『Abstract
　In situ denitrification rates were measured in a shallow unconfined glaciofluvial aquifer that had undergone large-scale nitrate contamination. Denitrification rates and isotopic enrichment factors, ε, were measured using three tracer tests in two aquifers in situ mesocosms (ISMs). Denitrification rates were also measured using a mass balance method using water samples from multiport samplers. First-order kinetic rates (k) best described the denitrification rates measured. ISM kinetic rates ranged from 0.00049/d to 0.0031/d and ε values ranged from -4.86‰ to -9.34‰; a linear relationship between k and ε values showed greater fractionation (more negative ε values) associated with higher rates. For the mass balance method, k values ranged from 0.0028/d to 0.0041/d. Combined mineralogical analysis, water quality data from the ISMs, and geochemical models using PHREEQC indicated that contributions of major electron donors to denitrification were 43-92% by organic carbon, 4-18% by pyrite, and 2-43% by non-pyritic ferrous iron, depending on the sample date and the type of amphibole used as the electron donor for ferrous iron. ISMs show promise as a tool for hydrogeochemical investigations. They are large enough to allow long-term sampling of aquifer denitrification tracer tests (＞2 years), they may be used, with the modeling methodology shown herein, to estimate relative e^- donor contributions, and they limit the influence of advection and mechanical dispersion on the amended water within the chamber.

Keywords: Aquifer denitrification; In situ mesocosms; Denitrification rates; Electron donor contributions to denitrification; Isotopic enrichment of nitrate; nitrate contamination』

1. Introduction
　1.1. Study area description
2. Methods
　2.1. Aquifer sediment samples
　2.2. In situ mesocosms
　2.3. Geochemical modeling
　2.4. Multiport samplers
　2.5. Mass balance computations
3. Results and discussion
　3.1. Aquifer sediment samples
　3.2. In situ mesocosms
　　3.2.1. Denitrification rates
　　3.2.2. Enrichment of ¹⁵N
　3.3. Geochemical modeling
　3.4. Multiport samplers
　3.5. Mass balance computations
4. Conclusions
Acknowledgments
References