Selected modeled output variables from ROMS 4DVAR model fits for the P1706 and P1604 CCE-LTER process cruises

Values for Kz and w were extracted from a three-dimensional physical model (9 km horizontal resolution, 42 vertical layers) developed using the Regional Ocean Modeling System (ROMS) four-dimensional variational data assimilation system (4DVAR) (Moore et al., 2011), then depth-interpolated linearly to calculate values at the MLD. Initial and boundary conditions were taken from the California Current System Reanalysis product produced by the University of California, Santa Cruz. Surface forcing was taken from the 9-km resolution Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS) (Doyle, Jiang, Chao, & Farrara, 2009; Hodur, 1997). We assimilated satellite remote sensing products (sea surface height from Ssalto/Duacs and sea surface temperature from AVHRR) and in situ temperature and salinity data vertical profiles collected during CTD measurements on our cruise. Assimilated data were used to update model initial conditions and surface forcing to generate a dynamically-consistent month-long model fit covering the duration of each cruise and producing 8-hr averaged outputs. For detailed descriptions of our ROMS 4DVAR approach, see (Song et al., 2012) and (Miller, Song, & Subramanian, 2015). Kz and w were interpolated to CTD cast locations from the 9-km ROMS 4DVAR solution using bilinear 2D interpolation from the four points comprising the nearest neighbors to the cast coordinates.