Basic HTML Version
International Journal of Marine Science 2013, Vol.3, No.34, 267-277
http://ijms.sophiapublisher.com
268
In this paper, the ability of reanalysis products to
represent coastal upwelling in the southern Benguela
is evaluated in the period 1980-2008 using mean maps
and depth sections. The annual cycle and anomalies
are studied in the St Helena Bay sub-area. The paper
addresses the following research questions: How is the
structure and variability of coastal winds and
upwelling represented? Which model databases best
capture the slope of isotherms over the shelf? What
temporal variability is evident in coastal upwelling?
What are the environmental influences on chlorophyll
enrichment and fish catch? An overarching goal of the
paper is to show that modern reanalysis products are
capable of describing coastal upwelling and its
variability and impacts in the southern Benguela.
2 Data and Methods
Three monthly reanalysis datasets are considered in
the period 1980-2008, two for the lower atmosphere
and one for the upper ocean. Winds, sea surface
temperatures (SST) and atmospheric features just west
of South Africa are analyzed from monthly Coupled
Forecast System (CFS, Saha et al., 2010, 0.3º
resolution) and European Community Medium-range
Weather Forecast (ECMWF, Dee et al., 2011, 0.7º
resolution) reanalysis fields drawn from the Climate
Explorer website. These two products are generated
using rather different numerical models and
techniques of data assimilation. Results over the
elevated plateau and deep ocean are masked to focus
on the shelf and coastal environment (28-36°S,
14-20°E). Oceanographic fields are provided by
monthly Simple Ocean Data Assimilation version
2.1.6 (SODA, Carton and Giese, 2008) from the IRI
Climate Library, that include temperature, salinity,
currents and vertical motion in the satellite era. SODA
is forced by ECMWF wind stress and data are
available at 0.5º resolution. Here, sub-surface data are
compared with the National Center for Environmental
Prediction (NCEP) Global Ocean Data Assimilation
System (GODAS, Behringer and Xue, 2004) and the
World Ocean Atlas (WOA, Locarnini et al., 2006).
The monthly reanalysis products incorporate in-situ
and remotely sensed data: coastal and ship
observations, scatterometer and passive microwave
winds, drifters and ARGO profilers, infrared and
microwave surface temperature (Smith and Reynolds,
2004), microwave altimeter sea surface height, and
satellite estimated air temperature, humidity and
cloud-drift winds. Data from the South Africa Data
Center for Oceanography and GODAS are incorporated to
provide adequate coverage of the Benguela shelf in
the satellite era. The reanalysis products assimilate
observations via a numerical model, which for CFS
and SODA is the Geophysical Fluid Dynamics Lab
Modular Ocean Model version 4 (Griffies et al., 2004)
and for ECMWF is the ORA-S3 model (Balmaseda et
al., 2008). The representation of coastal upwelling and
its forcing is evaluated from long-term means and
standard deviations 1980-2008. Oceanographic and
atmospheric structure is analyzed as annual hovmoller
plots and depth or height sections west of St Helena
Bay (32.5°S, 14°-19°E).
Surface upwelling and shelf productivity are analyzed
using monthly ~0.1° resolution SST and chlorophyll
estimates from MODIS and SeaWifs satellites in the
period 1998-2010, obtained from the National Aeronautics
and Space Administration (NASA). Past research has
found that chlorophyll is better related to winds and
currents than SST in west coast upwelling zones
(Demarcq, 2009; Freon et al., 2009). Index-to-field
correlations are calculated for the St Helena Bay
sub-area: 32.2-32.7°S, 17.2-18.2°E (cf. Figure 7a); a
nursery for South Africa’s pelagic fishery (Agenbag et
al., 2003) within the perennial Cape Columbine
upwelling plume. Temporal cross-correlations are
analyzed between SODA wind stress, 1-100 m
depth-averaged currents, vertical motion, temperature
and salinity (1990-2008), and chlorophyll (1998-2008),
considering 1. the annual cycle, 2. monthly values and
3. anomalies. Statistical significance is evaluated
according to the degrees of freedom as limited by the
record length and filters applied, which subsequently
affects the persistence.
Inter-relationships are studied using wavelet-smoothed
time series for SODA zonal wind stress, CFS wind
vorticity, chlorophyll and various climate indices
(from Climate Explorer). The wavelet filter eliminates
cycles <18 months to focus on inter-annual
fluctuations of importance to the biota. Environmental
fluctuations and fish catch in territorial waters west of
South Africa (from Food & Agriculture Organization,
cf. Jury, 2011, 2012) are evaluated in the period
1980-2008. The catch of hake, sardine (pilchard) and
anchovy ranges from 235 to 738 10
3
T/yr in the area: