As part of a longer term SeaWiFS/modeling data assimilation project, we examine the spatial-temporal decorrelation scales of daily, level-2 and level-3, standard mapped images of SeaWiFS ocean chlorophyll for the North Atlantic Ocean and the global ocean. The decorrelation scales, defined mathematically as the time or space distance over which neighboring data points are correlated at some specified confidence level, can be both non-stationary and anisotropic, providing important information on biological-physical interactions. They also are a necessary component of future work to objectively analyze the chlorophyll images on equal time slices prior to assimilation using the best linear, unbiased estimates from kriging. The primary focus here is on the spatial scales, which are determined from the semivariograms of the log-normalized SeaWiFS chlorophyll values. Two regimes are evident in the semivariograms, a mesoscale where decorrelation scale lengths vary with latitude comparable to the local Rossby deformation radius and a regional scale governed by basin-wide patterns of nutrient supply, light, and seasonal mixed layer depth. We find that, as expected, the zonal decorrelation scale lengths for the regional regime are considerably longer than the corresponding meridional scales. To help us resolve the shorter, mesoscale influences and test our confidence in the level 3 calculations, we examine also level-2, 1km data from the HRPT station at the Bedford Institute of Oceanography in the northwestern North Atlantic for specific periods. To complete the objective analysis a similar study with emphasis on the temporal decorrelation scale lengths, extracted from semivariograms of an entire annual cycle of satellite data, will be performed when the data is available.
The North Atlantic study served as a feasibility study and following this link will take you to a slide presentation of a talk that was presented at the 1998 AGU Fall Meeting in San Francisco and the 1999 SeaWiFS Science Team Meeting in Monterey, CA.
The global study is a complete semivariogram analysis of the year 1998 (Jan-Dec) of all of the daily, level-3 standard mapped images (SMI). Following this link will take you to a web-page we are using to rapidly disseminate results of compute intensive calculations to all of the investigators working on this project. This web page is subject to change on an almost daily basis.
This study is intent on quantifying the seasonal and interannual variability in mesoscale biological variability signals from a global multi-year (1998 - 2005) time-series records of level-3 SMI SeaWiFs and MODIS-Aqua data.
Characterizing the submesoscale component of ocean color variability to compare and contrast with mesoscale variability is main goal of this study. We use structure function (semivariogram) techniques applied to high spatial resolution (1 km), regional satellite data (initially centered on the Bermuda area).