SciELO - Scientific Electronic Library Online

 
vol.14 número2Sistema mecatrónico de desinfección de alimentos mediante luz LED UV-A índice de autoresíndice de materiabúsqueda de artículos
Home Pagelista alfabética de revistas  

Servicios Personalizados

Revista

Articulo

Indicadores

Links relacionados

  • No hay artículos similaresSimilares en SciELO

Compartir


Ingeniería agrícola y biosistemas

versión On-line ISSN 2007-4026versión impresa ISSN 2007-3925

Ing. agric. biosist. vol.14 no.2 Chapingo jul./dic. 2022  Epub 16-Feb-2024

https://doi.org/10.5154/r.inagbi.2022.02.020 

Scientific article

Parameter optimization of Green and Ampt equation using a nonlinear algorithm

Sebastián Fuentes1 

Carlos Fuentes2 

Carlos Chávez1  * 

1Universidad Autónoma de Querétaro, Centro de Investigaciones del Agua, Departamento de Ingeniería de Riego y Drenaje. Cerro de las Campas sn, col. Las Campanas, C. P. 76010, México.

2Instituto Mexicano de Tecnología del Agua, Coordinación de Riego y Drenaje. Paseo Cuauhnáhuac, núm. 8532, Jiutepec, Morelos, C. P. 62550, México.


Abstract

Introduction: The Richards equation or a simplification, such as the Green and Ampt equation, is mainly used to describe the flow of water in the soil. The Richards equation produces accurate results, but lacks general analytical solutions, so the Green and Ampt equation is used as an intermediate way of modeling the phenomenon by maintaining physical-mathematical bases in the representation.Objective: To optimize the parameters of saturated hydraulic conductivity and wetting front suction of the Green and Ampt equation by means of a nonlinear optimization algorithm, and to validate the solution to optimize the parameters as a function of soil textures.Methodology: The Levenberg-Marquardt algorithm was used to estimate the infiltration parameters of the Green and Ampt equation, and the initial pair of values was taken according to soil texture. Results: This model was used to calculate the hydraulic conductivity parameters at saturation and wetting front suction.Study limitations: A homogeneous soil column and a constant initial moisture content were considered for the whole column.Originality: The optimization algorithm was oriented to keep the parameter values within the range established in the textural classes.Conclusions: An optimization algorithm was implemented to estimate the parameters Ks and hf by solving the Green and Ampt equation, which included code performance review and experimental validation with infiltration tests.

Keywords: infiltration; saturated hydraulic conductivity; wetting front suction; Levenberg Marquardt algorithm; soil texture

Resumen

Introducción: Para describir el flujo de agua en el suelo se utiliza principalmente la ecuación de Richards o una simplificación, como la ecuación de Green y Ampt. La primera produce resultados precisos, pero carece de soluciones analíticas generales, por lo cual se utiliza la ecuación de Green y Ampt como una forma intermedia de modelar el fenómeno al mantener bases físico-matemáticas en la representación.Objetivo: Optimizar los parámetros de conductividad hidráulica a saturación y la succión en el frente de humedecimiento de la ecuación de Green y Ampt mediante un algoritmo de optimización no lineal, y validar la solución para obtener la optimización de parámetros en función de las texturas de los suelos.Metodología: Se utilizó el algoritmo de Levenberg-Marquardt para estimar los parámetros de infiltración de la ecuación de Green y Ampt, y en función de la textura del suelo se tomaron el par de valores iniciales.Resultados: El modelo obtenido permitió calcular los parámetros de conductividad hidráulica a saturación y la succión en el frente de humedecimiento.Limitaciones del estudio:Se consideró una columna homogénea de suelo y un contenido de humedad inicial constante en toda la columna.Originalidad: El algoritmo de optimización se orientó para mantener los valores de los parámetros dentro del rango establecido en las clases texturales.Conclusiones: Se implementó un algoritmo de optimización para calcular los parámetros Ks y hf mediante la solución de la ecuación de Green y Ampt, la cual incluyó la revisión del funcionamiento del código y la validación experimental con pruebas de infiltración.

Palabras clave: infiltración; conductividad hidráulica a saturación; succión en el frente de humedecimiento; algoritmo Levenberg Marquardt; textura del suelo

Introduction

Soil water flow is important in areas such as soil mechanics, irrigation, drainage, hydrology and agriculture (Fuentes et al., 2020). This phenomenon can be described physically with the Richards equation (1931), which produces accurate results if the hydrodynamic properties and boundary conditions are known, but it lacks general analytical solutions, and therefore highly complex numerical methods are usually required for its solution (Damodhara-Rao et al., 2006; Ma et al., 2010; Malek & Peters, 2011). Alternatively, there are physically based approximate models, which are the result of simplifying the initial conditions. In particular, the equation of Green and Ampt (1911) is an alternative to simulate the process of vertical infiltration of water into the soil, which has been used in surface irrigation. (Chávez et al., 2018, 2020; Naghedifar et al., 2020; Saucedo et al., 2015).

The application of the Green and Ampt equation involves soil physical property parameters (Ali & Islam, 2018; Damodhara-Rao et al., 2006); however, there are two key parameters that are not viable to measure experimentally: saturated hydraulic conductivity (Ks) and wetting front suction (hf). The value of Ks can be measured directly in the field using rainfall simulators, infiltration meters and constant or variable load permeameters (Gómez-Tagle et al., 2008), as well as Guelph permeameter. However, these methods need long periods of time to achieve a stable value, which leads to the use of large volumes of water for estimation (Castiglion et al., 2018). This uncertainty in the values of these soil hydrodynamic parameters can be considered an obstacle to optimal irrigation design due to the high degree of difficulty in calculating them, although most of the time it depends on the experience of the modeler. The parameters of the Green and Ampt equation have been optimized by least-squares fitting under rainfall conditions (Chen et al., 2015; Xiang et al., 2016). Chen et al. (2015) showed that these parameters are affected by the duration of precipitation used in their modeling. Currently, different theoretical and empirical approaches such as pedotransfer functions (Saxton & Rawls, 2006; Trejo-Alonso et al., 2020) and artificial neural networks are used to estimate these parameters (Trejo-Alonso et al., 2021). Nevertheless, pedotransfer functions are only applicable in the area where the model was calibrated, and neural networks represent a high computational cost (higher computational time).

Therefore, the objectives of this work were to optimize the parameters Ks and hf of the Green and Ampt equation by using a nonlinear optimization algorithm and to validate the solution to achieve the optimization of parameters as a function of soil textures.

Materials and methods

The Green and Ampt equation

The Green and Ampt equation is established from the continuity equation and Darcy's law (1856) with the following assumptions: a) the initial moisture profile in a soil column is uniform: θ = θo, b) water pressure at the soil surface is hydrostatic: ψ = h ≥ 0 (where h is the water depth on the soil surface), c) there is a well-defined wetting front characterized by a negative pressure: ψ = -hf () 0 (where hf is the wetting front suction) and d) the region between the soil surface and the wetting front is completely saturated (piston flow): θ = θs and K = Ks (where Ks is the saturated hydraulic conductivity; that is, the value of the hydraulic conductivity of Darcy's law corresponding to the volumetric water content at saturation). The resulting ordinary differential equation is as follows:

dI(t)dt=Ks1+h+hfθs-θ0I (1)

where I the cumulative infiltration, t is the time, and θs and θo are the initial and saturated moisture contents, respectively.

If the water depth over the surface is considered independent of time, Equation (1) is analytically integrated with the initial condition I = 0 at t = 0, which results in the following:

I=kst+λln1+Iλ, λ=h+hfθs-θ0 (2)

When the amount of water flowing at infinity is neglected, the infiltrated water depth as assumed by Green and Ampt is equal to the volume per unit area stored in the piston: I(t) = zf(t)Δθ with Δθ = θs - θo, and zf(t) is the piston front (Fuentes et al., 2012).

Parameter optimization

In nonlinear regression models, each observation yi is written as a response function f(xi; β). An important difference in nonlinear regression models is that the number of parameters β of the regression is not directly related to the number of variables xi in the model (Cornejo-Zuniga & Rebolledo-Vega, 2016). To estimate the parameters Ks and hf using an infiltration test, the Levenberg-Marquardt algorithm (Moré, 1978) is applied, which has been a standard technique for nonlinear least squares problems (Chávez et al., 2022; Fuentes, et al., 2022; Šimůnek & Hopmans, 2018). The parameter update is performed iteratively with the following expression:

p=-JTJ+μId-1JTr (3)

where J is the Jacobian matrix related to the variations of the infiltration function for each parameter to be optimized, Id is the identity matrix, r is the vector of differences between the measured infiltrated water depth and the one calculated with the optimization algorithm, and μ is the damping parameter that is fitted in each iteration. The Jacobian matrix is calculated as follows:

J=I1KsI1hfI2KsI2hfImKsImhf (4)

where Im is the equation of Green and Ampt and m is the number of measured data. In this algorithm the values of the Jacobian matrix are approximated numerically by centered derivatives to improve computational time.

The optimization of the parameters was coded in a program with functions and subroutines, which start with a pair of initial values, to subsequently calculate the Jacobian matrix and the error vector. This allows to solve Equation (3), which results in a new pair of values. This allows to calculate the error vector and check if it is smaller than the previous iteration due to a tolerance. If it is higher than the tolerance, the algorithm starts again with the pair found in the last iteration, and if this error vector is lower than the tolerance, the program stops and prints the results. The diagram is shown in Figure 1.

Figure 1. Levenberg-Marquardt flow chart for parameter optimization. 

Input parameters

The input data required are the initial and saturation moisture contents, as well as data derived from an infiltration test (time and infiltrated water depth). To determine the initial moisture content (θo), the input variables are bulk density of soil (ρt), gravimetric moisture content (ωo) and density of water (ρw):

θ0=ptpwω0 (5)

Porosity (φ) is the void space volume of the porous media, which is calculated from bulk density (ρt) and quartz particle density (ρs):

φ=1-ptps (6)

where ρs = 2.65 g∙cm-3.

Saturation moisture content (θs) is the volume of water retained in the pore space, usually assimilated to the volumetric porosity (φ) by the following inequality: 0θsφ; nevertheless, it is important to note that, for a soil apparently saturated with water, a certain amount of air is usually trapped. Therefore, the saturated water content can be taken θs = 0.9φ (Haverkamp et al., 2016; Rogowski, 1971). In this study, θs will be taken as the total porosity: θsφ.

Laboratory experiment

Soil samples were taken from agricultural plots in the region of San Juan del Río, Querétaro, Mexico, following the methodology proposed by Reynolds and Topp (2007). The soil was sieved with a number 10 mesh (2 mm) and was outdoor dried for one week. Before placing the soil in the column, a sample was taken and sent to the laboratory to determine the initial moisture content using the gravimetric method.

The infiltration test was carried out using two acrylic columns of different lengths and circular cross section. Both columns had a porous plate covered by a filter and placed at the bottom to retain the soil and allow water and air to escape (Figure 2). The inside of the columns were partially coated with wax to create roughness between the soil and the acrylic column. The soil was placed in the column in 5 cm thick layers at a density similar to that obtained in the field. A constant water depth was kept during the infiltration test.

Figure 2. Schematic representation of a soil column. 

Table 1, adapted by Saucedo et al. (2013), shows the mean values of some soil parameters (θo, θs, Ks and hf) based on soil texture, using the relationships proposed by Rawls et al. (1991). To determine these values, the soils used were analyzed at the laboratory with the Bouyoucos hydrometer using the methods of the Mexican standard NOM-021-SEMARNAT-2000 (Secretaría del Medio Ambiente y Recursos Naturales [SEMARNAT], 2002) and the texture triangle of the United States Department of Agriculture (USDA). Within the operation of the optimization algorithm, the following condition was included: if the difference of parameters in the current iteration with the previous one is lower than 0.00001 %, or the number of iterations is equal to 20, end the algorithm sequence and display the results.

Table 1. Mean parameters of the Green and Ampt infiltration equation according to soil texture. 

Textura del suelo /
Textura del suelo
Parameters / Parámetros
θo (cm3∙cm-3) θs (cm3∙cm-3) hf (cm) Ks (cm∙h-1)
Clay / Arcilla 0.36 0.49 140.26 0.05
Silty clay /
Arcillo-limoso
0.32 0.48 100.16 0.05
Silty clay loam /
Franco-arcillo-limoso
0.26 0.49 60.12 0.15
Clay loam /
Franco-arcilloso
0.25 0.48 36.00 0.4
Sandy clay /
Arcilla-arenosa
0.25 0.42 25.72 0.5
Silt /
Limo
0.14 0.50 30.52 0.8
Loam /
Franco
0.20 0.46 20.04 1.5
Silt loam /
Franco-limoso
0.17 0.55 30.07 1.0
Sandy clay loam /
Franco-arcillo-arenoso
0.18 0.42 35.61 2.0
Sandy loam /
Franco-arenoso
0.16 0.46 10.00 2.9

Results and discussion

Decoding the algorithm

The optimization algorithm was programmed using the Visual Basic.net language in the Microsoft Visual Studio 2019 integrated development environment. This program can be installed on any Windows 10 platform, and the speed in RAM required is minimal, so any computer with a mid-range processor and 4 GB of RAM supports it. The program consists of four main sections: a) main menu, b) displayed data from an infiltration test, c) graphical results of measured vs. optimized irrigation water depth and d) optimization results (Figure 3).

Figure 3. Main window of the computer program under development. 

Table 2 shows the routines used for the proper operation of the computer program we developed. The task performed by each routine is also described.

Table 2. Routines used in the computer program. 

Titel / Título Routine / Rutina Task / Acción
File / Archivo Import Excel data /
Importar datos de Excel
It opens data from an infiltration test in *.xls
format (time and cumulative infiltration). /
Abre datos de una prueba de infiltración en formato*
.xls (tiempo y lámina infiltrada acumulada).
Open examples /
Abrir ejemplos
Data from an infiltration test is loaded,
including input parameters. /
Se cargan datos de una prueba de infiltración, donde
también se incluyen sus parámetros de entrada.
Save optimized data /
Guardar datos optimizados
It saves the simulated data with the optimized parameters
in *.xls format (time and cumulative infiltration). /
Guarda los datos simulados con los parámetros optimizados
en formato *.xls (tiempo y lámina infiltrada acumulada).
Exit program /
Salir del programa
Closes the program. / Cierra el programa.
Soil data /
Datos del suelo
Input data /
Datos de entrada
It displays a window in which the user must enter the required
input parameters (bulk density, initial moisture, water
depth, infiltration test column length and soil texture). /
Muestra una ventana en la cual el usuario debe introducir
los parámetros de entrada necesarios (densidad aparente,
humedad inicial, carga superficial, longitud de columna
de la prueba de infiltración y textura del suelo).
Run /
Ejecutar
Run /
Ejecutar
It starts the parameter optimization process,
while charting the iterations in real time. /
Inicia el proceso de optimización de parámetros, al mismo
tiempo que grafica en tiempo real las iteraciones.
Calculation memory /
Memoria de cálculo
Calculation memory /
Memoria de cálculo
It displays a window with the results for each iteration
performed by the Levenberg-Marquardt algorithm. /
Muestra una ventana con los resultados en cada iteración
realizada por el algoritmo de Levenberg-Marquardt.
Help / Ayuda Help / Ayuda It opens a window displaying the authors' contact information. /
Abre una ventana donde se muestran los
datos de contacto de los autores.

Infiltration test

The initial conditions of the experiment are described in Table 3, showing the results of the analyses conducted at the laboratory, as well as the lengths of the columns (L) filled with soil, diameters (D) and surface water depths (h).

Table 3. Physical characteristics of the study sites. 

Soil sample /
Muestra de suelo
D (cm) ρt (g∙cm-3) θs (cm3∙cm-3) θo (cm3∙cm-3) L (cm) h (cm) Texture /
Textura
S1 9.2 1.0810 0.5921 0.1462 70.0000 6.0000 Sandy loam /
Franco-arenosa
S2 8.8 1.1588 0.5627 0.1259 85.0000 6.0000 Clay / Arcillosa
S3 8.8 1.1713 0.5580 0.0280 85.0000 5.0000 Sandy loam /
Franco-arenosa

With the data from the initial conditions of the experiments and the infiltrated water depth over a period of time, the parameters Ks and hf of the Green and Ampt equation were optimized according to the soil texture. The results of the optimization for each site are displayed in Figure 4, showing the fit of the Green and Ampt model to the measured values.

Figure 4. Comparison between infiltrated water depth measured and calculated with the Green and Ampt equation. 

Table 4 shows the optimization achieved, and the root mean square error (RMSE), resulting from the infiltrated water depth measured and the theoretical infiltrated water depth. Errors lower than 0.45 cm are observed, showing that the optimization of parameters with initial values related to soil texture works adequately. It is important to highlight that, from the previous graph, it is possible to observe the good fit of the optimized model with the data measured at the laboratory.

Table 4. Optimized values of the site analyzed according to the soil texture. 

Soil / Sitio Ks (cm∙h-1) hf (cm) RMSE (cm) / RECM (cm)
S1 2.4869 2.0000 0.1323
S2 0.2132 119.0250 0.3052
S3 2.8071 10.0561 0.4236

Numerical validation

The validation of the numerical algorithm was performed to remove programming errors and check the consistency of correct solutions. Data reported in the literature were used, and the results were obtained using the one-dimensional solution of the Richards equation (Chávez et al., 2016; Fuentes et al., 2020). The data used correspond to a sandy loam soil from Montecillo, Mexico (Zataráin et al., 2003). Parameter values for this soil are: ρt = 1.3607 g∙cm-3,θo = 0.1391 cm3∙cm-3, h = 1.50 cm, φ = 0.4865 cm3∙cm-3, θs = φ and L= 70 cm. The parameters Ks and hf of the Green and Ampt equation were optimized using the computational program according to soil texture (Figure 5).

Figure 5. Comparison of the infiltrated water depth derived from the Richards equation and the one modeled withthe Green and Ampt equation. 

When comparing the result and the one reported in the literature (Table 5), it is observed that there is no considerable difference, since RMSE (0.1953 cm) is considered the same, and this indicates that the result measured using the Green and Ampt model optimizing Ks and hf varies very little.

Table 5. Optimized parameter values compared with those in the literature. 

Experiment / Experimento Ks reported (cm∙h-1) /
Ks reportado (cm∙h-1)
Ks optimized (cm∙h-1) /
Ks optimizado (cm∙h-1)
Montecillo 1.84 2.28

The parameter hf is characteristic of the Green and Ampt equation; therefore, it cannot be compared with any other parameter of other equations. However, the value is adapted depending on the proportions of sand, silt and clay in soil, so it can be considered inversely proportional to Ks, because when Ks decreases, the value of hf increases.

If only soil texture is considered for the surface irrigation design, the average Ks value for the sandy loam soil analyzed would be 2.9 cm·h-1 (Table 1). However, this type of soil is dominated by sand characteristics, which leads to Ks values higher than 1.5 cm·h-1. The range of variation between both values is very wide, which causes a substantial change in the characteristic values of infiltration.

Therefore, it is recommended to perform infiltration tests in soil columns to adequately represent the flow in a one-dimensional manner, since other tests, such as the double ring infiltrometer shows a two-dimensional flow and, in addition, there is a preferential flow along the cylinder walls, which in most cases is attributed to the infiltration velocity. The advantage of using columns with soil samples altered at the laboratory allows control of all variables, such as bulk density, soil depth, water depth, initial moisture, among others.

Temporal sensitivity of optimized infiltration test parameters

A sensitivity analysis of the Green and Ampt parameters was performed to verify how the estimation errors of the two parameters propagate to the infiltration estimates. Optimizations were made with the experimental results at different times for each site. For S1 a cumulative increment of Δt = 1 h, for S2 a Δt = 4 h and for S3 a Δt = 1 h. Figure 6 shows that the parameters have the same trend, while Ks gradually decreases, the value of hf increases. The inverse process occurs with the same trend; this indicates that the algorithm optimizes the values according to the soil texture and the more infiltration data one has, the optimized parameters can vary considerably, either decreasing or increasing their values gradually and inversely proportional to each other. It also shows the importance of performing an infiltration test at long times, because, if it is performed at short times, the pores are not filled with water, and one would only have a value of hydraulic conductivity and not the saturated hydraulic conductivity.

Figure 6. Evolution of Ks and hf parameters.  

The program is delimited so that each pair of new values proposed are within the limits of the parameters, i.e., Ks )( 0 and 0 () hf ≤ 200. Parameters are adapted according to soil texture, with emphasis on the Equation (7):

S2=2KshfΔθ (7)

where S is the sorptivity of the porous medium (Philip, 1957). Without the use of initial parameters related to soil texture, there can be an infinite number of parameters to fit the data measured in an infiltration test.

In this study, a computer with the following characteristics was used: Intel® CoreTM i7-4710MQ CPU @ 2.50 GHz and 32 GB RAM memory. The computation time required to find the optimal values of Ks and hf in the experiments was 2 s. This is because the Green and Ampt solution requires no knowledge of values at the previous time level; therefore, it is possible to solve it exclusively in the required time, as in the exact times where it is compared with the measured data. On the other hand, the Richards equation needs the pressure values at the previous time level, which delays, in at least one iteration, the calculation of the infiltrated water depth.

Conclusions

The estimation of hydrodynamic parameters of saturated hydraulic conductivity and wetting front suction was performed with the Levenberg-Marquardt algorithm. The main advantage of the presented optimization is the short computational time required to optimize the parameters of the Green and Ampt equation, compared to the Richards equation reported in the literature. It was shown that, by using infiltration test data from the literature, the values derived for Ks with the proposed nonlinear optimization algorithm are very similar to those found with the Richards equation, which provides certainty and reliability when calculating the parameters.

The sensitivity analysis suggests that the duration time of the infiltration test significantly affects the Ks and hf values, so it is recommended to perform an infiltration test up to the maximum infiltration, since variations in the infiltrated water depth with time may cause the curve to be adapted to a different pair of parameters..

Acknowledgments

The first author thanks CONACyT for the grant No. 957179, and the Universidad Autónoma de Querétaro for the support provided to carry out his studies in the PhD in Engineering program.

REFERENCES

Ali, S., & Islam, A. (2018). Solution to Green–Ampt infiltration model using a two-step curve-fitting approach. Environmental Earth Sciences, 77(7), 1-9. https://doi.org/10.1007/s12665-018-7449-8 [ Links ]

Castiglion, M. G., Behrends-Kraemer, F, & Marquez-Molina, J. J. (2018). Conductividad hidráulica saturada determinada por distintos procedimientos en suelos con alta humedad inicial. Ciencia del suelo, 36(2), 158-169.https://www.researchgate.net/publication/336412222_Conductividad_hidraulica_saturada_determinada_por_distintos_procedimientos_en_suelos_con_alta_humedad_inicialLinks ]

Chávez, C., Fuentes, C., & Mota, C. (2016). Programa para modelar la transferencia de agua usando la ecuación de Richards 1D. Patente núm. INDAUTOR 03-2016-110812360600-01. [ Links ]

Chávez, C., Fuentes, S., Fuentes, C., Brambila-Paz, F., & Trejo-Alonso, J. (2022). How surface irrigation contributes to climate change resilience - a case study of practices in Mexico. Sustainability, 14(13), 7689. https://doi.org/10.3390/su14137689 [ Links ]

Chávez, C., Limón-Jiménez, I., Espinoza-Alcántara, B., López-Hernández, J. A., Bárcenas-Ferruzca, E., & Trejo-Alonso, J. (2020). Water-use efficiency and productivity improvements in surface irrigation systems. Agronomy, 10(11), 1759. https://doi.org/10.3390/agronomy10111759 [ Links ]

Chávez, C., Mota, C., Fuentes, C., & Quevedo, A. (2018). Modelación bidimensional de la infiltración del agua en surcos aplicando el gradiente conjugado. Tecnología y Ciencias del Agua, 9(1), 89-100. https://doi.org/10.24850/j-tyca-2018-01-06 [ Links ]

Chen, L., Xiang, L., Young, M. H., Yin, J., Yu, Z., & Genuchten, M. T. (2015). Optimal parameters for the Green-Ampt infiltration model under rainfall conditions. Journal of Hydrology and Hydromechanics, 63(2), 93-101. https://doi.org/10.1515/johh-2015-0012 [ Links ]

Cornejo-Zuniga, O., & Rebolledo-Vega, R. (2016). Estimación de parámetros en modelos no lineales: algoritmos y aplicaciones. Revista EIA, 13(25), 81-98. https://revistas.eia.edu.co/index.php/reveia/article/view/1019/976Links ]

Damodhara-Rao, M., Raghuwanshi, N. S., & Singh, R. (2006). Development of a physically based 1D-infiltration model for irrigated soils. Agricultural Water Management, 85(1-2), 165-174. https://doi.org/10.1016/j.agwat.2006.04.009 [ Links ]

Darcy, H. (1856). Les fontaines publiques de la ville de Dijon. Hachette Livre BNF. [ Links ]

Fuentes, C., Parlange, J. Y., & Palacios-Vélez, O. (2012). Teoría de la infiltración. In: Fuentes, C. & Rendón, L. (Eds.), Riego por gravedad (pp. 142-197). Universidad Autónoma de Querétaro. [ Links ]

Fuentes, S., Fuentes, C., Saucedo, H., & Chávez, C. (2022). Border irrigation modeling with the barré de Saint-Venant and Green and Ampt equations. Mathematics, 10(7), 1039. https://doi.org/10.3390/math10071039 [ Links ]

Fuentes, S., Trejo-Alonso, J., Quevedo, A., Fuentes, C., & Chávez, C. (2020). Modeling soil water redistribution under gravity irrigation with the Richards equation. Mathematics, 8(9), 1581. https://doi.org/10.3390/math8091581 [ Links ]

Gómez-Tagle, C. A., Gómez-Tagle, R. A. F., Batlle-Sales, J., Zepeda-Castro, H., Guevara-Santamaría, M. A., Maldonado, L. S., & Pintor A. J. E. (2008). Conductividad hidráulica saturada de campo: uso de un infiltrómetro de carga constante y anillo sencillo. Terra Latinoamericana, 26(4), 287-297.https://www.researchgate.net/publication/237036591_CONDUCTIVIDAD_HIDRAULICA_SATURADA_DE_CAMPO_USO_DE_UN_INFILTROMETRO_DE_CARGA_CONSTANTE_Y_ANILLO_SENCILLOLinks ]

Green, W. H., & Ampt, G. A. (1911). Studies on soil physics, I: The flow of air and water through soils. The Journal of Agricultural Science, 4(1), 1-24. [ Links ]

Haverkamp, R., Debionne, S., Angulo-Jaramillo, R., & de Condappa, D. (2016). Soil properties and moisture movement in the unsaturated zone. In: Cushman, J. H., & Tartakovsky, D. M. (Eds), The Handbook of Groundwater Engineering (pp. 149-190). CRC Press. https://doi. org/10.1201/9781315371801 [ Links ]

Ma, Y., Feng, S., Su, D., Gao, G., & Huo, Z. (2010). Modeling water infiltration in a large layered soil column with a modified Green-Ampt model and HYDRUS-1D. Computers and Electronics in Agriculture, 71(1), 40-47. https://doi.org/10.1016/j.compag.2009.07.006 [ Links ]

Malek, K., & Peters, R. T. (2011). Wetting pattern models for drip irrigation: New empirical model. Journal of Irrigation and Drainage Engineering, 137(8), 530-536. https://doi.org/10.1061/(ASCE)IR.1943-4774.0000320 [ Links ]

Moré, J. J. (1978). The Levenberg-Marquardt algorithm: Implementation and theory. In: Watson, G. A. (Ed), Numerical Analysis (pp. 105-116). Springer. https://doi.org/10.1007/BFb0067700 [ Links ]

Naghedifar, S. M., Ziaei, A. N., & Ansari, H. (2020). Numerical analysis and optimization of triggered furrow irrigation system. Irrigation Science, 38(3), 287-306. https://doi.org/10.1007/s00271-020-00672-5 [ Links ]

Philip, J. R. (1957). The theory of infiltration: 4. Sorptivity and algebraic infiltration equations. Soil Science, 84(3), 257-264. https://doi.org/10.1097/00010694-195709000-00010 [ Links ]

Rawls, W. J., Gish, T. J., & Brakensiek, D. L. (1991). Estimating soil water retention from soil physical properties and characteristics. In: Lal, R., & Stewart, B. A. (Eds.), Soil restoration (pp. 213-234). Springer New York. https://doi.org/10.1007/978-1-4612-3144-8_5 [ Links ]

Rendón, L., Saucedo, H., & Fuentes, C. (2012). Diseño del riego por gravedad. In: Fuentes, C. & Rendón, L. (Eds.), Riego por gravedad (pp. 324-357). Universidad Autónoma de Querétaro. [ Links ]

Reynolds, W. D., & Topp, G. C. (2007). Soil water analyses: Principles and parameters. In: Carter, M. R. & Gregorich, E. G. (Eds.), Soil sampling and methods of analysis (pp. 913-939). CRC Press. [ Links ]

Richards, L. A. (1931). Capillary conduction of liquids through porous mediums. Physics, 1(5), 318-333. https://doi.org/10.1063/1.1745010 [ Links ]

Rogowski, A. S. (1971). Watershed physics: model of the soil moisture characteristic. Water Resources Research, 7(6), 1575-1582. https://doi.org/10.1029/WR007i006p01575 [ Links ]

Saucedo, H., Zavala, M., & Fuentes, C. (2015). Diseño de riego por melgas empleando las ecuaciones de Saint-Venant, y Green y Ampt. Tecnología y Ciencias del Agua, 6(5), 103-112. [ Links ]

Saucedo, H., Zavala, M., Fuentes, C., & Castanedo, V. (2013). Gasto óptimo en riego por melgas. Tecnología y Ciencias del Agua, 4(3), 135-148. https://www.scielo.org.mx/scielo.php?script=sci_arttext&pid=S2007-24222013000300008Links ]

Saxton, K. E., & Rawls, W. J. (2006). Soil water characteristic estimates by texture and organic matter for hydrologic solutions. Soil Science Society of America Journal, 70(5), 1569-1578. https://doi.org/10.2136/sssaj2005.0117 [ Links ]

Secretaría de Medio Ambiente y Recursos Naturales (SEMARNAT). (2002). Norma Oficial Mexicana NOM-021-RECNAT-2000, que establece las especificaciones de fertilidad, salinidad y clasificación de suelos. Estudios, muestreo y análisis. SEMARNAT. https://biblioteca.semarnat.gob.mx/janium/Documentos/Ciga/libros2009/DO2280n.pdfLinks ]

Šimůnek, J., & Hopmans, J. W. (2018). Parameter optimization and nonlinear fitting. In: Dane, J. H. & Topp, G. C. (Eds.), Methods of soil analysis, Part 1, Physical methods (pp. 139-157). Soil Science Society of America. https://doi.org/10.2136/sssabookser5.4.c7 [ Links ]

Trejo-Alonso, J., Fuentes, C., Chávez, C., Quevedo, A., Gutierrez-Lopez, A., & González-Correa, B. (2021). Saturated hydraulic conductivity estimation using artificial neural networks. Water, 13(5), 1-15. https://doi.org/10.3390/w13050705 [ Links ]

Trejo-Alonso, J., Quevedo, A., Fuentes, C., & Chávez, C. (2020). Evaluation and development of pedotransfer functions for predicting saturated hydraulic conductivity for mexican soils. Agronomy, 10(10), 1516. https://doi.org/10.3390/ agronomy10101516 [ Links ]

Xiang, L., Ling, W., Zhu, Y., Chen, L., & Yu, Z. (2016). Self-adaptive Green-Ampt infiltration parameters obtained from measured moisture processes. Water Science and Engineering, 9(3), 256-264. https://doi.org/10.1016/j.wse.2016.05.001 [ Links ]

Zataráin, F., Fuentes, C., Rendón, L., & Vauclin, M. (2003). Propiedades hidrodinámicas efectivas del suelo en el riego por melgas. Ingeniería Hidráulica en México, 18(3), 5-15.http://revistatyca.org.mx/ojs/index.php/tyca/article/view/945/pdf_1Links ]

Received: February 12, 2022; Accepted: August 11, 2022

Corresponding author: chagcarlos@uaq.mx, tel. 442 192 12 00 ext. 6036.

Creative Commons License This is an open-access article distributed under the terms of the Creative Commons Attribution License/