The PWS surface current (as seen by the HF radar) fluctuates hourly, daily, and throughout the FE. To the first order, one can attribute these current fluctuations to three major factors: 1) atmospheric wind, 2) geostrophic current associated with the density and temperature/salinity structure in the upper ocean (mostly influenced by the fresh-water input from rainfall, land runoff, rivers, and glaciers), and 3) tides. It is a complex interplay (through non-linear interactions) among these three factors that is responsible for the flow pattern we see in the HF radar images.
The dominant tidal component is the 12 hourly M2 tide. There is also a pronounced diurnal changes (land-sea contrast associated with the solar heating), particularly near the coaslines. We can reduce and eliminate the tidal signal and diurnal variations by looking at the 24-hour averaged surface current map.
Thus, there are only two dominant factors of (1) wind and (2) geostrophic current that determine the surface current pattern.
In the summer seasons, the PWS circulation is often seen as a cyclonic flow or eddy, with strong inflow from HE (there is a tidal forcing for this as well). This cyclonic flow is consistent with the density/salinity distribution associated with the fresh-water input from the rivers/runoff/glaciers. It is also consistent with the conservation theory of the vorticity: northward (with increasing Earth rotation f) inflow is compensated by the decreasing relative vorticity of cyclonic motion. If there is no wind or wind is weak, this cyclonic eddy should be evident.
For the moderate to strong wind conditions, the wind-driven circulation should influence the final surface current pattern. The wind-driven
Ekman transport is 90 degrees to the right of the wind.
During the first week of our FE, the PWS is dominated by a very steady and strong easterly wind (blowing from east to west), and therefore producing a northward Ekman transport. This should strengthen the northward inflow from HE, and reduce or even eliminate the southward return flow in the western flank of the cyclonic eddy.
For the past few days, the wind in the central sound has been fluctuating, switching between easterly and westerly wind conditions. When the easterly wind switched to westerly a few days ago, a cyclonic eddy emerged. When the wind became easterly again yesterday, the cyclonic eddy disappeared, and the surface current exbihits rapid fluctuations and often with complex and smaller scale patterns.
Today's WRF forecast suggests that the weak easterly in the evening of July 30 Thursday will gradually switch to a strong westerly during the rest of today. If the above theory is correct, then we should see the return of this cyclonic eddy in the central sound later today Friday. If the westerly wind persists in the central sound as predicted by WRF, the cyclonic eddy should also persist as predicted by ROMS.
If this prediction can be verified in the next 12-24 hours, we should be able to gain confidence in the WRF/ROMS predictions, which can then be used to guide our field deployment strategy (also known as adaptive sampling) in the final days of our FE.
If you have any questions or need information/materials (that were used to make these conclusions), please feel free to contact me.
One of the most challenging task in modeling the PWS circulation is to ensure sufficient fresh-water input into the sound coming from rainfall, rivers, runoff and glaciers. We have been estimating the fresh-water input from rivers and runoff based on the WRF simulated rainfall over land (routing them through the various watersheds to put them back in the right place around the sound). However, the fresh-water inflows from glaciers remain unknown, as a result, not represented in ROMS. Thus, the ROMS simulated salinity is significantly saltier than observed.
Question: What is the circulation pattern (both at surface and depths) and the associated T/S gradient for the western flank of the cyclonic eddy?
As you can see from these images (color represents speed, arrow represents direction, and length of arrow is proportional to the speed) of current at surface (0-meter), 10-meter, and 20-meter, ROMS clearly shows a closed cyclonic eddy in the center of the sound at 20-meter depth. It is less evident in the surface and 10-meter current. With more cross-sections data being collected and assimilated, the surface circulation may improve in the coming days.
Here are two more sections proposed for sampling in the coming days.
In case you didn't see it yet, the WRF hourly wind map (left panel) is displayed at the JPL OurOcean/PWS portal (http://ourocean.jpl.nasa.gov/PWS, click WRF Wind). When clicking the weather buoy/station, users can display the comparison between the observed and WRF simulated wind speed and direction (right panel). By evaluating the performance during the past two days, one can estimate the uncertainty in the 2-day forecast. 
results will tell us. Among the crowd were fishermen, resource managers, some long-time residents, and several visitors from Anchorage who were in Cordova for a Native Village of Eyak sponsored conference on alternative energy options the community might explore.
Channel 11, the Alaska Public Radio Network, Congressman Don Young's staff, the MMS, Chugach National Forest, Petroluem News, and others. We even got to see a drifter. Great job Scott! 
Hello from the Sound.
Photo: Hank Statscewich
