A first look at a new interannual ROMS solution for the Canary Basin Evan Mason 1 , Ananda Pascual 1 , James C. McWilliams 2 1 IMEDEA (CSIC-UIB), Esporles, Illes Balears, Spain 2 UCLA, Los Angeles, California, United States The Canary Basin • Region of great interest to oceanographers of all disciplines (Figure 1). • Role in the Meridional Overturning Circulation. • Eastern boundary: elevated mesoscale activity; Canary Up- welling Ecosystem. • The Canary Islands: barrier to the Canary Current; the Canary island wake and Canary Eddy Corridor. • Entrance and spreading of Mediterranean Water. Figure 1: Principal oceanographic features of the Canary Basin in the northeastern Atlantic Ocean.The boundary of the model domain is outlined in red. Red dots show the positions of five Subduction Experiment (SubExp; Weller et al., 2005 ) moorings. Motivation for an interannual solution: • Support for cross-disciplinary studies, e.g., larval exchange across the Canary-African transition zone. • Open questions, e.g., Canary Current variability. • Support for individual cruises, e.g., the PUMP project cruise in autumn 2014. • Building on the previous climatological modelling work of Ma- son et al. (2011); see Figure 2. Figure 2: Annual mean sea surface height variance from (a) the climatological ROMS solution of Mason et al. (2011) and (b) AVISO. The 300 and 3000 m isobaths are plotted in white. Model setup Regional Oceanic Modelling System (ROMS): Version 3 of the ROMS-AGRIF code. • Four open boundaries (Mason et al., 2010). • Gibraltar strait: parameterised barotropix flux; Neumann con- dition. • Weak sponge (50 m 2 s −1 ) in order not to suppress variability along western boundary. Period: 1985-2014. Domain: Rotated 7.5 km grid (Figure 1). Boundary forcing: Interpolation from global interannual model so- lutions with data assimilation such as SODA, ECCO2 and Mer- cator. • SODA 2.2.8: Monthly 0.5 ◦ averages from a 30 member en- semble; 1985-2010. • SODA 2.1.6: 5 day 0.5 ◦ snapshots; 1985-2008. • ECCO2: 3 day averages at 0.25 ◦ ; 1992-2012. • Mercator: 1 day averages at 0.08 ◦ ; 2010-2014. Surface forcing: Six-hourly bulk fluxes from CFSR/CFSv2 (next generation NCEP). • CFSR: Resolutions between 0.5 ◦ and 0.3 ◦ ; 1985-2010. • CFSv2: Resolutions between 0.5 ◦ and 0.2 ◦ ; 2011-2014. Presently, two solutions exist: 1. ROMS + SODA2.2.8: 1985-1999 2. ROMS + SODA2.1.6: 1985-2004 Preliminary results with SODA 2.2.8 Sea surface and mixed layer temperature • SST: consistently good agreement for SST snapshots. • Annual and seasonal SST means yet to be checked. Figure 3: Sea surface temperature snapshots on 9 December 1992 from (left) satellite SST and the ROMS+SODA 2.2.8 solution (right). The positions of the SubExp moorings in Figure 5 are marked. • Coherence in seasonal evolution between model and observa- tions. • The model responds well to local forcing. Figure 4: Comparison of temperature evolution between 1999 and 1993 in the top 200 m from the five SubExp moorings (left) and ROMS+SODA 2.2.8 solution (right). Periods/depths with missing mooring data are left blank. See Figure 3 for the locations of the moorings. Annual mean sea surface height variance comparisons ROMS+SODA 2.2.8 and ROMS+SODA 2.1.6 • Increases in model variability across entire domain (Figures 5 and 6). • Can be explained by high-frequency surface forcing with respect to climatological solution (cf. Figure 2a) • But, patterns are inconsistent; in particular, the Azores Current is weak and displaced northwards. Figure 5: Annual mean sea surface height variance from (a) ROMS+SODA 2.2.8 and (b) AVISO. The 300 and 3000 m isobaths are plotted in white. Figure 6: Annual mean sea surface height variance from (a) ROMS+SODA 2.1.6 and (b) AVISO. The 300 and 3000 m isobaths are plotted in white. Diagnosis of the SODA solutions • SSH variance patterns in Figures 5a and 6a are unsatisfactory. • Suggests that boundary forcing and/or conditions are at fault. • Surface forcing appears to be reliable. • Strong coincidence between wintertime periods of high dis- charge and high SLA over the Ebre shelf (ρ=0.56, p<0.01)). Annual mean surface height variance: • Predictor of Western Intermediate Water (WIW) presence in Ibiza Channel (Monserrat et al. 2008). • WIW is thought to block the channel (Pinot et al. 2002). Ebre river outflow: • Figure 7 suggests strong link between Ebre freshwater dis- charge and mean seasonal SLA patterns (cf . Figure ??c,d; Salat et al., 2002). • Anticyclonic anomaly coincides with rapid Ebre discharge in- creases that are characteristic of the autumn season. • Strong coincidence between wintertime periods of high dis- charge and high SLA over the Ebre shelf (ρ=0.56, p<0.01)). Figure 7: Time series of monthly mean Ebre river discharges at Tortosa near the Ebre delta from SAIH (begins 2003), and monthly mean SLA over the Ebre shelf. Winters with large SLA over the Ebre shelf are shaded in green. Ibiza Channel variability as function of WIW: • Wintertime southward current anomalies stronger in absence of WIW (Figure 8). • Demonstration of viability of altimeter data for study of Balearic channel variability. Figure 8: Seasonal PDFs of meridional surface geostrophic velocity anomalies through the Ibiza channel. Two time ranges are sampled that correspond to years when Western Intermediate Water is present (WIW T ) or absent (WIW F ) in the channel. Conclusions • Use of long-term altimeter SLA dataset to study multi-scale vari- ability in the Balearic Sea. • Northern Current anomaly over Ebre slope is maximum in au- tumn rather than winter (Figure ??). • Balearic Sea EKE is highest in autumn and weakest in spring. (Figure ??). • Ibiza channel geostrophic velocity anomalies indicate that mean Northern Current through the channel is strongest in win- ter (Figure ??). • Geostrophic velocity anomalies for years with and without Ibiza channel WIW are consistent with theory that WIW blocks the channel in winter (Figure 8). • This work demonstrates the value of homogenized altimetric time series for studies of oceanic mesoscale processes and variability. Acknowledgments Evan Mason is supported by a Spanish government JAE-Doc grant (CSIC), cofinanced by FSE. The altimeter products were produced by Ssalto/Duacs and distributed by Aviso, with support from Cnes (http://www.aviso.oceanobs.com/duacs/). CFRS/CFSv2P Reanalysis data provided by NCAR’s Data Support Section from their Web site at http://rda.ucar.edu/. IMEDEA - Mediterranean Institute for Advanced Studies  emason@imedea.uib-csic.es WWW: http://www.imedea.uib.es