Analyzing Full and Deficit Irrigation Systems for Industrial Crops Using the WINDS Model and Remote Sensing Technology

Abstract

Currently, the agriculture sector accounts for about 70% of the global water withdrawals and with increasing urbanization and climate change, there will be a much higher strain on the water resources in the future. If the situation remains unaltered, reports from Food and Agriculture Organization (FAO) predict that by 2050, there could be a 50% increase in food demand and a 40% decrease in the water supply. Adopting water conservation alternatives will help sustain agriculture and the population. To do so, accurate computational tools, such as soil water balance and crop growth models, are needed to provide guidance in irrigation management decisions. The WINDS (Water-use, Irrigation, Nitrogen, Drainage, and Salinity) model is a soil water balance simulation model that divides the soil profile into layers and calculates crop evapotranspiration (ETc) with the FAO56 dual crop coefficient method. In this research, the assumption was that the WINDS model can accurately simulate the ETc and soil moisture contents in soil layers. A foundational purpose of the research was to provide a comprehensive document describing the theory and algorithms used in WINDS. Data from a carefully monitored irrigation experiment conducted in 2007 in Maricopa, Arizona, were then used to calibrate and validate the WINDS model for cotton. Further, the project used the WINDS model to evaluate a guar deficit irrigation experiment conducted in 2018 and 2020 in Clovis, NM. The WINDS model is a daily time-step model that uses tipping bucket algorithms during infiltration events and the Richards’ equation between infiltration events. If the hypothesis is true, then it has a unique ability to accurately model soil moisture content in layers. Input data includes weather data used in FAO-56 dual crop coefficient calculations, soil characteristics, and crop evapotranspiration parameters. In this study, it was validated by comparing simulations with neutron probe water content readings in soil layers collected during the growing season. The WINDS model can simulate surface and subsurface irrigation methods on multiple soil types. One of the features of the model is the ability to assess input data and fill gaps for an infrequent dataset. This research effort was part of the Sustainable Bioeconomy for Arid Regions (SBAR) project. The SBAR project aims to evaluate and improve bioeconomic system for two industrial crops using models, field experiments, advanced computational tools, and outreach. From the project, work on the guar crop for one of the sites has been done. Many modifications were made to the WINDS model to simulate the range of experiments, irrigation systems, and soils in the SBAR experimental portfolio. For the WINDS model validation, an experiment done by USDA has also been modeled. The 2007 cotton experiment included two crop coefficient methods. One calculated the basal crop coefficient (Kcb) using a standard cotton crop coefficient curve from FAO56 (FT treatment), and the other estimated Kcb from observed Normalized Differential Vegetation Index (NDVI) values (NT treatment). There were four replicates in each treatment, all of which had soil texture and moisture retention characteristics analyzed in the laboratory. Extensive soil moisture content data by neutron probe were collected along with multiple NDVI readings. The field had varying soil types within the soil profile, ranging from clay loam to sandy loam. The calibrated WINDS model, using the laboratory-measured soil parameters, accurately simulated soil moisture content in layers during the growing season in all replicates for both the FT and NT treatments. This study helped confirm the model strength toward data assessment and soil moisture simulation. The next step was to use the WINDS model to augment and enhance the dataset from the guar deficit irrigation experiment in Clovis, NM. The goal was to provide an input dataset for the AquaCrop crop growth model. During 2018 and 2020, the guar field experiments had 3 factors in a randomized block design. The first factor included two pre-irrigation conditions, where the half field was irrigated before sowing and the other half was not. The second factor included four irrigation treatments: full irrigation (F), stress at the vegetative stage (Vst), stress at the reproductive stage (Rst), and rainfed (R). Each treatment had four replicates. The third factor included two cultivars, Kinman and Monument, but this study only evaluated the Kinman treatments. In this study, the second replicate (R2) for the 2018 Kinman, pre-irrigation, full seasonal irrigation treatment, had more extensive neutron probe measurements than other plots. Thus, it was evaluated with the WINDS model to parameterize the soil, assess the partition between rainfall and runoff, and detect any possible irregularities in neutron probe measurements. In the second phase, the other irrigation treatments in 2018 and 2020 were simulated with the WINDS model and compared with neutron probe measurements. In general, the simulated soil moisture with the WINDS model showed agreement with neutron probe measurements. Statistically, both experiments showed that the WINDS model provides accurate simulations of moisture in soil layers in irrigation experiments. This indicates the potential of the WINDS model to be used in parallel with a limited number of in-situ soil measurement devices and remote sensing to simulate moisture content in the soil profile and across fields.