Following the disastrous 2010 Deepwater Horizon oil spill, numerous studies have 19 been carried out to investigate the impact of the oil spill on a variety of environments. 20 However, it is currently unknown whether the spilled oil transported to the coastal ocean 21 has caused any discernible perturbation to the inorganic chemistry of the water column. 22 In this work we compared and contrasted a multiyear dataset (2006-2012) collected in the 23 northern Gulf of Mexico continental shelf, an area subject to frequent summer hypoxia. 24 Before and after the oil spill, apparent oxygen utilization (AOU) and dissolved inorganic 25 carbon (DIC) in bottom water samples all showed consistent relationship that was close 26 to Redfield reaction stoichiometry. However, we observed a possible oil degradation 27 signal in the bottom waters during a July 2010 cruise as manifested by a significant 28 deviation from all other years in the relationship between water column apparent oxygen 29 utilization (AOU) and dissolved inorganic carbon (DIC). Based on stable carbon isotope 30 analysis of bottom water DIC from a July 2011 cruise in the same region, oil carbon 31 degradation in the water column was likely negligible and the shelf water had returned to 32 the pre-spill conditions. 33 1. Introduction 34 Dissolved inorganic carbon (DIC), nutrients (nitrogen and phosphorus), and dissolved 35 oxygen (in the form of "apparent oxygen utilization" or AOU, the difference between 36 saturated oxygen concentration and observed oxygen concentration) have long been used 37 to calculate the composition of the organic matter (or the "Redfield ratio") undergoing 38 remineralization in the ocean (e.g., Takahashi et al., 1985). Here, AOU in a water mass 39 represents the net consumption of oxygen from the time when the parcel of water was last 40 exposed at the air-sea interface. The "isopycnal method" to examine evolution of 41 chemical species along individual isopycnals (Minster and Boulahdid, 1987; Takahashi et 42 al., 1985) or "neutral" surfaces (Anderson and Sarmiento, 1994), as well as water mass 43 mixing models coupled with a multilinear regression approach (Li et al., 2000; Li and 44 Peng, 2002) are appropriate methods used to derive relatively confined ranges of C, N, P 45 ratios in remineralizing organic matter as well as the reaction stoichiometry between the 46 reacting organic matter and dissolved oxygen. Conversely, the Redfield ratio obtained in 47