دانشگاه محقق اردبیلی مدل سازی و مدیریت آب و خاک 2783-2546 6 1 2026 03 21 Evaluation of point pedotransfer functions in estimating field capacity and permanent wilting point water content ارزیابی توابع انتقالی نقطه‌ای در برآورد رطوبت ظرفیت مزرعه و نقطه پژمردگی دائم 192 207 4097 10.22098/mmws.2025.18164.1652 FA پریسا کهخا مقدم استادیار گروه مهندسی آب، دانشکده آب و خاک، دانشگاه زابل، زابل، ایران Journal Article 2025 08 23 Introduction The amount of available water in the soil for crop use, called available water (AW), is described as the capacity of the soil to retain water and make it available for root uptake. Irrigation management, agricultural projects, and soil water balance are some of the well-known practical applications related to AW. It is generally accepted that matric potentials at -33 kPa and -1500 kPa represent FC and PWP, respectively. An attractive alternative to direct and often laborious measurements of these soil hydraulic properties is their estimation using soil pedotransfer functions (PTFs). The efficiency of PTFs is not the same in different locations and conditions, so the performance of each PTFs should be evaluated according to the soil characteristics of each region. The Sistan Plain is one of the strategic border areas of eastern Iran and the edge of the desert. The main objective of this study is to evaluate the performance of 23 different PTFs models in predicting soil water retention at matric potentials of -33 kPa and-1500 kPa in the Sistan Plain. In order to determine the best PTF that is appropriate for the conditions of the study area, 100 soil samples from the Sistan region were used. Materials and Methods In order to conduct this study, 100 soil samples were collected from a part of the Sistan Plain for laboratory analysis and transported to the laboratory. Particle size distribution (sand, silt, and clay) was measured by hydrometric method, organic carbon by Walkey-Black method, and bulk density by cylinder method. Twenty-three PTFs were evaluated to predict water content (gravimetric and volumetric), including twelve functions for matric potential of -33 kPa and eleven functions for matric potential of -1500 kPa. PTFs can also be categorized based on the type of prediction, namely point and continuous PTFs, and in this study, point PTFs were evaluated. The selection of PTFs is limited to PTFs that use soil properties that are present in the data set of the present study. The sum of squared error method was used to calibrate the models under study. In this study, NRMSE, ME, r, and RES criteria were used to fully evaluate the models. Results and Discussion The results showed that the changes in the amounts of sand, silt, and clay ranged from 14 to 56, 27 to 52, and 14 to 57 percent, respectively, which resulted in the creation of textural classes of loam, clay loam, sandy clay loam, clay, silty loam, and sandy loam. The function developed by Aina and Periaswamy (1985) with NRMSE, ME, r and RES values of 0.61, 0.15, 1.48 and -0.148, respectively, had the least closeness to the measured values of θ33. The results also show that among the PTFs examined for estimating θ33, the PTF developed by dos Reis et al. (2024) with NRMSE, ME, r and RES values of 0.10, 0.01, 1.04 and -0.012, respectively, and in the next rank and with a slight difference, the PTFs developed by Oliveira et al. (2002) and Minasny and Hartemink (2011) with NRMSE, ME and r values of 0.10, 0.02 and 0.05 (same for both functions) had the best agreement with the measured values of θ33. This is while the function developed by Aina and Periaswamy (1985) with NRMSE, ME, r and RES values of 0.61, 0.15, 1.48 and -0.148 respectively had the least closeness to the measured values of θ33. The summary of the evaluation of eleven PTFs for estimating θ1500 shows that the Dijkerman (1988) and Aina and Periaswamy (1985) functions provided the best performance among the functions studied with ME of 0.00 and -0.01, NRMSE values of 0.15 and 0.16, and r of 1.02 and 0.95, respectively. The function developed by Oliveira et al. (2002) showed good performance for estimating θ33, but it was the least close to the measured values of θ1500 with NRMSE, ME, r and RES values of 0.84, 0.14, 1.83 and -0.138, respectively. Then with the aim of improvement, the high-performance economic functions were selected and recalibrated. In estimating θ33, the performance of both the dos Reis et al. (2024) and Oliveira et al. (2002) functions improved (albeit only slightly) by reducing NRMSE (0.08) and ME (0.00) to reach r of 1.00. The Dijkerman (1988) and dos Reis et al. (2024) functions also improved the estimation of θ1500 with NRMSE, ME, and r of 0.14, 0.00, and 1.00, respectively. Conclusion The results showed that among the PTFs used to estimate θ33, the functions developed by dos Reis et al. (2024) and Oliveira et al. (2002) show the best agreement with the measured values. The functions developed by Dijkerman (1988) and Aina and Periaswamy (1985) also provide the highest performance in estimating θ1500. The results of this study also show that some of the new functions presented in this field can provide good performance compared to the basic functions in predicting soil moisture content. The performance of PTFs for the study area is not affected by the number of their input components, and PTFs that require fewer inputs will not necessarily have lower performance, and this is because PTFs depend on the location where they are developed and also on the soil structure. In order to improve the results of the functions of dos Reis et al. (2024) and Oliveira et al. (2002) were recalibrated to estimate θ33 and the functions of Dijkerman (1988) and dos Reis et al. (2024) were recalibrated to estimate θ1500. The selection of the aforementioned PTFs was based on high performance and minimum required input components, or in other words, the economy of the function. Recalibration further improved the performance of the aforementioned functions. θ33, θ1500 and AW are key components in a wide range of studies such as crop modeling, hydrological modeling, water resources management, soil nutrient cycle modeling and soil pollution modeling, therefore the results presented in this study can be used in predicting soil moisture content and AW for the study area, using the soil transfer function technique as an input parameter in various modeling. مدل‌های گیاهی و هیدرولوژیکی اغلب به مؤلفه‌های ورودی در مورد آب قابل‌دسترس برای گیاه (AW) نیاز دارند. با این وجود به‌دست آوردن چنین مؤلفه‌هایی پرزحمت و پرهزینه است. توابع انتقالی خاک (PTFs) یکی از روش‌های جایگزین برای تعیین خواص فیزیکی خاک مانند محتوای رطوبت خاک است. لذا هدف این پژوهش ارزیابی بیست و سه PTFs نقطه‌ای برای تخمین رطوبت خاک در پتانسیل‌های ماتریک ۳۳- (θ33) و ۱۵۰۰- (θ1500) کیلوپاسکال، برای صد نمونه از خاک‌های دشت سیستان است. در این پژوهش تلاش بر این بوده است که از جدیدترین توابعی که در این زمینه ارائه شده است نیز استفاده شود. نتایج حاکی از آن بود که تابع دوس‌ریس با مقادیر NRMSE، ME، r و RES به‌ترتیب برابر 10/0، 01/0، 04/1 و 012/0- و سپس تابع اولیویرا و همکاران با مقادیر NRMSE، ME، r و RES به‌ترتیب برابر 10/0، 02/0، 05/1 و 016/0-، بیش‌ترین انطباق را با مقادیر اندازه‌گیری شده θ33 داشته است. برای تخمین θ1500 نیز تابع دیجکرمن با مقادیر NRMSE، ME، r و RES به‌ترتیب برابر 15/0، 00/0، 02/1 و 003/0- و در رتبة دوم تابع آینا و پریاسوامی با مقادیر NRMSE، ME، r و RES به‌ترتیب برابر 16/0، 01/0-، 95/0 و 009/0- بهترین عملکرد را ارائه کردند. هم‌چنین نتایج این پژوهش نشان می‌دهد بهترین توابع در تخمین θ33 و θ1500 به‌ترتیب تنها به درصد شن و درصد رس بستگی دارند، لذا برای منطقة مورد مطالعه PTFs که به ورودی‌های کم‌تری نیاز دارند، لزوماً عملکرد پایین‌تری نخواهند داشت، بلکه عواملی مانند تعداد نمونه‌های خاک، ساختمان خاک و مکانی که توابع مورد بررسی توسط آن‌ها توسعه یافته‌اند نیز بر عملکرد آنها مؤثر است. در ادامه PTFs اقتصادی‌تر و با عملکرد بالاتر مورد واسنجی مجدد قرار گرفتند. در تخمین θ33 عملکرد هر دو تابع دوس‌ریس و اولیویرا و همکاران با کاهش NRMSE (08/0) بهبود یافته‌اند. توابع دیجکرمن و دوس‌ریس نیز با مقادیر NRMSE برابر 14/0، تخمین θ1500 را بهبود بخشیده‌اند. θ33، θ1500 و AW مؤلفه‌های کلیدی در طیف وسیعی از مطالعات مانند مدل‌سازی گیاهی، مدل‌سازی‌ هیدرولوژیکی، مدیریت منابع آب، مدل‌سازی چرخه مواد غذایی خاک و مدل‌سازی آلودگی خاک هستند؛ بنابراین، نتایج این پژوهش می‌تواند در مباحث مربوط به مدیریت آبیاری و حفاظت از خاک در منطقه مورد مطالعه به‌کار گرفته شود.

https://mmws.uma.ac.ir/article_4097_3a30d37e4bacf86af02534dda4bd20a6.pdf