UC Davis Study Finds Nitrogen Data Choice Can Redirect California Groundwater Protection Priorities
14 November 2026, Davis, US: A new study from the University of California, Davis has found that the data source used to estimate how much nitrogen fertilizer is left unused in farm soil can swing the results widely enough to change which crops and regions get prioritized for groundwater protection in California’s Central Valley.
Nitrogen surplus is the gap between the nitrogen applied to a field through fertilizer, manure and other inputs and the nitrogen removed when the crop is harvested. The leftover amount can leach into groundwater as nitrate, and California has been tightening regulation of farm nitrogen surplus in recent years to protect drinking-water wells. In parts of the Central Valley, 40 percent to 50 percent of private domestic wells exceed the US Environmental Protection Agency’s nitrate limit of 10 milligrams per liter, according to state and federal data cited in the study.
The research, published November 10 in the journal Nature Sustainability, was led by Thomas Harter, a Cooperative Extension specialist and professor in UC Davis’s Department of Land, Air and Water Resources, with postdoctoral researcher Baibaswata Bhaduri as lead author. Co-authors were Giorgos Kourakos of UC Davis and Kenneth Miller of Formation Environmental LLC. The California State Water Resources Control Board and the US Department of Agriculture’s National Institute of Food and Agriculture funded the work.
How the numbers diverge
The team used California’s Central Valley, which the study notes covers less than 1 percent of US land but produces more than 250 different crops, supplies over a quarter of the country’s food and accounts for roughly 18.5 percent of national milk production, as what the researchers called a living laboratory. Most nitrogen surplus assessments elsewhere in the country rely on just one or two data sources, typically the US Department of Agriculture’s agricultural survey and census. California’s Central Valley is unusual in having several additional, more localized sources to compare against.
The researchers identified four main data sources used in practice: the USDA survey and census; grower-reported nitrogen use, which farms in sensitive areas have filed annually since 2019 under the state’s own tracking requirements; UC Davis Cost and Return Studies, which are expert-estimated, crop-specific fertilizer rates and representative yields rather than site-specific measurements; and county-level crop acreage and harvest records kept by the California Department of Food and Agriculture since the early 1900s. Each source has its own weaknesses. Grower self-reports can carry recall bias and inconsistent interpretation, the Cost and Return Studies can rely on outdated survey data and skew toward high-input farming systems, and the USDA data is infrequent and spatially coarse, with limited detail for specialty crops.
To measure how much these differences matter, the team built a spatially distributed nitrogen mass-balance model running at 50-meter resolution across 17 crop categories covering around 20 of the Central Valley’s dominant crops, including almonds, walnuts, pistachios, peaches, nectarines, cherries, citrus, wine and table grapes, wheat, cotton, corn, oats, dry beans, processing tomatoes, onions, garlic and carrots. Together these crops account for about 75 percent of the valley’s irrigated cropland and more than 90 percent of its estimated potential nitrogen loading. Rice was excluded because flooded rice fields lose little nitrogen to leaching through a process called denitrification.
The model ran 24 scenarios, each pairing a high or low estimate of fertilizer applied with a high or low estimate of nitrogen removed at harvest, drawn from different combinations of the four data sources, with separate variations for whether perennial crops’ nitrogen storage in woody tissue was counted and for how much manure nitrogen was assumed. The researchers describe the 24 cases as a bounded range of plausible outcomes rather than 24 equally likely futures. Even so, the spread was large: for some crops, the researchers found the standard deviation across the estimates was more than twice the average surplus value itself, meaning the uncertainty was bigger than the number it was supposed to describe. Corn and other field crops that commonly receive manure, such as cotton, wheat and oats, showed the widest divergence, while orchard and vineyard crops such as grapes and cherries tended to produce narrower, lower estimates.
Harter said the findings support the approach California’s nitrogen policy is already taking by requiring grower self-reporting, but he cautioned regulators to “be cautious” about leaning on any single dataset, since combining different sources can reshuffle which areas get prioritized for conservation spending. Bhaduri said the disagreement between datasets “reveals where our knowledge is weakest,” adding that resolving it matters for groundwater decisions that are fair to the farmers and communities affected.
The study argues that credible nitrogen assessment anywhere requires explicitly accounting for this kind of data-source uncertainty rather than treating a single estimate as fact, and notes that the Central Valley’s problem is not unique to California: it is simply one of the few farming regions with enough overlapping datasets to reveal a gap that single-dataset assessments elsewhere carry unseen.
Also Read: Crystal Crop Protection to Launch Rs 455 Crore IPO on November 18
Global Agriculture is an independent international media platform covering agri-business, policy, technology, and sustainability. For editorial collaborations, thought leadership, and strategic communications, write to pr@global-agriculture.com
