Title: Evaluation of a Landscape-Scale Approach to Cotton Modeling Authors
Submitted to: Agronomy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: July 25, 2014
Publication Date: N/A
Interpretive Summary: Agriculture in the Texas High Plains is in a state of transition where irrigation water from the Ogallala aquifer continues to decline and the incidence of drought continues to increase. These two factors combined with increased costs of energy have contributed to develop management strategies to determine what to do with the remaining available water. For example, on a ¼-mile center pivot, is it better to concentrate the available water on the most productive land or is it preferable to irrigate the entire area? The answer to this question is not simple and producers need tools that would provide them with information to support them on their decision-making process. In this study, we used a concept that was introduced in the 1960’s and that is the use of simulation models to quantify crop growth over time and calculate how crop yield is affected by the environment. To do so we used a state-of-the art landscape scale model called PALMS for Precision Agricultural-Landscape Modeling System, which was combined with the cotton model Cotton2k to produce PALMScot. This integrated model essentially calculates al variables associated with carbon, water, energy and nitrogen on a three-dimensional scale across the landscape, i.e., a ¼-mile center pivot irrigated field. For this purpose, the landscape is divided into cells (30 × 30 feet) and to a soil depth of 3 feet or deeper, and for each cell, the model calculates cotton lint yield as affected by inputs of water and energy. In this study, we have made our first attempt to verify if the values calculated by the PALMScot model are correct. This process starts by first verifying if calculated values of soil water content are correct. For this purpose, we did experiments on cotton fields from Lamesa and Bushland, TX, which represent the range from sandy soils in the south (Amarillo soil series) to clayey soils in the north (Pullman soil series). At both sites, cotton was grown at different levels of irrigation and we compared calculated and measured values of soil water content across the landscape and over the growing season. Our results showed that the PALMScot model correctly calculates soil water content at both soil series. This is the initial verification of the model and the work continues as further verification of the cotton lint yield needs to be evaluated and this is the subject of future work. The overall objective is to have a working cotton model that can be used by producers to evaluate what management decisions would lead to more lint yield per unit of irrigation water and nitrogen applied, which are the most important determinants of lint yield. Of course, these decisions have to be supported by economics and net returns to the producers. This work represents that first step towards this objective.
Technical Abstract: The Southern Texas High Plains (SHP) is an important semiarid region for cotton (Gossypium hirsutum L.) production that depends on irrigation from the Ogallala Aquifer. The aquifer is recharged at rates less than depletion, leading to regulations and novel management strategies for irrigation that may be explored through modeling, if the model represents and captures the spatiotemporal variability associated with irrigated cotton production. The Precision Agricultural-Landscape Modeling System (PALMS) is a grid-based model that accounts for variability across complex cropping fields, but lacked a cotton growth model. Our objective was to integrate the cotton growth model Cotton2K with PALMS to produce PALMScot, and compare calculated values of soil water storage obtained with PALMScot to field measured values for two years and two irrigation treatments at two locations with contrasting soil series, an Amarillo fine sandy loam near Lamesa, TX and a Pullman clay loam near Bushland, TX. Model effectiveness was assessed using the modified coefficient of efficiency; and these values were 0.2 for both years in the Amarillo soil and 0.4 in the Pullman soil. Within the 1.4-m profile of both soil series, however, the model tended to overestimate root water uptake in the surface horizon and underestimate it in deeper horizons. Our initial evaluation implies that PALMScot is a tool that can be used to track the spatial and temporal variability of soil water storage across the landscape while minimizing field-installed sensors. Further evaluation of the cotton growth component of PALMScot at production scale is an ongoing effort.