Photochemical production of CO and CO2 in the Northern Gulf of Mexico: Estimates and challenges for quantifying the impact of photochemistry on carbon cycles

The photochemical production of carbon dioxide (CO2) and carbon monoxide (CO), derived from marine colored dissolved organic matter (CDOM), is considered a significant oceanic loss process for the large and variable fluxes of terrigenous dissolved organic carbon (DOC) in river dominated coastal systems. Using samples collected over 4 seasons (2009-2010) in a grid centered on the Mississippi-Atchafalaya River system in the Northern Gulf of Mexico, we attempt to improve constraints for these direct photochemical impacts on DOC cycles by contributing the largest coherent set of photochemical apparent quantum yield (AQY) spectra for CO to date (n = 99), 18 of which had CO2 AQY spectra determined in the same sample for use in calculating CO2:CO photoproduction ratios. Simple correlations, previously reported from much smaller data sets, between CO photoproduction and CDOM optical properties showed weak or no correlations within our much larger spatio-temporal study. However, grouping samples with an optical transition point (a(g)(320) = 1.3 m(-1)) allowed definition of two distinct inshore and offshore CO AQY spectra. Inputting these in regional photochemical models that use remotely sensed ocean color data and modeled water optical properties dramatically improves results over using a single CO AQY spectrum (modeled vs. measured r(2) = 0.73 vs. 0.18 for single AQY), as has been done previously. Monthly average CO photoproduction rates ranged from 6.0 to 17.7 mu mol m(-2) d(-1), amounting to a conservative estimate of 3.35 Gg C yr(-1) for our study region. Our data also shows that the efficiency for CO2 photoproduction decreases over prolonged irradiations up to 48 h. Consequently, no assessable method exists for direct determination of initial rates in clear waters with low CO2 photoproduction. Therefore CO2 photoefficiency in the Gulf of Mexico was inferred using CO photoproduction estimates and CO2:CO ratios. In agreement with previous studies, CO2: CO ratios determined here were also poorly constrained, ranging from similar to 6 to 66 providing a median value of 24.4. This approach estimates that direct photochemical production of CO2 plus CO can remineralize DOC on the order of 85 Gg C each year. This estimate provides an improvement over the use of a single inshore CO2 AQY spectrum for a largely blue water system but caution should be used when interpreting CO2 photochemical flux estimates that rely heavily on the selection of CO AQY spectra and/or poorly constrained relationships between CO2 and CO photoproduction. Our results stress that the continued expansion of the CO2:CO ratio database to new oceanic regimes is likely of very limited value and that assessable direct methods or better proxies to quantify CO2 photochemistry in clear marine waters are needed. (C) 2015 Elsevier B.V. All rights reserved.