Irrigation management has evolved into a top priority issue since available fresh water resources... more Irrigation management has evolved into a top priority issue since available fresh water resources are limited. Water production functions (WPFs), mathematical relationships between applied water and crop yield, are useful tools for irrigation management and economic analysis of yield reduction due to deficit irrigation. This study aimed at (i) designing and evaluating site-specific WPFs (using k nearest neighbors (k-NN), multiple linear regression, and neural networks), (ii) simulating yield maps for uniform, sector control VRI, and zone control VRI center pivot systems using the site-specific WPFs, (iii) using the best WPF to investigate different cotton irrigation and zoning strategies using integer linear programming, and (iv) comparing soil-based and WPF-based zones for sector control VRI systems. A two-year cotton irrigation experiment (2013–2014) was implemented to study irrigation-cotton lint yield relationship across different soil types. The site-specific k-NN WPFs showed the highest performance with root mean square error equal to 0.131 Mg ha-1 and 0.194 Mg ha-1 in 2013 and 2014, respectively. The result indicated that variable rate irrigation with limited sector control capability could enhance cotton lint yield under supplemental irrigation when field-level spatial soil heterogeneity is significant. The temporal changes in climate and rainfall patterns, however, had a great impact on cotton response to irrigation in west Tennessee, a moderately humid region with short season environment. We believe site specific WPFs are useful empirical tools for on-farm irrigation research.
s u m m a r y A detailed understanding of soil hydraulic properties, particularly soil available ... more s u m m a r y A detailed understanding of soil hydraulic properties, particularly soil available water content (AWC) within the effective root zone, is needed to optimally schedule irrigation in fields with substantial spatial heterogeneity. However, it is difficult and time consuming to directly measure soil hydraulic properties. Therefore, easily collected and measured soil properties, such as soil texture and/or bulk density, that are well correlated with hydraulic properties are used as proxies to develop pedotransfer functions (PTF). In this study, multiple modeling scenarios were developed and evaluated to indirectly predict high resolution AWC maps within the effective root zone. The modeling techniques included kriging, co-kriging, regression kriging, artificial neural networks (NN) and geographically weighted regression (GWR). The efficiency of soil apparent electrical conductivity (EC a ) as proximal data in the modeling process was assessed. There was a good agreement (root mean square error (RMSE) = 0.052 cm 3 cm À3 and r = 0.88) between observed and point prediction of water contents using pseudo continuous PTFs. We found that both GWR (mean RMSE = 0.062 cm 3 cm À3 ) and regression kriging (mean RMSE = 0.063 cm 3 cm À3 ) produced the best water content maps with these accuracies improved up to 19% when EC a was used as an ancillary soil attribute in the interpolation process. The maps indicated fourfold differences in AWC between coarse-and fine-textured soils across the study site. This provided a template for future investigations for evaluating the efficiency of variable rate irrigation management scenarios in accounting for the spatial heterogeneity of soil hydraulic attributes.
This study aimed at investigating the performance of multiple irrigation zoning scenarios on a 73... more This study aimed at investigating the performance of multiple irrigation zoning scenarios on a 73 ha irrigated field located in west Tennessee along the Mississippi river. Different clustering methods, including k-means, ISODATA and Gaussian Mixture, were selected. In addition, a new zoning method, based on integer linear programming, was designed and evaluated for center pivot irrigation systems with limited speed control capability. The soil available water content was used as the main attribute for zoning while soil apparent electrical conductivity (ECa), space-borne satellite images and yield data were required as ancillary data. A good agreement was observed among delineated zones by different clustering methods. The new zoning method explained up to 40% of available water content variance underneath center pivot irrigation systems. The ECa achieved the highest Kappa coefficient (=0.79) among ancillary attributes, hence exhibited a considerable potential for irrigation zoning.
Knowing the soil water retention curve (WRC) is essential for analyzing soil hydraulic behavior w... more Knowing the soil water retention curve (WRC) is essential for analyzing soil hydraulic behavior within the vadose zone. The van Genuchten (VG) soil hydraulic equation is one of the most frequently adopted models to parameterize the WRC. Some measured water retention points are needed to fit the VG model, but direct measurement of water content versus matric potential is expensive and time consuming. A pedotransfer function (PTF) enables indirect determination of a WRC from basic soil information. The typical method employed to derive a PTF using the VG model (VG-PTF) is to establish a mathematical relationship between the parameters of the VG model and basic soil data. However, both establishing and reusing a VG-PTF for new soils are challenging due to several reasons, such as over-parameterization, low correlation between basic soil data and VG parameters, and interdependency among parameters. In this study, a nonparametric approach based on the k nearest neighbor technique was designed and tested to establish a VG-PTF. A subset of soils from the UNSODA database (n = 554)) and a data set from Belgium (n = 69) were used as the model development and validation data sets, respectively. The proposed PTF showed reasonable accuracy and reliability and was comparable to well-known parametric VG-PTFs available in the literature.
The pedotransfer function (PTF) concept has been widely used in recent years as an indirect way t... more The pedotransfer function (PTF) concept has been widely used in recent years as an indirect way to predict soil hydraulic properties, particularly the water retention curve (WRC). The pseudo continuous (PC) approach allows us to predict water content at any predefined matric head, resulting in an almost continuous WRC. When combined with powerful pattern recognition approaches, a PC-PTF can be trained to learn the shape of WRC from a discrete set of measured points, unlike traditional parametric PTFs which follow a predefined WRC shape dictated by the selected soil hydraulic equations. The purpose of this study was to investigate the impact of two elements on the performance of a PC-PTF: (i) data mining method (neural network, NN, versus support vector machine, SVM) and (ii) distribution and density of the provided water retention data in the training phase. Two datasets from Turkey and Belgium, consisting of mainly fine and coarse-textured soils, respectively, were employed. Multiple scenarios containing different combinations of measured water retention points in the training phase were defined. The lower root mean square error (RMSE) on average (0.044 cm 3 cm −3 ) attained with the NN-based PTF shows that it is a better option than SVM (RMSE of 0.052 cm 3 cm −3 ) for deriving PC-PTFs. The accuracy of PC-PTF was firmly dependent on the presence of measured water retention points in the entire range of WRC. Applying different scenarios revealed that a well distributed set of measured water retention points in the training phase could result in up to 0.03 cm 3 cm −3 reduction in RMSE values.
Production functions (PFs) are practical tools for not only irrigation scheduling but also in eco... more Production functions (PFs) are practical tools for not only irrigation scheduling but also in economic analysis as a mathematical relationship between relative grain yield and factors like evapotranspiration, irrigation water and salinity. This study was carried out in the Mashhad region of Iran during cropping years 2010 and 2011 to evaluate the performances of two data mining methods, decision tree and neural network, for deriving PFs of spring wheat under simultaneous drought and salinity stress compared with four well known regression-based PFs. The four well known PFs were: Jensen-PF (Jensen, 1968), Minhas-PF , modified Stewart-PF , and Nairizi-PF . Heading and flowering were the most sensitive growth stages followed by the stem elongation and booting. Salinity stress also affected grain yield and therefore was an important parameter for deriving PFs. In general, all the PFs were in agreement concerning the sensitivity of spring wheat to water stress. The neural network-based PF performed the best with a root mean square error equal to 44.27 g m À2 while the decision tree-based PF ranked fourth out of six in terms of accuracy. The most important advantage of the neural network-based PF was the flexible number of input parameters.
In this study, a new approach, which we called pseudo-continuous, to develop pedotransfer functio... more In this study, a new approach, which we called pseudo-continuous, to develop pedotransfer functions (PTFs) for predicting soil-water retention with an artificial neural network (ANN) was introduced and tested. It was compared with ANN PTFs developed using traditional point and parametric approaches. The pseudo-continuous approach has a continuous performance, i.e. it enables to predict water content at any desirable matric potential, but without the need to use a specific equation, such as the one by van Genuchten. Matric potential is considered as an input parameter, which enables to increase the number of samples in the training dataset with a factor equal to the number of matric potentials used to determine the water retention curve of the soil samples in the dataset. Generally, the pseudo-continuous functions performed slightly better than the point and parametric functions. The root mean square error (RMSE) of the pseudo-continuous functions when considering local data for training and testing, and with both bulk density and organic matter as extra input variables on top of sand, silt and clay content, was 0.027 m 3 m À3 compared to 0.029 m 3 m À3 for both the point and parametric PTF. The increased number of samples in the training phase and the selection of matric potential as an input variable enabling to predict water content at any desired matric potential are the most important reasons why pseudo-continuous functions would need more intention in the future. Uniformity in the training and test dataset was shown to be important in deriving PTFs. We finally recommend the use of pseudo-continuous PTFs for further improvement and development of PTFs, in particular when datasets are limited.
Uploads
Papers by Amir Haghverdi