Nitrogen Leaching From California's Central Valley Irrigated Lands: A Multi-scale Modeling Study

The complex linkage between agricultural activities and groundwater quality is of worldwide concern, where 40% of agricultural production occurs in irrigated settings. Recent and new institutional policies are beginning to regulate the discharge of nitrate from agricultural lands to groundwater in the US, Canada and the EU. Growers are facing new regulatory pressures not only to adopt advanced irrigation and nutrient management practices, but also to develop monitoring and assessment systems to demonstrate their performance with respect to the field nitrogen budget.

Nitrogen leaching under cyclic irrigation and fertilization practices can be represented by a mass balance approach. However, water content and solute distribution in soil profiles under high efficiency irrigation and fertilization practices can be heterogeneous with small changes in management having an effect on solute leaching and their representation requires more complex numerical models based on Richards' equation, such as HYDRUS-(2D/3D).

The main objective of this study was to develop an integrated modelling framework for assessing enhanced irrigation and fertilization practices by informing and evaluating the large-scale SWAT model through site specific simulations of the transient water flow and reactive nitrogen transport model HYDRUS-(2D/3D). The comparison between the HYDRUS and SWAT models allowed for an evaluation of the physical and chemical interactions, fluxes, fate and transport of nitrogen in the agriculture-groundwater nexus.

Results from the SWAT model, calibrated by the Management Practices and Evaluation Program (MPEP) for the 40 thousand Km² regional California's Central Valley, were compared to transient water flow and nitrogen transport simulations for almond, citrus, tomato and corn crops in varying soils. Uncertainty in HYDRUS predictions resulting from uncertainty in soil hydraulic parameters was accounted for in this study. Modelled fluxes were generally found to be in the same range for both models, if accounting for specific uncertainty. Simultaneous evaluation of SWAT and HYDRUS helped to assess the range of modeled N system output (leaching) using different modeling methods.