The Prince William Sound (PWS) is located in the northern Gulf of Alaska(GOA). While relatively exposed to open ocean on its southern boundary, to the north, east and west, PWS is embayed by a rugged, elevated, and highly glaciated mountain barrier. In the cold season this barrier tends to separate the cold, dry interior continental air mass from the relatively warm, moist maritime airmass in the northern GOA. Frequently occurring low pressure systems in combination with the high terrain act to induce a variety of strong off-shore and along-shore local winds [12
]. In the cold season, the interior of Alaska often is dominated by high pressure near the surface, and the PWS relatively low pressure. Consequently, pressure gradients are maintained around the PWS and gap flows are common in major gaps and fjords in PWS.
A further unique characteristic of Prince William Sound is that it contains Port Valdez along its north boundary. Port Valdez is the southern terminus of the Trans Alaskan Pipeline System (TAPS). Hence, almost all the crude oil extracted in Alaska transits the full N-S extent of PWS in supertankers en route to ports of call beyond the GOA. The potential perils of such a method of oil transport were partially realized in 1989, with the infamous accidental grounding of the tanker Exxon Valdez
and attendant massive oil spill. While the Exxon Valdez
incident did not ultimately result from adverse weather conditions, it brought into sharp focus the potential consequences of navigating the PWS with its highly localized adverse weather conditions [23
Gap flows are common and gusty [12
] in PWS. The strength of these gap winds is strongly affected by the imposed pressure gradients. Strong pressure gradients often generate strong gap winds. Many previous studies on gap winds in the Northern GOA have been conducted over major gaps and consisted largely of in-situ aircraft observations. For an example, Macklin et al. [12
] documented airplane observations of gap flows in Copper River and Resurrection Bay; Macklin et al. [13
] also reported an airplane observation of a case involving gap flow in the lower Cook Inlet. These studies rely on in-flight observations limited in areal extent and are further constrained by the limitation of the observing instruments and instrument platforms. For example, an instrumented aircraft must fly a finite height above ground, inherently missing observations in close proximity to the surface.
Numerical weather prediction (NWP) models and satellite observations have also been combined to study gap winds. Streenburgh et al. [24
] studied a gap wind in Tehuantepec of Mexico using MM5 and satellite images. Pan and Smith [18
] studied gap winds in the Alaska Aleutians using a shallow water model and space-borne Synthetic Aperature Radar (SAR) images. Sandvikv and Furevik [21
] studied high latitude gap winds at Spitsbergen, Norway. Winstead et al. [25
] used SAR image and mesoscale models to better understand barrier jets in Northern Gulf of Alaska. Liu et al. [9
] reported a climatology of simulated gap winds in Cook Inlet and Shelikof Strait of Alaska.
Space-borne Synthetic Aperture Radar (SAR) observed normalized radar cross section (NRCS) depends upon scattering from surface roughness elements within the beam footprint. The NRCS from water represents the roughness of the water body surface which is a non-linear function of near-surface wind stress. With a prior estimate of the wind direction, wind speed can be inverted from the NRCS [25
]. Comparisons with buoys, models and scatterometer estimates have demonstrated that the accuracy of wind speed estimated from SAR is comparable to those of operational scatterometers when the proper wind direction is used in the NRCS inversion [15
]. At the Johns Hopkins University (JHU) Applied Physics Laboratory (APL), the Alaska SAR Demonstration Project has been operating for almost a decade. High resolution “snapshots” of the near-surface wind speed field have been processed using the Navy Operational Global Atmospheric Prediction System (NOGAPS) winds as a first guess along with the NRCS intensity observation [7
] and compared well with buoy observations in GOA [6
This site archives a vast number of the near-surface wind field images associated with a wide variety of mesoscale and synoptic-scale atmospheric phenomena. Using this rich archive, several studies have investigated boundary layer convection [1
], gap flows [9
], barrier jets [11
], polar lows [22
], and synoptic-scale fronts [26
This study uses SAR wind imagery from the JHU APL archive and a high-resolution (1 km in horizontal grid spacing) mesoscale model to document gap winds for gaps of small lateral dimension in PWS. Section 2 describes two cases of gap winds captured by SAR in PWS. Section 3 presents the mesoscale modeling of these two events. Section 4 discusses the simulated characteristics of the gap wind, and is followed by the conclusions in Section 5.
2. SAR observed Gap winds in Valdez and Wells Passage
The Valdez Arm is a north-south trending gap located in the northern coast of PWS while the Wells Passage is a broader east-west oriented sea level gap (Fig.1
) among the elevated coast lands and islands in western PWS. The Wells Passage is the extension of the relatively low elevation areas connecting the Cook Inlet and PWS. The Valdez Arm is an opening in the mountains of the north-central coast of the PWS, winds frequently blow from these channels— sometimes simultaneously— into the center of PWS.
is a SAR wind image which captures a weak gap wind event in the Valdez Arm and the Wells Passage respectively at 3:29 UTC 18 Feb 2007. The arrows in the image represent NOGAPS wind directions used in the computation of the SAR wind speeds [14
]. This image shows that the wind speed in the Valdez Arm and the Wells Passage are in the 12 to 17 ms-1
range. Relatively weak westerly winds dominate the PWS suppressing the northerly channeled wind exiting the Valdez Arm. The boundaries between the westerly flow and the north-easterly flow are clearly shown as a narrow zone of weak wind on the image. (Note the SAR wind image is a bit off in position compared to the land mask; the image should be moved a few pixels northeastwards. This slight offset in navigation induced thin red areas around coastlines in the image which can be falsely interpreted as strong winds.) Figure 2b
shows the 0Z surface analysis which corresponds closely in time to this SAR image. The low system in the Northern Gulf of Alaska is moving southeastwards (not shown) away from PWS causing the pressure gradient in the PWS to gradually relax. At the point in time considered here, the pressure gradients induced by the high pressure in the Interior and the low pressure in PWS are weak. Consequently, the gap winds are weak in the PWS.
The Figure 3a
SAR wind image shows a strong northeasterly gap wind of approximately 25 ms-1
in the Valdez Arm and a westerly strong gap wind of approximately the same strength from the Wells Passage. These two perpendicularly aligned jets meet near the center of the PWS. The combined wind jet further extends southeastwards out the PWS through the Hinchinbrook Entrance with a weaker speed of 15 ms-1
. The extension of the Wells Passage jet turns right as the result of pushing from the Valdez jet and the right-turning nature of a gap wind [24
]. Note, as in Figure 2
, there is also a slight offset of the image relative to the land mask which results in some narrow red areas along the coastlines. The NOGAPS winds used in the SAR wind retrieval inversion are northerly in the PWS and its vicinity. Clearly, the coarse NOGAPS wind does not resolve the local winds well in the western PWS, which may cause significant wind speed errors [6
]. (How the deficiency of NOGAPS wind direction affects the accuracy of the wind retrieval is beyond the scope of this study.) Apparently, the Wells Passage winds are very strong in this case. However, the SAR-wind retrieval for the Valdes Arm has proper wind direction for both cases, which are our focus in this study. Figure 3b
shows the surface synoptic pattern. A low system in the Gulf of Alaska is moving further northwest towards PWS. The high pressure in the Interior and the falling pressure in the northern Gulf form strong pressure gradients around the PWS which in turn induce strong gap winds in the Valdez Arm and the Wells Passage.
The above SAR images capture a weak wind event and a strong wind event respectively; and the surface pressure charts show there are subtle yet significant differences between these two to reveal the details of processes governing these two events, we use a numerical mesoscale atmospheric model to perform high resolution simulations. The following section introduces the model and discusses the simulation results.
High resolution SAR wind imagery and a mesoscale model were combined to better understand gap winds in the PWS. SAR wind imagery reveals detailed wind features over the water body surface while numerical mesoscale model simulations can be used accordingly to understand the highly localized atmospheric interactions over complex coastal terrain. Our focus was on the northeasterly gap wind in the Valdez Arm which is the largest northerly gap in PWS and the origin of the oil tanker route to the continental USA. The SAR wind imagery revealed the gap wind's areal extent and resembled the mesoscale model simulations. The high resolution mesoscale model simulations revealed that the Valdez gap winds were fueled by the down slope winds from the elevated areas surrounding the Arm. Both the weak and strong cases of this gap wind regime showed indications of a barrier jet, stronger winds adjacent to a physical barrier. The west-east vertical cross-sections crossing the middle of the Valdez Arm showed the air is descending on the eastern and rising on the western part of the Arm. The blocking of the barrier induced the upward motion of the air on the western part of the Arm. The eastern bank of the Arm also served as a barrier which accelerated the gap wind in the strong wind case. The westerly winds from the Wells Passage have a significant impact on the Valdez gap wind, especially when the northerly flow is weak. The westerly Wells Passage winds turned southwesterly into the Valdez Arm and suppressed the development of the Valdez gap wind. In an ideal case, pressure gradients are solely responsible for a gap wind, but in “messy” complex coastal areas multiple mesoscale mechanisms likely act in combination to dynamically impact the gap wind. With the addition of SAR-wind imagery, a high-resolution mesoscale model can be used to reveal several of the 3 dimensional details of these interactive flows.
The Alaska SAR demonstration project has found many applications ranging from atmospheric boundary layer measurements, oceanic measurements and sea ice observations [7
] — wind retrieval is a part of the atmospheric boundary layer measurements. The SAR wind archive of JHU APL provides a valuable source to begin understanding flows in the extremely complex flow found in PWS. The high spatial resolution of SAR imagery in combination with the mesoscale model's flexible temporal and spatial resolution have proven to be useful in helping elucidate complicated low level flows over complex coastal regions. Reliable availability of SAR snapshots may help improve modeling the atmospheric boundary layer when assimilated into a model. However, this speculation needs to be tested. We anticipate more studies on complicated flows in the Gulf of Alaska using SAR wind retrievals and mesoscale modeling.