Nitrogen Leaching and Groundwater Recharge Modeling for the On-Farm Flood Flow Capture Project in Fresno, California

Groundwater management is an important subject in California where people depend on groundwater for irrigation and drinking water. On-Farm Flood Flow Capture (OFFC) is a new method of groundwater recharge, which diverts flood flow in a river or canal to agricultural lands. OFFC's feasibility is currently investigated through an on-going project at Terranova Ranch in Helm near Fresno, CA. One of the concerns in OFFC is the groundwater pollution by nitrate (NO 3-) leaching from agricultural soils. This study is to develop a numerical model to investigate the impacts of OFFC to the NO 3- concentration and flux in the soil and groundwater recharge with different OFFC application rates, irrigation rates, and soil parameters. Also this study aims to find if NO3- can be captured by immobile pores, which reduces NO3- leaching. .;Field infiltration tests using double-ring infiltrometers were conducted to assess if immobile pores should be considered in the model. Based on the infiltration test results, a numerical model for 60 m vadose zone, which estimates groundwater recharge and NO3- leaching as its bottom water and NO3- fluxes, was developed using HYDRUS-1D considering OFFC, irrigation, precipitation, plant water uptake, plant solute uptake, evaporation and transpiration from vineyards. The model inputs were developed using surface flow and climate data from 1983 to 2002. The model was calibrated with volumetric water contents and water inflow from OFFC experiment data from 2011. The model base scenario (BASE) assumed OFFC application at 0.060 m/day when surface water was available during the non-growing season from late-October to mid-March and was only used for in-lieu recharge during the growing season from mid-March to late-October. Scenarios of an over irrigation (OI) with 1.2 times more OFFC rate from BASE and lower OFFC rate of 0.048 m/day (LOW), with 2 different longitudinal dispersivity coefficients of 1 and 0.1 m (DL1 and DL0.1) were also modeled. The model calibration data was only available for the surface 1.22 m. Therefore, the deep vadose zone soil parameters, a higher and a lower residual volumetric water contents (HthetaR and LthetaR), saturated volumetric water contents (HthetaS and LthetaS), and saturated hydraulic conductivities (HK and LK) were also used in BASE to assess their impacts to the groundwater recharge and NO3 - leaching.;The calibrated model showed low root mean square errors of less than 0.02 and high Nash-Sutcliff coefficients of above 0.65 for volumetric water contents measured at 0.15, 0.61 and 1.22 m from the soil surface. The field infiltration leached most NO3- in the 0.20m-thick soil, which cannot be explained by reasonable NO3- encapsulation in immobile regions. Therefore the model was developed with a simple equilibrium transport model without consideration of immobile pores. However, data collected from the relatively shallow depth (0.20 m) to determine the transport model type may not be able to capture preferential flow, and further investigation may be needed.;How quickly groundwater responds to surface water input varied by the quantity applied at the surface in the modeling results. However, in wet years when more than 4 m of water is applied at the surface, the groundwater recharge increased within 1 year in all tested scenarios. The modeling results showed small differences in bottom water and NO3- leaching between BASE and OI, while LOW showed lower and delayed recharge and NO 3- leaching. The root zone NO3- concentrations decreased by more than 90% at the end of the modeling period for all scenarios. NO3- concentrations at the vadose zone-groundwater interface became lower than the groundwater NO3 - concentration after 15 m of groundwater recharge in most DL0.1 scenarios, and this indicates groundwater quality improvement. However, it was not observed in any DL1 scenarios. BASE with different vadose zone parameters did not show significant differences in the recharge and timings, but showed differences in the NO3- leaching quantities and timings. HthetaR, HthetaS and LK delayed the NO3- leaching while LthetaR, LthetaS and HK accelerated NO3- leaching. Results from all scenarios and parameters showed at least 70% of initial residual soil NO3- leached to the groundwater after the modeling period.;OFFC seems to increase NO3- loads to the groundwater when applied on a soil with high residual NO3-, while it removes NO3- accumulated near the surface soil. BASE and OI showed only small differences in NO3- leaching, so irrigation rate differences may have small impacts to NO3- leaching compared to OFFC rates. The models showed that OFFC is likely to leach a high portion of the initial soil NO 3- for the study site, and NO3- reached to the groundwater from annual fertilizer application is much smaller than that from the initial soil NO3-. Therefore, initial soil NO3- concentrations and OFFC rates have more impacts to groundwater NO3- contamination than irrigation and NO3- fertilizer management for the scenarios investigated in this study.