OS31D-1023:
Upwelling Scales off the Coast of Peru: Comparison of Observation and Model

Wednesday, 17 December 2014
Jorge Vazquez1, Toshio Michael Chin2 and Edward M Armstrong1, (1)NASA Jet Propulsion Laboratory, Pasadena, CA, United States, (2)Jet Propulsion Laboratory, Pasadena, CA, United States
Abstract:
Upwelling regions of the world’s oceans are home to some of the most productive fisheries. Yet these coastal areas provide unique challenges for remote sensing from satellite platforms because of both their proximity to land (radar interference) and typically small horizontal scales (< 50km) of upwelling processes.

Comparisons are performed on the gradient of sea surface temperature (SST) fields derived from multiple sources:

1) the 0.25 degree resolution National Climatic Data Center (NCDC) Optimally Interpolated AVHRR+in-situ or AVHRR_OI, data set.

2) the 1km resolution Multi-scale Ultra-high Resolution (MUR) gridded SST data set,

3) the 0.25 degree resolution SST derived from the WindSat microwave sensor,

4) a 2km version of the Estimating the Circulation and Climate of the Ocean Model (HECCO2).

Temporal and spatial correlations between HECCO2 and MUR, as well as between HECCO2 and NCDC, are examined through the dominant singular vectors (eigenmodes) of the covariance matrix for each pair of data sets. In both cases the first mode of covariability accounts for over 90% of the total variance. A simple technique based on SST gradients is then applied to the first mode to determine the upwelling scales based on HECCO2, MUR, and NCDC.

Longitudinal sections at 8S, 20S, and 30S indicate that the upwelling scale decreases between 8S and 20S. At 20S the first mode of covariability between MUR and HECCO2 indicate an upwelling scale between 25 and 50km. Results are consistent when compared with chlorophyll-a data from MODIS-Aqua. Such upwelling scales are not seen in the WindSat data and reduced in the NCDC data. A reduction of the upwelling scale by a factor 0.2 between 8S and 20S is consistent with a dependence on the Coriolis parameter. A major conclusion of the work is that magnitudes of SST gradient and upwelling scales derived from MUR are consistent with those of the HECCO2 for the test period of October-November 2011. Additionally, it is shown that to resolve upwelling scales near the coast high resolution infrared data must be used in the analysis. Microwave derived SSTs, such as those from WindSat are of limited value when upwelling scales are less then 50km.