1. Evaluation of operational altimeter-derived ocean currents for shelf sea applications: a case study in the NW Atlantic D. Vandemark1, H. Feng1, and J. Wilkin Ocean Process Analysis Lab ,University of New Hampshire, NH, USA 2Institute of Marine and Coastal Sciences, Rutgers University, NJ, USA
II. Data and Methods
I. Introduction
Altimeter-derived gridded surface current products (Globcurrent) from (Version 3) produced using an optimal interpolation of multiple altimeter tracks, including both east and north components of
Surface geostrophic current Ugeo on a 1/4 degree (~25 km) spatial grid and with daily time resolution
Eulerian “total” current at significant wave height from the sum of surface geostrophic Ugeo and Ekman currents Uek on a 1/4 degree (~25 km) spatial grid and a 3-hour resolution
Note that the effective resolution of Ugeo and Uek currents is of the order of 5-10 days and km
In-situ buoy measurements: (IOOS-NERACOOS)
ADCP-derived ocean current measurements at six GoM buoys (Fig. 2) is used for Globcurrent validation( ).
hydrographic measurements at these buoys (depths of 1, 20, and 50 m) used for water mass advection analyses
Low-passed time series: formed by applying a 70 day running mean low pass filter to remove higher frequency aliasing but preserve seasonal to interannual signals.
The satellite altimeter record now extends for decades and can support examination of the mean state of ocean circulation dynamics in complex coastal regions such as the shelf of the Northwestern Atlantic (NWA) and the semi-enclosed marginal sea Gulf of Maine (GoM). 
This technical study evaluates the strengths and weaknesses of altimeter-derived multi-mission global gridded ocean current products (GlobCurrent and soon OSCAR) in their application to coastal process studies using the NWA shelf as a test case.  Specific objectives are:
to evaluate the quality of Globcurrent ocean vector currents ( ) against in situ current measurements from the US Integrated Ocean Observing System (IOOS)
to test data application to a known topic - how subsurface GoM hydrography is impacted by the upstream (remotely-forced ) advection
Fig 1 The regional map with mean surface current vectors on the NWA shelf. Some key locations include the Gulf of St. Lawrence (GSL), Scotian Shelf (SS) and Gulf of Maine. (GoM). Two inset boxes are addressed in III and IV below.
Globcurrent product : Buoy Validation
GoM subsurface salinity variability tied to southwest Scotian Shelf inflow and its controls ?
Fig. 5 Climatological mean surface geostrophic current vectors (red arrows) and ellipses (blue) from Globcurrent ( ) on the Scotian Shelf, in the Gulf of St. Lawrance, and along shelf-slope domain. Long-term satellite altimeter (Topex and Jason) repeat ground tracks are shown in black. In particular, data near six tracks (#100, 252, 074,191, 217, and 141) are selected for further analysis, with focus on shelf sections as indicated with dark black dots. Along-shelf geostrophic current Vg for these tracks are estimated by projecting the vector currents normal to the local isobath
Fig. 2 The climatological mean surface geostrophic current vectors from Globcurrent, (black) and mean buoy-measured upper ocean (average over 5-50 m depths) current vectors (blue) at six separate GoM buoy stations.
Track 141 Cabot Strait
(b)
Track100, S. Scotian Shelf
A recent study (Feng et al, 2016) found that interior GoM subsurface salinity (and thus biochemistry) variability at day time scales is closely tied to Scotian Shelf inflow that is, in turn, correlated with Vg change measured on altimeter track The study showed that satellite altimeter data allow monitoring/prediction of remote forcing of GoM hydrography due to this upstream advection. Here we examine Globcurrent data to further investigate potential sources of control on this variable advection process.
Fig 3 . Globcurrent surface geostrophic current (Ugeo, RED) vs. buoy current (BLUE) at station N (left) , I (middle) , and E (right). The upper panels show eastward and the lower the northward current components.
Fig. 6. Filtered time series of GoM salinity and Scotian shelf currents. Blue curve is 70-day moving average of Buoy B measured subsurface salinity at 50m depth and black is the lagged alongshelf altimeter-based (Globcurrent) current Vg near track 100 on the Scotian shelf. The 130 day time lag represents the offset for maximum correlation (R=0.8) between salinity and Vg100. Positive lag indicates Vg100 leads. Also note that negative Vg100 corresponds to downstream flow. This result is consistent with that were reported by Feng et al.( 2016) using dedicated altimeter processing of along-track data including near coast corrections.
(a)
Fig 4. GlobCurrent Utotal=Ugeo(surface geostrophic) + Uek (surface Ekman ) (RED) vs. buoy (BLUE) current at station N(left), I(middle),and E(right). The upper panels show the eastward and the lower the northward components .
Conclusions related to these preliminary Globcurrent validations
on the NWA shelf and near the shelf break, altimeter currents show excellent agreement with data from Buoy N in both mean and variation, including high correlation, low bias and RMSE.
but inside the GoM, Globcurrent data accuracy is more variable, showing better performance in the east (Buoy I) than in the west (Buoy E). The EW (u) component accuracy is higher than the NS (v) component. This is likely because of strength in u, aligned with shelf as well as NS altimeter track orientation. Moreover,, there is a higher bias inside GoM, clear indication that the local mean dynamic topography for these Globcurrent data is incorrect as seen in Fig.2.
in considering the added surface Ekman term, Globcurrent performance does not show large improvement, but this depends upon site location, relative strength of the wind-induced term, etc... Further work is required and can exploit the buoy currents in detailed analyses.
Fig. 8. Time series (70 day moving average applied) of alongshelf altimeter-based (GlobCurrent) current Vg across track 100 on the Scotian shelf and for track 141 along the Cabot Strait. Time lag and max correlation are both indicated. Positive lag indicates Vg141 leads. Negative alongshelf currents are downstream flow. The estimated advection time from Cabot Strait (track 141) to the southwest SS (track 100) is about 100 days. Result is consistent with that reported by recent regional modeling of Urrego-Blanco and Sheng (2014).
Fig.7. Time lagged cross-correlation between the alongshelf current VgXXX and Vg100 at the key Track 100 location (Fig.5). Positive time lag indicates VgXXX leads Vg100. The highest correlation is observed between Vg141 and Vg100, suggesting the along shelf current near the Cabot Strait in G. St. Lawrence is a key modulator of currents on the southern Scotian Shelf.
References
Feng, H., D. Vandemark, and J. Wilkin (2016), Gulf of Maine salinity variation and its correlation with upstream Scotian Shelf currents at seasonal and interannual time scales, J. Geophys. Res.Oceans, 121, 8585–8607.
Urrego-Blanco, J., and J. Sheng (2014), Study on subtidal circulation and variability in the Gulf of St. Lawrence, Scotian Shelf, and Gulf of Maine using a nested-grid shelf circulation model, Ocean Dyn., 64(3), 385–412.
Acknowledgements
Work is funded through the NASA Science Mission Directorate and NASA Ocean Surface Topography Science Team.
First conclusions related to use of Globcurrent for coastal advection studies in NWA
Preliminary results show similar correlation levels to that using coastally-processed data
New Globcurrent results indicate added insight into remote control of key regional dynamics
10th Coastal Altimetry Workshop, Florence, Italy, Feb, 2017 ct