Excessive nutrients have penetrated water bodies worldwide yet our forecasting capabilities remain elusive. This work tests the shallow and deep hypothesis: chemical contrasts in the subsurface shape nitrate export patterns. We use data synthesis for 228 U.S. watersheds and reactive transport modeling (500 simulations) spanning broad climate, geology, and land conditions. Data synthesis showed that human perturbation has amplified chemical contrasts in shallow versus deep waters, inducing primarily flushing patterns (concentration increase with streamflow) in agriculture lands and dilution (concentration decrease with streamflow) patterns in urban watersheds. Data and model reveal a general relationship between export patterns and shallow versus deep nitrate contrasts. This underscores the often-overlooked role of N distribution over depth. The results challenge commonly-held perception that legacy stores in agricultural lands induce chemostasis with negligible concentration variations with discharge. They suggest nutrient export will exacerbate as extreme events such as flooding intensify in the future climate.