Characterization of SAR Image Features of the Ocean as a Function of Wind Speed and HF Radar Products
Vicente, Ricardo1; Paduan, Jeffrey2; Holt, Benjamin3; Cook, Mike2
1Instituto Hidrografico, PORTUGAL; 2NPS, UNITED STATES; 3JPL, UNITED STATES
Assessment of coastal ocean conditions is valuable for both military and civilian operations. Remote sensing of those conditions can be even more valuable, particularly in the case of all-weather sensor types. The potential for better understanding of ocean conditions through the combination of remote sensing results was recognized here with the focus on Synthetic Aperture Radar (SAR) imagery and High Frequency (HF) radar-derived surface currents.
The hypothesis that combining remote sensing products may improve results was tested using SAR imagery and available HF radar surface current maps along central California. Data were obtained from 2007-2010 when the network of HF radar stations was operating relatively continuously. Over the same time period, 780 archived SAR images were identified and, of those, 31 images were chosen for detailed assessment by identifying representative images under weak, moderate, and strong wind conditions.
As expected, wind strength played a dominant role in determining the physical processes visible in the SAR imagery. Moderate wind speed of 2-4 m/s exhibited the most obvious ocean-related processes and the best correlation with features in the HF radar surface current maps. Surprising is the discovery that oceanographic features in the SAR imagery represent recent history of tracer advection over hours to days. As such, individual hourly, surface-current snapshots are not, perhaps, the best product for comparing with those features. Features in the daily-average currents, for example, appear more highly correlated with features in SAR imagery under moderate wind conditions.
SAR FEATURES SEEN IN HF-RADAR SURFACE CURRENT
The SAR image in Figure 1 provides examples of oceanic surface features within the HF-radar large scales. The image was sensed on 25 May 2008 at 0555 UTC. The local wind is weak (~ 3m/s) and from the southeast.
Figure 1. SAR image sensed on 25 May 2008 at 0555 UTC, HF radar currents (red) and surface winds (green) from the closest hour. Both SAR- and HF-radar products detect a frontal signature offshore of Point Sur (details in Figure 2) and an anti-cyclonic eddy offshore of Monterey (details in Figure 4). The average wind is ~ 3m/s (SAR data provided by ESA).
Offshore of Point Sur, Figure 2 shows a frontal signature seen in SAR with a brighter edge where one of the mechanisms at work is convergence. The visible flow field sensed by the HF-radar indicates the presence of a strong southeastward surface current that turns south and then east, in total agreement with the plume edge. The water mass closer to shore evidences the presence of biogenic slicks which might indicate recent upwelled water.
Figure 2. Detail of the SAR image sensed on 25 May 2008 at 0555 UTC, HF radar currents (red) and surface winds (green) from the closest hour. A frontal signature offshore of Point Sur is depicted by both SAR- and HF-radar products (SAR data provided by ESA).
Given that, we propose that HF radar and SAR are capable of measuring the same ocean phenomenon under low wind conditions and strong surface currents, providing the features are large enough. The divergence map in Figure 3 illustrates converging HF-radar currents (inside the circle) in the Point Sur area, which is coincident with the brighter frontal signature seen in the SAR image in Figure 2.
Figure 4 shows a detail of an anti-cyclonic eddy offshore of Monterey Bay. The HF-radar surface currents clearly demonstrate a vortical feature, and the SAR image illustrates a darker core with brighter boundaries and the presence of slicks. With relatively low wind speed, the surface current enhancement role is captured on the SAR image.
A principal hypothesis of the present investigation (Vicente, 2012) is to assess the existence of ocean features in the SAR imagery that are correlated to the surface currents as perceived by HF radar. Some results show ocean features retrieved by HF radar and imaged by SAR provided the local wind has low intensity, the current field is strong, and the features are large enough. However, a consistent pattern has not been established. The divergence validation test performed in the selected cases is positive and encouraging, yet a general conclusion could not be made.
There are several possible reasons contributing to a scarcely direct correlation between HF radar-derived surface currents and SAR ocean features, including:
One possibility to bridge the temporal differences between instantaneous HF-radar currents and SAR images is to use HF-radar, daily-average, surface-current map in which tidal and sea breeze effects are naturally removed, instead of the hourly-average product.
Wind speeds below 2-3 m/s cause SAR ocean features to fade and merge with noise levels, though the appropriate enhancement of surface roughness by other imaging processes can lead to the improved sensing of oceanic features. This study revealed that surface winds of 2-4 m/s constitute the suitable level of wind speed required to make assessments of oceanic surface currents influence on SAR images. Hence, future studies in this area should consider assessing SAR data in seasons and time of day when wind speeds are moderate. Strong wind speeds of 10-12 m/s may cause wind clutter to mask most of the ocean surface features and, in cases of large fetch, it may lead to a well-organized atmospheric imprints on the surface layer.
Vicente, R. (2012). Characterization of SAR image features of the ocean as a function of wind speed and HF radar products (Masterís Thesis). Retrieved from: http://calhoun.nps.edu/public/handle/10945/7424