Data and model code for "Mapping the potential for denitrification during infiltration with machine learning informed by field and laboratory experiments"

Managed aquifer recharge (MAR) can increase groundwater storage and, under some conditions, improve groundwater quality simultaneously. However, the conditions under which water quality improvements can be achieved during MAR have not been systematically examined in a spatially explicit manner and remain weakly defined. We synthesize observations from laboratory tests, field experiments and operational MAR facilities at four different MAR locations within the Pajaro Valley, in central coastal California, USA to develop a predictive model of nitrate removal during infiltration. We combine data on soil and fluid properties with conditions during infiltration and compare three modeling approaches: multiple linear regression, random forests, and boosted regression trees. The preferred model (R2 = 0.54) uses boosted regression trees based on four predictor variables: total soil carbon, soil clay fraction, soil residence time, and initial nitrate concentration. We apply this model to simulate the spatial distribution of potential nitrate removal (NRp) across a heterogeneous and mixed-use landscape. We find that areas of high NRp (≥ 1.0 mg/L) are more common on floodplains and riparian areas, and urban areas tend to have higher NRp than do forested or agriculturally developed areas. We integrate these results with existing datasets at multiple scales that are relevant to increasing groundwater supply to demonstrate the potential benefit from considering where simultaneous groundwater quality and supply improvements may be achieved. Combining a map of NRp in the Pajaro Valley with independently simulated hillslope runoff, we find that potential load reduction is highest in urban areas (median = 18.2 kg-N/yr) where large runoff volumes are collocated with soils of high nutrient cycling capacity compared to forested and agricultural areas (median = 1.6 and 3.5 kg-N/yr, respectively). We also apply the NRp model to agricultural areas across California to illustrate the utility and flexibility of this approach and find that across California, areas that are classified as moderately good to excellent for recharge are collocated with > 7,000 km2 where NRp ≥ 1.0 mg/L and > 400 km2 where NRp ≥ 2.0 mg/L. We compared modeled results to previous analyses of recharge suitability focused on water quantity, which could help guide decisions in resource management and identify promising MAR sites. NRp simulations also suggest where additional experimental data might be especially useful, show where soil augmentation to improve nutrient cycling could be efficacious and help to understand heterogeneity in biogeochemical cycling across landscapes. NOTE: The data and code withing this resource will be updated upon final publication.