Surface Ocean Temporal and Spatial Variability of Chl-a and SST on the South Atlantic Bight: Revisiting with Cloud-free Reconstructions of MODIS Satellite Imagery Travis Miles1 (tnmiles@marine.rutgers.edu) and Ruoying He2 1Rutgers IMCS, 2North Carolina State University Abstract: Daily, cloud-free Data Interpolating Empirical Orthogonal Function (DINEOF) reconstructions of sea-surface temperature (SST) and chlorophyll (Chl-a) satellite imagery are compiled into monthly mean images for a six-year period (2003 to 2008) and used to identify SST and Chl-a variability on the South Atlantic Bight. Reconstructing Missing Data: Cloud-cover can be as high as 60% onshore annually. To reconstruct cloud-covered satellite images, we utilize Data Interpolating Empirical Orthogonal Functions (DINEOF) (Alvera-Azcarate et al., 2007). DINEOF statistically reconstructs missing data temporally and spatially, resulting in daily reconstructed satellite images. DINEOF subtracts the temporal mean from X and flags missing values The procedure performs a singular value decomposition of X Iteratively using: Each successive iteration reconstructs missing data using the previous best guess DINEOF employs a cross-validation technique to determine the optimum number of iterations to perform We set our 295-day data set to matrix X We add the temporal Mean back to the final reconstructed data set The South Atlantic Bight (SAB) stretches from Cape Canaveral, Florida, to Cape Hatteras, North Carolina, or from approximately 28o N to 35o N respectively, and accounts for 700 km of coastline. Distinct dynamic regimes subdivide the SAB into the inner shelf (~ 0-20 m), middle shelf (~ 20-40 m), outer shelf (~ 40-70 m) and shelf slope (~ 70-200 m) (Atkinson and Menzel, 1985). Figure 5. EOF of natural log of Chl-a (mg/m3) (upper left and center panels), Chl-a total six year mean in (mg/m3) and principal components (center and bottom panels). Figure 6. SST (oC) EOF (upper left and center panels), total six year mean (upper right panel) and principal components (Center and bottom). Empirical Orthogonal Function analysis shows a clear seasonal cycle in the first mode of variability for both SST and Chl-a. SST and Chl-a mode 1 accounts for 95.75% and 46.35% of variability respectively. Chl-a has a highly regionalized pattern with values on the central SAB out of phase with the northern and southern SAB. Further study with subsurface Chl-a data is necessary to better understand this pattern. SST EOF mode 2 exhibits a seasonal cycle as well, which previous studies have suggested to be a function of seasonal stratification and local wind.(Miles et al. 2009). Chl-a EOF mode 2 is well correlated with the cumulative river transport onto the SAB, but accounts for a relatively small 10% of Chl-a variability on the shelf. Multivariate DINEOF: We use a multivariate version of DINEOF in this study, which includes SST, SST 1-day lag and concurrent Chl-a measurements in the same state vector. The procedure identifies the patterns of co-variability in this state vector and uses them to reconstruct missing data points. Figure 1. Blue arrows indicate rivers, red and green indicate a north and south transect, respectively, referenced in Figure 6. A study by Barnard et al. (1997) was one of the first to use long-term satellite imagery to study seasonal variability of ocean color on the SAB, using 372 images from the Coastal Zone Color Scanner to represent five-years (1981-1986). We revisit these findings using over 1800 images from six-years of Moderate Resolution Imaging Spectroradiometer (MODIS) satellite data. Figure 7. SST (upper panel) and Chl-a (lower panel) time-series of monthly means for 2003 to 2008 at transect locations indicated in Figure 1. Figure 8. Chl-a PC 2 with a 1-month lag (blue dashed line) and the cumulative river transport from the Cape Fear, PeeDee, Altamaha and Savannah rivers (green solid line). References: Alvera-Azcarate, A A. Barth, J.-M. Beckers, and R. H. Weisberg (2007), Multivariate reconstruction of missing data in sea surface temperature, chlorophyll, and wind satellite fields, J. Geophys. Res., 112, C03008, doi:10.1029/2006JC003660. Atkinson, L.P., and D.W. Menzel (1985), Introduction: Oceanography of the Southeast United States Continental Shelf, in Oceanography of the Southeastern U.S. Continental Shelf, Coastal and Estuarine Sciences 2, edited by L.P. Atkinson et al., pp.1-9, AGU, Washington D.C. Barnard, A.H., P.M. Stegmann and J.A. Yoder (1997) Seasonal surface ocean variability in the South Atlantic Bight derived from CZCS and AVHRR imagery, Cont.Shelf Res., 17(10), 1181-1206. Miles, T. N., R. He, and M. Li (2009), Characterizing the South Atlantic Bight seasonal variability and cold-water event in 2003 using a daily cloud-free SST and chlorophyll analysis, Geophys. Res. Lett., 36, L02604, doi:10.1029/2008GL036396. Acknowledgements: We are grateful for the research support NASA provided through grant NNX07AF62G and the NC NASA Space Grant. We also acknowledge NASA GSFC for providing MODIS data used in this research, NCEP NARR for providing wind data, NOAA NDBC for providing buoy data and Rutgers IMCS for providing presentation support. Six-year Monthly Means: Monthly mean SST (oC) (left panel) and ln(Chl-a) (mg/m3) (right panel) are computed using daily cloud-free DINEOF reconstructed MODIS satellite imagery. Daily reconstructions for 2003 to 2008 are available at: http://omglnx2.meas.ncsu.edu/travis/DINEOF/ Figure 2. Raw SST (oC) and Chl-a (mg/m3) (upper left and lower left respectively) with DINEOF reconstructed SST (oC) and Chl-a (mg/m3) (upper right and lower right respectively).