Satellite altimetry dedicated to coastal ocean.

 

Satellite altimetry is limited near the coastlines due to land contamination in the altimetry and radiometric footprints (10 km and 50 km of footprint diameter, respectively) but also to inaccurate geophysical corrections. Despite this, the altimetric measurements are present and may contain useful information for coastal studies.
This data use case proposes to give some instructions on how to use the Coastal along-track Delayed Time Sea Level Anomaly (CoastalDT-SLA)products distributed by Aviso over the South African coasts, in the Agulhas current.

We will show that more altimetry data are recovered in coastal areas and that this high resolution products allow retrieving small scale oceanic structures, like the Natal Pulse signature. Later, some comparisons with independent data (SAR and in situ measurements: current meters, tide gauges data) will show the good correlation.

Data used

The Jason-2 Coastal along-track Delayed Time Sea Level Anomaly (CoastalDT-SLA) product over the Agulhas area is used here.
The CoastalDT-SLA are Level-3 products, conceived as an evolution of the Jason-2 Level-2 coastal products, commonly called PISTACH products. The Level-2 products rather address altimetry experts and are quite difficult to be used by non-expert users.
The CoastalDT-SLA products are easier-to-use and offer high resolution (20-Hz sampling rate) sea surface height on reference tracks. Differently from Level-2 products, the CoastalDT-SLA products are only computed over specific ocean areas (see details on the Aviso website).

Download CoastalDT-SLA data files from the Aviso FTP server for the Agulhas area. An authentication is required to access products. It can be obtained by filling in the form on Aviso and by selecting the CoastalDT-SLA product on the second page. Once connected to the server, files are in the sub-directories named /experimental/coastal_dt_sla_j2.

Ssalto/Duacs Along-Track Sea level anomalies are used here for comparison with the CoastalDT-SLA products. They are also distributed by Aviso and can be downloaded from the same authenticated ftp server with the same login/password (see details on the Aviso website and in the dedicated User Manual). Once connected to the server, files are in sub-directories named /global/dt/upd/sla/j2_cf and files are tipically named dt_upd_global_j2_sla_vxxc_20090805_20090812_20110329.nc.gz

Methodology

We use the Broadview Radar Altimetry Toolbox to observe the data and do some computation.

Data chosen

For the CoastalDT-SLA product, three Jason-2 passes are used in this data use case:

  • the Jason-2 interesting passes over the Agulhas area are descending passes: #020, #096, #172 (one file per pass). The descending track alignment is almost perpendicular to the Agulhas Current flow, this presents a good configuration for repeat-track analysis of the system with altimetry.
  • the time series used for this data use case streches out from the cycle_001 (July 2008) to the cycle_110 (July 2011) (all the cycles are included in a given file).

Ssalto/Duacs Along-track Sea level anomalies used for comparison with the CoastalDT-SLA products contain all the available passes during a 7-days period. The downloaded time series streches out from the cycle_011 (October 2008) to the cycle_110 (2011-07-08). Although BRAT can handle numerous files, it appeared preferable in terms of efficiency for BRAT to perform an extraction of the files based on a portion of the given track. We performed this extraction with the following script, based on the use of the “ncdump”, “ncgen” and “ncea” commands. The latter needs the installation of NCO (NetCDF operators) on the computer. The NCO homepage contains more information. Here, the extraction is done between -45°S and -33°S.

agulhas_area_tracks_sm

fig 1: Jason-2 ground tracks over the Agulhas area included in the CoastalDT-SLA product.

More altimetry data close to the coasts

Dataset

Once the relevant files are all downloaded, create a dedicated workspace in BRAT and then, create two datasets in the “Dataset” tab:

  • a first one, named “agulhas_coastal_096” includes the CoastalDT-SLA product with only one file: coastal_dt_sla_j2_agulhas_096.nc. It contains all the time serie between 2008-07-17 to 2011-07-07.
  • the second, named “agulhas_duacs_096” includes the 101 Duacs SLA files between 2008-10-18 to 2011-07-08.

Operations

On “Operations” tab, we create one operation for each dataset, CoastalDT-SLA and Duacs-SLA, in order to plot an hovmöller diagram, i.e a latitude-cycle diagram. These operations are respectively named “Op_agulhas_coastal_096” and “Op_agulhas_duacs_096”.

  • Operation “Op_agulhas_coastal_096”: in “Operations” tab, click on “New” button to define a new operation, give the choosen operation name in the “Operation name” box.
    In Data expression, drag and drop the variables “cycle” in “X”, the latitude in “Y” and the “SLA_Filtered_41pts_MLE4” in “Data”.
    The spatial resolution for this dataset is ~1/106° in latitude. But as the resolution is set to 1/3°, by default, in BRAT, we have to change it to reflect the native resolution of the CoastalDT-SLA product. In “Set Resolution/Filter” (still in the Operations menu on the bottom right), the “Step” in Y resolution sets to 1/99° (note that a such small spatial resolution as 1/106° doesn’t work).
  • Operation “Op_agulhas_duacs_096”: The previous Operation can be copied to create the operation corresponding to the “agulhas_duacs_096” dataset. To do that, select the previous operation (“Op_agulhas_coastal_096”) and click on the “Duplicate” button. Change the name to refer to the new operation name (“Op_agulhas_duacs_096”) and be aware to well select the corresponding Dataset with the corresponding Operation.
  • In Data expression, drag and drop the variables “cycle” in “X”, the latitude in “Y” and the “SLA” in “Data”. Since the file resolution is different, change the “Step” in Y resolution at 1/20°.

Once each Operation has been defined, click on the “Execute” button.

Views

Once the Operations finished, define one View per Operation in the “Views” tab:

    • View “Vi_agulhas_coastal_096”: Drag and drog the corresponding operation “Op_agulhas_coastal_096” from the list of operations in the “Available” box, on the left, to the “Selected” box on the right. Finally, click on “Execute”. For comparison between both dataset, the SLA limits are set to -0.9 and 0.9 m in the “Range” boxes.
    • View “Vi_agulhas_duacs_096”: Do the same for the operation “Op_agulhas_duacs_096”.
agulhas_JA2_tr096_MLE4_SLA_hov_sm agulhas_JA2_tr096_MLE4_SLA_hov_Duacs_vxxc_sm
fig 2: Latitude-time (cycle) diagram of Sea Level Anomalies, from -35°S up to the nearest point to the coast, for the Jason-2 pass #096. LEFT: CoastalDT-SLA product between August 2008 (cycle 1) and July 2011 (cycle 110). RIGHT: Ssalto/Duacs Along-track product between October 2008 (cycle 11) and July 2011 (cycle 110). The vertical white areas represent missing data.

The northern limit of the CoastalDT-SLA product reaches -33.32°S whereas it is -33.42°S for the Ssalto/Duacs Along-track product.
Thanks to CoastalDT-SLA product, a better along-track sampling is obtained and more data are recovered in coastal areas compared to the Ssalto/Duacs Along-track product, by roughly 11 km.

Signature of small structures

The aim is to plot the geostrophic currents with hovmoeller diagrams for several passes and retrieve the small oceanic structures near the coast.

Dataset

Create new datasets in the “Dataset” tab, one per pass:

  • The previous dataset defined for the CoastalDT-SLA product, pass #096 is also used here : “agulhas_coastal_096”. It includes the CoastalDT-SLA product with only one file: coastal_dt_sla_j2_agulhas_096.nc. It contains all the time serie between 2008-07-17 to 2011-07-07.
  • On the basis of this dataset, do the same for the following passes by creating the datasets “agulhas_coastal_020” and “agulhas_coastal_172”.

Operations

On “Operations” tab, we create a first operation for a given dataset, by filtering the variable SLA. Then, this operation is exported and integrated in a new dataset. A second operation will compute the geostrophic currents from the filtering SLA which will be plotted as an hovmoeller diagram, i.e a latitude-cycle diagram. These steps are detailed below for only one pass. They will be done again for all the passes in order to obtain one plot hovmoeller diagram for each pass.
  • Operation #1 “Op_agulhas_coastal_096_Filt”: in “Operations” tab, click on “New” button to define a new operation, give the choosen operation name in the “Operation name” box.
    In Data expression, drag and drop the variable “time” in “X”, nothing is reported in the optional “Y” field. Then drag and drop the following variables in “Data”: SLA_Filtered_41pts_MLE4, cycle, latitude, longitude.
    Then, a Lanczos filter is appplied to the SLA. Select the “SLA_Filtered_41pts_MLE4” in the Data expressions and click on the grey box “Expression”. Then click on the “Insert algorithm” button. A pop-up window opens. Define the Lanczos filter computation from the proposed algorithms: BratAlgoFilterLanczosAtp. Click on “OK” to close the pop-up window. A function-like expression is inserted in the Expression box. You have to check that the variable “Expr” matches the existing variables within your dataset (here, you have to drag and drop the “SLA_Filtered_41pts_MLE4” from “Fields” up to the “Expr” field). The Window Length, the Cut-Off and the minimum of valid points have to be defined. We respectively set 41,30 and 1. Finally, the filtered SLA computation is :
    exec(“BratAlgoFilterLanczosAtp”,SLA_Filtered_41pts_MLE4,41,30,1, 0)
  • Export: this “Op_agulhas_coastal_096_Filt” operation is exported by clicking on the “Export” button (upper right). A pop-up window opens. Set the “NetCDF format”, define the “Output file” name “Op_agulhas_coastal_096_Filt.nc” (don’t change the path). Click on “Execute operation before export” and on “Execute”. Once finished, the output file is available in the sub-directory “Operations” of your workspace.
  • In the “Dataset” tab, create a new dataset “agulhas_coastal_096_Filt” by retrieving the previous operation saved in the the sub-directory “Operations” of your workspace.
  • Operation #2: create a new operation named “Op_agulhas_coastal_096_UV” corresponding to the dataset “agulhas_coastal_096_Filt”. In Data expression, drag and drop the variable “cycle” in “X”, the “latitude” in “Y” and the “SLA_Filtered_41pts_MLE4” in “Data”. With a right-mouse click on Data, select ”Insert empty expression”, and rename this new expression with “U” (for East component). Click on the grey box “Expression” and then on the “Insert algorithm” button. A pop-up window opens. Choose the BratAlgoGeosVelAtp algorithm. A function-like expression is inserted in the Expression box. You have to check that the variable “Height” matches the existing variables within your dataset (here, you have to drag and drop the “SLA_Filtered_41pts_MLE4” from “Fields” up to the “Height” field). Finally, the expression for “U” is:
    exec(“BratAlgoGeosVelAtp”, latitude, longitude, SLA_Filtered_41pts_MLE4)
    The “Y” spatial resolution is set to 1/99° in latitude in “Set Resolution/Filter”.
    Do the same for the North component. Execute the operation “Op_agulhas_coastal_096_UV”.

Views

Once the Operation finished, define a View in the “Views” tab:

  • Create a new view and name it “Vi_agulhas_coastal_096_UV”. Then, drag and drog the U and V of the operation “Op_agulhas_coastal_096_UV” from the list in the “Available” box, under the Z=F(X,Y) section, located on the left, to the “Selected” box on the right. Click on the “Selected” U expression and below, click on the box “East component”. Define the Min/Max for “U” under “Display Expression Properties”: -1.5/1.5 (unit=m/s). Then click on the “Selected” V expression and below, click on the box “North component”. Define the same Min/Max for “V”. Finally, click on “Execute”.
  • The display window opens, and your hovmöller diagram is created.
agulhas_JA2_tr020_096_172_41pts_MLE4_UV_hov_sm
fig.3: Latitude-time (cycle) diagram of coastal filtered geostrophic velocity anomalies, perpendicular to the track, up to the nearest point to the coast, for the Jason-2 passes #020 (bottom), #096 (middle) and #172 (top), between August 2008 (cycle 1) and July 2011 (cycle 110). The vertical white areas represent missing data.

Results

The Agulhas current is a major western boundary current flowing in the southwestern Indian Ocean. Near the South African coasts, it follows the narrow continental shelf where is characterized by stable flow conditions. Nevertheless, some variability patterns occurs during the “Natal Pulses”, some solitary cold cyclonic meanders. These meanders are generated at the Natal Bight (off Durban), close to the coast, and move downstream along the Agulhas Current by going away from the coast.
Natal pulses can be identified with satellite altimetry as depressions in the sea surface topography (see also negative sea level anomalies on fig.2, left) and as negative flow anomalies in the geostrophic velocity anomalies, describing southwestward flow anomalies (fig.3). Evidence of such variability are several times plotted on the latitude-time diagrams, in the northern extremes of each figure: around -35.5°S for pass #020, around -34.5°S for pass #096 and around -33.3°S for pass #172. The Natal Pulse is well observed either on the western part of the area (on the pass #020) when it comes off the coasts, or in the eastern part (pass #172) when it flows closer to the coasts.

 

Validation

To assess the quality of the CoastalDT-SLA product, some comparisons with external data have been done (note that these figures have not been plotted with BRAT).

SAR

The small scale structures (50 km wavelength) observed with the CoastalDT-SLA product are confirmed by the SAR currents. The coastal anomalies retrieved with CoastalDT-SLA in the last 50 km are confirmed by SAR data.

agulhas_SAR

fig. 4: Comparison of velocity anomalies during cycle 65 (April 2010) between SAR projected velocities and CoastalDT-SLA product, Jason-2 pass #096 (note that the coast is on the left).
SAR velocities anomalies are plotted in dashed lines: the green one corresponds to the original data (total current) and the black one corresponds to the SAR minus the mean current (MDT CNES-CLS09_v1.1).
CoastalDT-SLA velocities anomalies are plotted in solid lines: the blue one corresponds to the MLE4 retracking, the green corresponds to RED3 and the red one corresponds to the OCE3 retracking. Credits CLS.

ADCP

The strong negative current anomaly (~ -34.5°S) occured during a Natal Pulse event is coherent in position and amplitude between the CoastalDT-SLA product and ADCP in situmeasurements. The same Natal Pulse event is also observed with the SAR currents the same day (see fig.4).

agulhas_ADCP

fig. 5: Comparison of velocity anomalies between ADCP projected velocities (middle) and CoastalDT-SLA product, Jason-2 pass #096 (top and bottom) on April 2010 (note that the coast is on the right).
Cyan lines correspond to the mean current (over the 3 figures).
Green lines (top and bottom) correspond to the MLE4 retracking, the blue lines correspond to RED3 and the red lines correspond to the OCE3 retracking.
Blue line (middle) corresponds to ADCP in situ measurements.
Credits Noveltis, ADCP data provided by ACT oceanographic cruise (RSMAS Univ. of Miami).

Tide gauge

By approaching the coast, the CoastalDT-SLA product provides additional data which are consistent with tide gauge observations (correlation >0.85)

agulhas_TideGauge_sm

fig. 6: Correlation between the altimeter data (Jason-2 pass #248) and tide gauge measurements done in Durban (see details on the tide gauge identification, on the Aviso website).
Blue line corresponds to Ssalto/Duacs Along-Track SLA vfec altimeter data, red line correspond to vxxc, green line correspond to MLE4 retracking in CoastalDT-SLA product, red line correspond to RED3retracking in CoastalDT-SLA product and brown line correspond to OCE3 retracking in CoastalDT-SLA product. Credits CLS.

Further information:

  • Other Data Use Case using CoastalDT-SLA products: The Florida Keys by using coastal dedicated products
  • User Handbook: CoastalDT-SLA User Manual
  • S.Labroue et al.: Level-3 PISTACH Products for Coastal Studies (pdf, Presentation at the San Diego Coastalt meeting, 2011)
  • C.Dufau et al.: The PISTACH product: altimeter data for coastal ocean, Presentation at the 2012 EGU meeting (EGU2012-12649).
  • J.Tournadre et al.: Agulhas Current Estimation from High Resolution Altimetry (pdf, Presentation at the San Diego Coastalt meeting, 2011)
  • F.Collard et al.: Surface Current in the Coastal Region of the Agulhas Current. (pdf, Presentation at the San Diego Coastalt meeting, 2011)