity and saturated moisture with their estimated values.
Results: According to the results, the highest standard deviation among the estimated hydraulic properties occurred for Ks in all three examined layers. The estimated Ks values were significantly highest in the third layer, followed by the second and first layers. In contrast, the estimated θs and θr values showed a significant decrease with increasing depth from the soil surface, while the n parameter showed a significant increase with depth. No significant difference was observed for α between the profile layers. A significant correlation was obtained between the estimated hydraulic properties and the measured soil properties, including bulk density, and the percentages of sand, silt, and clay. The measured values of saturated water content (θs), field capacity (FC), permanent wilting point (PWP), and plant available water (PAW) showed positive and significant correlations with clay and silt content. There was no significant correlation between the measured and estimated Ks values. Furthermore, the RMSE and NSE values from the regression relationship between measured and estimated Ks were 16.41 and -0.11, respectively. The RMSE and NSE values for the regression relationship between measured and estimated saturated water content were 0.06 and -1.11, respectively. The negative NSE values indicate that the mean of the measured data provides better predictions than the ROSETTA estimates, confirming the high uncertainty of the ROSETTA model in these salt-affected soils.
Conclusion: Analysis of the RMSE and NSE values from the regression relationship between measured and estimated Ks revealed high uncertainty in ROSETTA estimates for salt-affected soils, indicating a potential need for the development of regional pedotransfer functions that incorporate soil chemical properties (such as salinity and sodium content) as input variables. Also, more detailed considerations are suggested for soil and water management in the region.
کلیدواژهها [English]