APPLICABILITY OF THE NLOS MODEL FOR PREDICTIONS OF SOIL WATER MOVEMENT AND NITROGEN TRANSPORT IN AN AGRICULTURAL SOIL, AGASSIZ, BC

The Abbotsford-Sumas aquifer is a shallow, unconfined aquifer in northern Whatcom County, WA and southern British Columbia, Canada that is contaminated with nitrates due to agricultural land use. To mitigate the problem, US and Canadian government agencies are considering the NLEAP on STELLA (NLOS) leaching model as an additional tool for assessing nutrient management strategies. NLOS is an adaptation of the Nitrogen Leaching and Economic Analysis Package (NLEAP) model. We examined the applicability of the model by calibrating it to an agricultural field plot in southern British Columbia.

NLOS incorporates fertilizer application events, climatic data, and soil properties, to simulate one-dimensional water flow and nitrogen fate and transport. Field data from a trial of silage corn located at the Pacific Agri-Foods Research Centre in Agassiz, BC was used to calibrate the model. Sampling included soil, soil pore water, nitrous oxide emissions, and groundwater chemistry parameters. The field soil (a silt loam) was subjected to a nutrient loading and crop management scenario comparable to regional farming practices. Although the model was calibrated in Agassiz, BC, we expect that NLOS will perform similarly in the Abbotsford-Sumas aquifer due to similar soil types and climatic conditions.

The ability of NLOS to predict water and nitrate transport during seasonal precipitation events was examined by comparing simulations to monthly field data. NLOS was found to be useful for predictions of nitrate leaching and soil nitrate and ammonia in the upper 12 inches of the soil profile, and soil water from 12 to 24 inches deep. Model predictions accounted for 72% of the observed variability in nitrate leached in the upper 24 inches and 67% from 24 to 36 inches deep. Simulated soil nitrate and ammonia in the upper 12 inches of the soil profile accounted for 85% and 87%, respectively, of the variability in the observed values. Simulated soil water from 12 to 24 inches deep accounted for 73% of the variability in the observed values. Field observations and model simulations indicate that nutrients in the soil and soil pore water fluctuated in direct response to fertilizer applications, crop events, and precipitation. Although the model performed reasonably well, more frequent field data collection is recommended for further model calibration and validation.