Version 69 (modified by steiger5na, 4 years ago) (diff)


  2. 0 - Publications, reports from the project
  3. 1 - Objectives
  4. 2 - Experimental setup:
    1. 2.1 General description
    2. 2.2 Topography
      1. 2.2.1 Topography for the shelf break experiments
      2. 2.2.2 Topography for the ice shelf experiments
    3. 2.3 Reference axis
      1. 2.3.1 Reference axis for shelf break experiments
    4. 2.3.2 Reference axis for ice front experiments
    5. 2.4 References axis along the wall (horizontal and vertical) - Nadine …
    6. 2.5 Fixed Parameters
    7. 2.5 Variable Parameters
    8. 2.6 Additional Parameters
    9. 2.7 Definition of the relevant non-dimensional numbers (UPDATE WITH …
  5. 3 - Instrumentation and data acquisition
    1. 3.1 Instruments
    2. 3.2 Definition of time origin and instrument synchronisation
    3. 3.3 Requested final output and statistics
  6. 4 - Methods of calibration and data Processing
    1. 4.1 Calibration for shelf break experiments
      1. Probes calibration
    2. 4.2 Calibration for ice shelf experiments
  7. 5 - Good to know…
    1. 5.1 How to turn on and off the lasers
    2. 5.2 How to connect to the cluster
    3. 5.3 How to connect to the server from home
    4. 5.4 How to process the data
  8. 6 - Organization of data files
  9. 7 - Diary:
    1. 7.1 Thursday 07 September
    2. 7.2 Friday 08 September
    3. 7.3 Monday 11 September
    4. 7.4 Tuesday 12 September
      1. Test experiment EXP01
      2. Experiment EXP02
    5. 7.5 Wednesday 13 September
      1. Experiment EXP03
      2. Experiment EXP04
      3. Experiment EXP05
      4. Experiment EXP06
      5. Experiment EXP07
    6. 7.6 Thursday 14 September
      1. Experiment EXP08
      2. Experiment EXP09
      3. Experiment EXP10
      4. Experiment EXP11
      5. Experiment EXP12
    7. 7.7 Friday 15 September
      1. Experiment EXP13
      2. Experiment EXP14
      3. Experiment EXP15
      4. Experiment EXP16
      5. Experiment EXP17
    8. 7.8 Monday 18 September
    9. 7.9. Tuesday 19 September
      1. Experiment EXP18
      2. Experiment EXP19
      3. Experiment EXP20
      4. Experiment EXP21
      5. Experiment EXP22
    10. 7.10 Wednesday 20 September
      1. Experiment EXP23
      2. Experiment EXP24
      3. Experiment EXP25
      4. Experiment EXP26
      5. Experiment EXP27
      6. Experiment EXP28
    11. 7.11 Thursday 21 September
    12. 7.12 Friday 22 September
      1. Experiment EXP29
      2. Experiment EXP30
      3. Experiment EXP31
      4. Experiment EXP32
      5. Experiment EXP33
    13. 7.13 Monday 25 September
      1. Experiment EXP34
      2. Experiment EXP35


InfrastructureCNRS_Coriolis Topographic Barriers and warm Ocean currents controlling Antarctic ice shelf melt
Project (long title)Topographic Barriers and warm Ocean currents controlling Antarctic ice shelf melt
Campaign Title (name data folder)17ICESHELF
Lead AuthorElin Darelius (part I) and Anna Wåhlin (part II)
ContributorsNadine Steiger, Lucie Vignes, Joel Sommeria, Samuel Viboud
Date Campaign Start04/09/2017
Date Campaign End27/10/2017

0 - Publications, reports from the project

1 - Objectives

The warm water threatening to melt the Antarctic ice shelves originates from the deep ocean north of the continental shelf. In order for the warm water to reach the ice shelf cavities it has to pass to topographic barriers: the shelf break and the ice shelf front. In a series of experiments we will explore the role of topography in controlling the onshelf flow of warm water.

2 - Experimental setup:

2.1 General description

The two Barriers - the shelf break and the ice shelf front - will be studied in two separate sets of Experiments.

Part I: Divergent isobaths at the shelf break

An idealized topography representing a widening continental shelf and a trough crosscutting the Continental shelf break is used and the effect of changing 1) water depth 2) radius of curvature and 3) flow speed will be explored. The experiments will be repeated with a) a barotropic and b) a baroclinic current.

Part II: Flow across an ice shelf front

2.2 Topography

2.2.1 Topography for the shelf break experiments

Sketch of the topography seen from above with the distances to the (0, 0) coordinate. The blue circle and the blue line are the tank wall and the centerline, respectively.

Sketch of the topography seen from the side with the heights of the different areas.

From experiment 18 the source was moved 2.00m further back! The position of the source relative to the topography stayed as before.

2.2.2 Topography for the ice shelf experiments

2.3 Reference axis

2.3.1 Reference axis for shelf break experiments

By definition we will use Ox and Oy axis to define the along slope and the cross slope axis. The central reference point (0,0) along the slope is chosen to be the first "land" corner downstream of the source. Positive u - direction corresponds to the along slope flow direction, while positive v - direction is directed onshelf. The position of the topography relative to this coordinate system is shown in 2.2.1

2.3.2 Reference axis for ice front experiments

By definition we will use Ox and Oy axis to define the along channel and cross channel axis. The central reference point (0,0) is chosen to be at the front of the ice shelf and in the middle of the channel. Positive u - direction corresponds to the flow direction towards the ice shelf and positive v - direction is directed to the "west".

2.4 References axis along the wall (horizontal and vertical) - Nadine add image!

By definition we will use Ox and Oy axis to define the along shore and the cross shore axis. The central reference point (0,0) along the wall is chosen to be the closest point to the center of the tank (also labeled M0?). Positive direction corresponds to the mean wave or the mean flow direction.

We use seven references points along the wall to quantify the impact of the free surface deformation and the possible vertical deviation of the laser sheet???

All these points will be measured every day before starting the Experiment??

H_coast (cm)''
H_laser_up (cm)''
H_laser_down (cm)'

2.5 Fixed Parameters

$T_{rotation}$Rotation period$50 \ s-1$
$H_{shelf}$Shelf height$0.5\ m$
$H_{through}$Trough height$0.42\ m$
$W_{trough-slope}$Width of slope in troughm
$L$Total length of the wall$ \ m$
$W_{Source}$Width of source (inner)$ 23\ cm$

2.5 Variable Parameters

NotationDefinitionUnitInitial Estimated ValuesRemarks
$H_{water}$Total water depth$cm$60 - 70At the point where waterlevel is not influenced by rotation: r=?? m.
$H_{shelf}$Depth on shelf$cm$10 - 20estimated where?
$R_c$Radius of curvature$m$0 - 0.5
$Q$Flux$L min{-1}$20 - ??- ?? - 135
$\Delta \rho$Density difference (ambient - inflow)$kg$0 - 3 - 10

2.6 Additional Parameters

NotationDefinitionUnitInitial Estimated Values
$g'$Reduced gravity$m s{-2}$
$Rd$Baroclinic deformation radius$cm$

2.7 Definition of the relevant non-dimensional numbers (UPDATE WITH OUR NOTATION!!)

Rossby number, $ Ro = U/W/f$.

Relative step size, $H* = \delta H/ (H-\delta H) $. (Cenedese et al, 2005)

Depth of baroclinic current ($H_{BC}=sqrt(2Qf/g')$ (Chapman & Lenz 1994; Sutherland et al, 2009)

Internal rossby radius ($\lambda_i = sqrt (g' H_{BC})/f$.

3 - Instrumentation and data acquisition

3.1 Instruments

Conductivity Sondes (CS)

Particle Imaging Velocimetry (PIV) A Spectra-Physics Millennia ProS 6W YAG continuous laser (532 nm) in conjunction with 2 cameras was used to provide PIV images. The laser light sheet was brought in parallel to the bottom of the tank in case of the slope front experiments and tilted by 2% to match the slope of the channel in case of the ice front experiments. The light sheet can then be racked in the vertical through a series of steps through the use of a motorized traverse and a mirror set at 45 degrees. The laser has another set of optics to point the light sheet down at the mirror, producing the light sheet. The laser light sheet positions are then synchronized with the PIV cameras. The laser light sheets cover the whole topography, but are slighlty bended towards the sides, so that they are closer to the bottom at the source and at the end of the topography compared to the middle. Also a vertical laser was used together with a vertical camera to observe the flow in a cross section. The three PIV cameras consist of:

  • one Falcon1 camera (Falcon 4M, CMOS 2432*1728 pixels, 10 bits) as the vertical camera – with a 35 mm objective lens.
  • PCO1 (PCO.edge5.5 CMOS cameras (2560*2160 pixels)) with a 35 mm objective lens overlooking the part in front of the source in the slope front experiments and the ice shelf in the ice front experiments.
  • PCO2 (PCO.edge5.5 CMOS cameras (2560*2160 pixels)) with 20 mm objective lens overlooking the continental self and trough in the slope front experiments and the channel in the ice front experiments.

After experiment 26 of the slope front expriments, 60 micron particles were used for the flow seeding. The number of slices were adjusted to the need of the experiment and is listed in the file LIST_OF_EXPERIMENTS.xlxs. The number of slices, dt between the images, exposure time (either 20-50), the number of images and the number of scans had to be decided before each experiments. The slope front experiments also contained series of images at one horizontal slice to better observe the evolution of the flow. The vertical laser was turned on after a steady flow was established/ at the end of the experiment.

During horizontal PIV, the vertical camera was turned on (with the same acquisition as the PCO1 and PCO2) to produce an .xml file that contains information on the time, dt, exposure time, times for the scanning. During some experiments, this .xml file was missing (technical mistake or if we stopped the acquisition before it was done), so that .xml files from other experiments have to be used and modified to fit the setup. However, the starting time is not correct then.

3.2 Definition of time origin and instrument synchronisation

3.3 Requested final output and statistics

Batch processed camera data in to .png files for those experiments from 18-43 that have PIV data, so that images are in a non-proprietary format. PIV analysis of the flow field through multiple horizontal slices in different Z-positions, for the non-rotating case, and for the rotating cases (experiments 18-43 as above), dependent on the quality of the captured PIV images. Average velocity vectors for the channel slices. Potentially information on vorticity would enable the smaller-scale vortical structures that are obvious in some of the videos, to be identified.

4 - Methods of calibration and data Processing

The PIV data will be processed using the MATLAB tool UVMAT available on

All images from the experiments will be stored in the directory /fsnet/projects/coriolis/2017/17ICESHELF/DATA under the name for the experiment.

4.1 Calibration for shelf break experiments

The images for PIV (PC01, PC02) are calibrated from images of a horizontal grid, in 3D with an inclined grid (tilted angles) and in the vertical (NAME OF CAMERA!) for the camera fixed on the tank wall.
The calibrated images with the grids are stored in 0_REF_FILES with the
- 2D calibration in the directory CALIB_07_09
- 3D calibration in the directory CALIB_07_09_3D
- Vertical calibration in the directory CALIB_VERTICAL_08-09

The images are divided into the two different cameras as PCO1/ and PCO2/.

This sketch shows the topography together with the view of the two cameras.

The 3D calibration is done with the geometric calibration function on UVMAT to create 'intrinsic parameters' for each camera that will be used for all images. Calibration points along the inclined grid are marked and translated from image coordinates to physical coordinates defined in section 2.3.1 to later take into account the different height of slices. See for details of the method. The calibration parameters are copied in a xml file beside the folders of each camera (for instance PCO2.xml for PCO2/). The xml files also containing all the timing information and have to be copied next to the folders of all images, to use the same intrinsic calibration data for each image taken with the corresponding camera.

The 2D calibration is used as a reference horizontal grid with a grid height of
- z = 80.8 cm for PCO1
- z = 58.2 cm for PCO2

Calibration of camera 1 (PCO1) on the left and of camera 2 (PCO2) on the right, with the calibration grid produced with PIV in purple.

The calibration is checked with a calibration ruler (images saved in CALIB_RULER) to see that the transition between the two cameras is smooth.

For the vertical calibration the images are saved in a different file format with extension '.seq'. The format can be changed in the UVMAT software, but the folder with the vertical images should not be renamed (e.g. '2017-09-08T16.04.32'). To change the image format, choose RUN -> Field Series. Chose the right input file, an select 'extract_rdvision.m', click INPUT and RUN. The image will be saved in a new folder as a .png file and is ready for use in UVMAT.

===Laser sheet calibration=== ==Slope front experiments==

Laser setting$H_{real}
65055 cm
70049.5 cm
80040 cm
85035 cm
90030 cm

==Ice front experiments==

Probes calibration

For the calibration of the probes 3 measurements were done before the experiments on the 5 probes and 4 measurements were done after. For each C you can see in the table the voltage measured. Done the 21/09/2017

$\rho$=998.9 T=20.3$\rho$=1004.5 T=20.7$\rho$=1015.1 T=21.0

Done the 29/9

$\rho$=998.2 T=22.8$\rho$=1004.3 T=22.3$\rho$=1009.5 T=22.4$\rho$ = 1014.9 T=22.4

4.2 Calibration for ice shelf experiments

5 - Good to know…

5.1 How to turn on and off the lasers

The laser for horizontal slices is turned on and off in the rotating office via the computer. You only have to switch the shutter between on and off.

The laser for the vertical slices is turned on and off outside the wall, where the laser is located. Press the button in the lower right corner of the instrument in the photo.

5.2 How to connect to the cluster

On linux : ssh -CX servcalcul7a

Then install uvmat (see the legi software page)

oarsub -I -l walltime=10:0:0 (to change your access time to 10h instead of 2)

Now you can open matlab

In uvmat>Run>field series>run mode change to "cluster oar"

5.3 How to connect to the server from home

On Linux:

  • Create a tunnel on command line: ssh -L !2222:servcalcul7a:22 username@…
    - Install X2GoClient (Remote desktop client) and start it
    - In upper left corner, you can create a new session. The setup for the server is: Host: localhost; Login: username; SSH port: 2222
    - Once the session is created, you can click on it, enter you password and then you are working on one of the LEGI computers.

On Windows: As Elin!

5.4 How to process the data

All experiments are saved in fsnet>project>coriolis>2017>17ICESHELF>DATAZ>EXPXX. DATAZ corresponds to DATA for the slope front experiments and to DATA2 for the ice shelf experiments. The file EXPXX is for the horizontal laser sheet, EXPXX_SCAN is for the horizontal scanning and EXPXX_VERT for the vertical sheet. Within each EXPXX and EXPXX_SCAN you will get three files : FALCON, PC01 and PC02. These are the three cameras we are using during the experiments, the FALCON is the camera for the vertical sheet and will only be used to get the time parameters. The data from the PC01 and PC02 need to be treated in order to perform the PIV.

  1. Copy the files from the computer of the cameras.
    - The images from the vertical camera are saved on the local computer in G: and have to be copied into the folder FALCON in the corresponding experiment.
    - The images from PCO1 (camera at the inflow) are saved on D: of the remote desktop connection that ends with 151. The images have the be copied into the folder PCO1 of the corresponding experiment.
    - The images from PCO2 (camera above the trough) are saved on D: of the remote desktop connection that ends with 150. The images have to be copied into the folder PCO2 of the corresponding experiment.
  2. Get the .xml file with timing information
    In uvmat select the Run>field series then open>browse and select an experiment. Go in the folder within the FALCON and select the .seq file. Choose extract_rdvision_times in the Action part, click on Input and Run. This will create a Falcon.xml within the FALCON folder. Check that the Falcon.xml file contains the right information on Dti (time between two images), Exposure time and Number of images (NbDti)!
  3. Convert .tif into .png
    In Run>Field series you can open any .tif of the desired folder (Warning : you should not take the im.tif but any im@… is working). Then in the Action part you should select extract_multitif, if the Input button is in pink you should click on it and give to the software the Falcon.xml that you extracted earlier (If that doesn't exist, copy one .xml file from a previous experiment and change the setting manually. Once it is done you can click on RUN, don't forget to select cluster_oar in the run mode, the action will be much faster. When it is done a new folder called PCOY.png is created in your EXPXX folder and contains all the png.
    For SCANS: Check that the right number of slices is registered! To do so, click on check_data_files in the Action part and type the right number of slices in nb_slices i
  4. Remove background
    Browse for the right experiment again and chose the extracted .png images. The action sub_background that will remove all the motionless structures. For the input you have to make sure that the image rescaling coefficient is equal to 2 and then make is run. (This process takes a long time, don't forget to make it run on the cluster). You can check on Status, how many images are done. The output of this action is a file PCOY.png.sback filed with png images.
    For SCANS: Here, the image rescaling coefficient should be equal to 0! Also, set the Number of images for sliding background to the number of images per slice!
  5. Do PIV
    Browse for the images with the removed background (PCOY.png.sback.). Choose the Action civ_series. For in Input, you have to import an existing .xml file to get the right parameters. You can use one .xml file from another experiment on which the piv is already done. Browse for: EXPXX/PCOY.png.sback.civ/0_XML and choose any of the img_#.xml. Make sure in the Input parameters that you chose the correct mask (right camera and with/without corner).
  6. Add calibration grid
    To get all images in cm relative to our coordinate system, the calibration we did in the beginning has to be applied on all experiments. This can be done in the very end.

Important to do during PIV processing:

  • Check that the Falcon.xml file contains the right information on Dti, Exposure Time and NbDti!
  • Check that all files are processed before you continue!
  • If there is a message about the wrong number of i before starting the civ_series, then the NbDti? in the Falcon.xml is wrong and you have to change it!
  • Before pressing the RUN button always double check that it is actually analysing all images (the values in first and last in the middle section).
  • If you are analysing scans and slices, always check that you have the right number of slices.

6 - Organization of data files

All data related to the project are in:
SERVAUTH4\share\project\coriolis\2017 or\share\project\coriolis\2017

  • 0_DOC: miscellaneous documentation and reports
  • 0_MATLAB_FCT: specific matlab functions
  • 0_PHOTOS: photos of set-up
  • 0_PIV
    • Each ‘PIV’ folder contains subfolders for each of the 3 PIV cameras: Dalsa (sometimes Falcon1 – it’s the same thing); PCO1; PCO2 [these are named after the different brands of camera]. Other folders include PCO1.png and PCO2.png which contain processes images of the PCO cameras that are in a non-bespoke format. Other folders that can be within the Camera folder include: Dalsa.sback; Dalsa.sback_1; PCO2.png.civ; PCO2.png.civ_1; PCO2.png.civ_2; PCO2.png.sback: PCO2.png.sback_1; PCO3.png.sback_1. .sback files refer to those files where the background has been subtracted, then civ_1 contains images with the first PIV iteration as processed in UVMAT (Joel’s code) and shows the raw data – with or without the rejected vectors; vectors are shown in four colours, blue = best, green = medium, red = poor, and pink = false. A box can be clicked to hide the false vectors. Civ_2 uses a spline interpretation to interpolate between vectors, so long as they are close enough to the surrounding vectors. Then interpolates all the vectors onto a regular grid. Times for the .png images are in the XML files, or netcdf files.
  • 0_Processing: UVP processing scripts in Matlab
  • 0_REF_FILES: files of general use (calibration data, grids ...)
  • EXP1, EXP2, folder for each experiment with names given in the table below. The names refer to ‘fix’ for non-rotating fixed case, ‘rot’ for rotating case, ‘str1’ for the first straight position (also called position X1), and ‘apex 2’ , for the apex in bend 2 (also referred to as position X4).
    • Within each experiments, there is a folder with PIV imagery called ‘Camera’, one for ADV data – ‘ADV’, one for UVP data – ‘UVP’, and one for the data coming directly off of the Coriolis table control system ‘LABVIEW’. Some experiments also contain an ‘Images’ folder or a ‘Gopro folder’ containing Gopro videos.
      • Each ‘Camera’ subfolder contains subfolders for each of the 3 PIV cameras: Dalsa (sometimes Falcon1 – it’s the same thing); PCO2; PCO3 [these are named after the different brands of camera]. Other folders include PCO2.png and PCO3.png which contain processes images of the PCO cameras, that are in a non-bespoke format. Other folders that can be within the Camera folder include: Dalsa.sback; Dalsa.sback_1; PCO2.png.civ; PCO2.png.civ_1; PCO2.png.civ_2; PCO2.png.sback: PCO2.png.sback_1; PCO3.png.sback_1
      • Each ‘ADV’ subfolder, contains two sub-folders: ‘nkt_files’ containing raw Nortek files, and the ‘mat_files’ are the exported raw data in Matlab format.
      • Each ‘UVP’ subfolder contains two folders – one with the experiment name (which is the downstream velocity data) recorded downstream of the velocity inflection downstream of bend apex 2, and one with experiment name ‘_cross’ which contains the cross-stream UVP data recorded at bend apex ‘2. These two folders contain text files for each of the probes. The convention is that Probe 1 is the basal probe, with each subsequent probe being successively higher. There are also .mfprof files which are the raw UVP data in native format. All probes are also integrated into single Matlab files. Lastly, there is a Logfile with the header file for the UVP detailing all of the parameters used in the run.
      • Each ‘LABVIEW’ subfolder contains: 1) a .lvm file which is a text file and contains a time-stamp, two voltages for the Conductivity probe on the traverse (C0 – Conductivity, and T0 – temperature [this latter one doesn’t work]), a Trig_cam heading representing the Trigger for the PIV Cameras, Conductivity probe in the input box (C1 and T1), and C2 (this was conductivity for a second probe in the input box which was briefly used before breaking. There is always a record for this but it is just background noise. 2) _position.lvm file which is an XYZ file with a times for the movement of the traverse. 3) Some folders also contain files. These are netcdf files and contain the vector data from the processed PIV images.

7 - Diary:

7.1 Thursday 07 September

The cameras still have to be adjusted. After that is done, we take photos with the calibration grid in the horizontal and in 3D with both cameras that are fixed on the roof. With PIV, we conduct the calibration of both cameras as described in 4.1.

7.2 Friday 08 September

To verify the calibration, we took photos with both cameras with one ruler to see if the calibrations of the cameras agree (which it does!). Then, we added a wall into the tank to prevent the water of the freshwater experiments to return to the source. The laser height was adjusted.

7.3 Monday 11 September

The table starts rotating for the first time with a velocity of 50 rounds/ sec. It gets slowly filled with water to a level of 60 cm. This takes about 3h.

After filling the tank Thoma measured the water level to be point where water level does not change with rotation 60.4 cm Center of tank – 42.5 + 16 cm = 58.5 cm Tank wall (6.5 m) – 62.4 cm. Water temperature 23.3C.

Particle sedimentation test : Water was taken from the tank, mixed with particles (30 + wetting agent) and set to settle in a glass “vase” with a narrow (ca 3cm) and tall (ca 10 cm) neck. The upper level with particles were observed as follows:

13.054 mm
13.305 mm

At 14h40 the border was now longer visible. There were still particles in the upper part of the flask, but a lot less than there were intitally.

Calibration of vertical laser sheet

While changing the settings on the laser, the actual height of the laser sheet was measured directly in the tank.

Laser setting$H_{real}
65055 cm
70049.5 cm
80040 cm
85035 cm
90030 cm

We will use 650 :50: 900 to get Slices every 5 cm from 30 cm to 55 cm (total 6 Slices).

TEST experiments

The desired flow rate of 50 L/min should according to 0_DOCS\flow_rate.xclx should be achieved With a diaphragme of diameter 12.9 mm. We used 13 mm. The flow rate was estimated by measuring the time it took for the water Level in the feeding tanks to descend from 170 to 150 (with the pumps turned off) - it took 41s giving a flowrate of 40 L / 41 s = 58.5 L/min.

The pipe was first opened to remove boubles from the system. The Sources was designed in a way that the hose stopped just below the top of the Source which was out of the pater, thus creating a small waterfall, turbulence and a lot of boubles that got throught the honeycomb and flowed along the slope With the current. The Source will be redisgned so that the water is inserted below the water surface.

A second trial (still with the original source) was made later (when Samuel was back). The same ddiaphragme was used and the same amount of boubles was created. Flourescent dye (NAME) was injected first in the feeding tank above, then in directly in the water just upstream of the vertical laser sheet.

Time series of photos was obtained from the vertical camera.

Aperture 150ms Time between photos 450 ms.

There were more particles in the inflow than in the ambient, and the border between the inflow and the ambient was seen to fluctuate up and down the slope. The timescale of this motion will be estimate from the photos tomorrow.

When turning to the horizontal cameras it was decided that sloping topography gave too many uncontrolled reflections that made it unsafe to increase the intensity. To remove these,parts of the bathymetry has to be Paint in black. The tank was set to empty during the night, while recycling the water.

7.4 Tuesday 12 September

The bathymetry has been painted black, a pipe has been introduced into the source and the drain has been lowered to that it can be used also for 60 cm water depth.

The tank is filling: $T_{water}=22.9C, \rho=998.5 kg m^{-3}, T_{air} = 23C$.

The reflection is now ok and we can start our experiments!

Test experiment EXP01

We conducted experiments with a long HS from the beginning after turning on the influx (EXP_01). The influx was 41 L/min with a diaphragma of 12.6 mm diameter. The influx water had a density of 998.2 and was enriched with particles of size 30mikrom. The laser sheet was at 11cm over the trough, 4 cm over the shelf, so at 54 cm when the laser setting was at 650. We decreased it to 640 to have the laser sheet at 55 cm. Water Level (supervision - this changes witha few mm when People come enter/leave the rotating Office) was at 58.4 cm. The exposure time of the cameras was at 50ms and the time interval 500 ms.

We switched to the VS at 14:58 (EXP01_VERT).

terwards we started the Scan at 15:15 (EXP01_SCAN) with the same camera settings as for the HS. The slices were at the heights of 55 cm to 30 cm in 5 cm steps. We took 4 images at each height with a time interval of 500 ms.

Observations: On the horizontal scans we can clearly see turbulent/vertical motion along the coast. They turn left at the curvature and follow along the topography to the wall. Some particles turn right above the trough/ shelf and flow out to the slope again.

Data check: We checked these first results with the PIV and decided to decrease the exposure time to get sharper images of the particles. With PCO2, the flow around the curvature can be captured with the PIV, but on PCO1 the PIV is not very good, because the flow is more turbulent and more 3 dimensional. We decide to decrease the time between the photos. Later we realized that the cameras only took every second photo, so with a time interval of 1000ms - > useless PIV.

Experiment EXP02

We repeated the HS with a reduced time to 200 ms between the photos and and exposure time of 30ms. We realized afterwards that the cameras only took every second photo, so with a time interval of 400ms!

7.5 Wednesday 13 September

The cameras are fixed now and taking every photo and the particles are spread out in the ambient water.

Experiment EXP03

The setup is still the same as for the previous experiments with 200ms time difference between the photos (this time working!) and particles in the ambient water. This time, we measured the flow rater AFTER the experiment and it was $Q = 50 l/min$ (measurements repeated twice), as expected. There are still bubbles coming out of the source!

The current reached the first corner after about 1 minute, and the trough corner after about 3 minutes.

Experiment EXP04

We started a new experiment at !10:31 with a reduced flow rate of about $Q = 20 l/min$ (diaphrama diameter of 6.8 mm). The first HS is run for 15 min n(!10:31- !10:46). The flow field was still evolving so we extended the five more minutes of hs (!10:51-10:56 this data are stored in EXP04_B) before we started the scan ( !10:59).

Note: During the scan, the laser is shaking too much! This needs to be changed.

Observations: The bubble at the source disappeared due to the lower flow rate. The current follows along the slope and turns left at the 2nd curvature into the trough. After the 15 min after the flow was turned on, the flow bifurcates at the continental coast and at the trough and fills up the trough. The return flow does not follow along the upstream slope of the trough, but occurs in the middle of the trough. The main current along the slope is getting wider with time.

Experiment EXP05

The vertical velocity of the laser was decreased and the timestep between the first and the second image at each level increased to 1s (first image will be discarded).

The tank from where the inflow is pumped was mistakenly filled with too much Cold water (T=19.2C) - i.e. about 4 c colder than the water in the tank.

There where a lot of Bubbles coming out from the Source - and we observed large (10 cm) airbubbles travelling Down the hose during the Experiment.

All the water turned at the first water, onto the shelf.

The current was at the first corner about 1 minute after the valve was opened.

Experiment EXP06

The squared cornes were inserted - the land corner was fastened with tape and the submerged corner with a screw..

To reduce to amount of Bubbles coming out from the Source we added a thin sheet of foam behind the honey comb. In addition, the joints of the tube where the diaphragm is inserted was greased. This appeared to help - the amound of Bubbles was greatly reduced and no Bubbles was seen travelling Down the hose.

The flow separated at the first corner and turned at the second corner.

The vertical scann was started 15 min after the valve was opened.'

Experiment EXP07

An eddy initially forms behind the first corner, but it then disappears.

The flow turns into the through. Large eddy on the Continental shelf (as in Kjerstis model Experiments!).

When scanning dt was set to 100 ms, but only every second photo was aquired so dt= 2 x 100.

7.6 Thursday 14 September

Experiment EXP08

The squared corners are still on, for the first 3 minutes of the horizontal slice the light in the entrance was turned on.

The flow turned left at the second corner following the slope in the depression keeping a "tight" shape on the topography.

After a while, a straight flow established, following the slope. The circulation is very slow compared to yesterday.

For the scan there is dt =1000ms between the first and the second image at each level, then dt=400ms. First image is not used to PIV

This is to avoid shaking of laser.

Experiment EXP09

The squared corners are removed and the flow rate is reduced. There were a lot of particles in the inflow in the beginning.

There is a clear split of the flow at the 2nd corner, the main part crosses the depression following the slope and a much smaller part followed the topography inside the depression.

The laser sheets were started 26 minutes after the valve was opened.

Experiment EXP10

The flow rate is increased, there are again a lot of bubbles.

Time to reach the first corner: less than 1minute.

After the reaching the first corner a major part of the flow followed the boundary and circulated on the shelf.

A smaller part continued to the corner of the slope and the depression (after 3minutes) and entered the depression, circulated along the topography and escaped by the western side of the depression

Experiment EXP11

The water level is increased by 10 centimeters to 68.5 cm. With the higher water level, we need to calculate new flow rates with the ration of the outflow area before and after the higher water level. Now, we are aiming to a flow rate of 33 l/min, which corresponds to the 20 l/min for the lower water level.

Conversion of flow rate from low to high water level:

Q (low water level)Q (high water level)

Note: In the previous experiment, the measured flow rate was always higher than the expected one with their diaphragme - flowrate curve. We plotted the diaphrame vs. flowrate again with the values from our experiments. For the flowrate of Q = 33 l/min, we therefore used a diaphragme with diameter of 10 mm, instead of 8.6 mm (which it was for the 20 l/min flux before).
We measured the flux 4 times during this experiment. It gave:
- 32.4 l/min
- 34.5 l/min
- 33.3 l/min
- 34.9 l/min

Measured water level after the experiment: 72.7cm at the edge and 68.5cm over the topography.

Time to reach the first corner: 3 min.

Observations: There are no more bubbles. The flow crossed the shelfbreak between the 1st and the 2nd corner. After the flow is established it follows along the rough topography and returns along the continental coast.

There was no straight flow along the slope.

There were very few particles in the end of the horizontal slice. On the vertical slice the flow does not look barotropic.

Note: After Exp11 they realized that there was water in the laser, they had to go and move it up.

Experiment EXP12

This experiment was the same as EXP11 with a water level of 68.5 cm but with the corner. The expected flowrate was 33 l/min and the measured flowrate was 34.6 l/min.

The exposure time was increased to 40 ms, because the flow is slow - so no need to have a short exposure time. It makes the particles more visible.

7.7 Friday 15 September

Experiment EXP13

During night, we increased the rotation rate to get a rotation period of 30s! We kept the corners.

There were still many particles in the ambient water, but Thomas sprayed more particles on the surface.

New measured water levels:
- wall: 69.8 cm
- center: 59.5 cm
- invariant: 65 cm

Water level on the screen: 63.4 cm

For the desired fluxrate of 50 l/min, we used the diaphragme with 12.6 mm diameter. The measured fluxrate after the experiment was 54.5 l/min.

Observations: In the beginning of the experiment, there were too many particles in the influx, but got less with time. The main current turned left at the second corner into the trough.

Note: The source is leaking, which created bubbles and a little "waterfall" due the the water surface difference inside and outside the source. Thomas fixed it by putting a wooden block in front of the leakage.

Experiment EXP14

We increase the influx to a desired value of 80 l/min with still the higher water level and increased rotation rate. Due to the leak there are water bubbles at the surface of the water at certain places.

Due to the high flux rate, we use an exposure time of 30 ms and a dt of 100ms.

Notes: The water level increased constantly. After this experiment, we had to adjust the sink to let out more water. It takes some time, so we had to wait until the water level was good again.

Water level on the screen : 63.4cm. The desired flux rate was 80L/min, we used the 16mm diaphragm and measured a rate of 94.7L/min.

Observations : The flow took 20s to reach the first corner, nothing crosses the shelf, a circulation is created on the shelf by water turning on the 1st or 2nd corner.

After a while some part of the current started following the slope after the 2nd corner.

Experiment EXP15

We wait for the water level to decrease (for about 2 hours) and reach 62.7cm on screen.

The desired flow rate is 50L/min and with the diaphragm of 12.6mm we measured a flow rate of 52.2L/min.

Observations : The flow reaches the first corner very quickly, then turned to the left and circulated on the shelf quite fast. Then a second branch started circulating inside the depression following the bathymetry.

After that a third smaller branch started to follow the slope.

Note : After this experiment, inertial oscillations appeared and the water started to move so we had to wait until further experiments.

Experiment EXP16

When we started this experiment the inertial oscillations were reduced but not suppressed.

The desired flow was 35L/min, with the diaphragm of 10mm we measured 36L/min.

Observations : At the first corner a large part of the flow turned left and went on the shelf, circulating in a larger vein than the previous experiment.

Some water reaches the 2nd corner splitting in a major part entering the depression and a smaller and slower part flowing along the slope.

There was no recirculation inside the depression, all the flow was evacuated at the end of the depression bypassing the "land".

Experiment EXP17

At the beginning of the experiment there were almost no oscillations left.

The desired flow was 20L/min, with a diaphragm of 8.2mm we measured 21.6 L/min.

Observations : At the 1st corner the flow continued straight ahead. At the 2nd corner the flow split in 2 branches of equal size, one following the slope and on entering the depression.

There was a very clear and narrow jet following the bathymetry inside the depression, with a lot of meanders on its sides.

After a while a branch detached from the slope current at the western corner of the depression entered the depression and recirculated joining the eastern side of the depression (drawing on the experiment paper).

We wanted to keep the same setup and to increase the flow rate to 80L/min in order to have consistency but unfortunately we were running out of time.

7.8 Monday 18 September

The whole day was used to build the extension of the topography to move the source 2m further back. The source sits now at the same position relative to the topography, only 2m further back. The cameras stayed at the same positions. We used this day to process data and to learn more about UVMAT.

7.9. Tuesday 19 September

Experiment EXP18

Today, first the source had to be attached and the water filled up. Then the particles had to be spread out in the water again. So the experiment didn't start before 11:30.
At least, the source was now moved 2m further back and the foam of the source was improved to avoid bubbles leaving the source, which also worked with this flow rate.

Because the source was mainly too close for the high flow rates, we will only do experiments with flow rates higher than 50 and with a rotation rate of 50. We started with Q = 50l/min and no corner.

Observations: The flow was much more established, when it reached the view of PCO1. Regular waves occurred along the slope until it reached the first corner. Almost the whole current turned south into the trough.

Notes: The images for HS were accidentally only taken for 7min, but restarted again for 7 min (EXP18_B).

Experiment EXP19

We used the same setup at EXP18, just with a flow rate of Q=80l/min. Note that we still used the diaphragme of the original diameter-flow rate curve, which gives an actual flow rate of >90 l/min. We do that to be consistent with previous experiments.

From the PIV of previous experiments with a flow rate of 80, we have decided to decrease the dt from 100 to 50ms and also the exposure time from 30 to 20ms, because the correlations were not very high. However, the dt = 50 ms did not work during this experiment, so that in fact it was dt = 100ms, because only every second image was taken.

During this experiment, the honeycomb fell out of the source after the HS. We did therefore no vertical scan.

Note: This experiment had too many errors that we had to redo it!

Experiment EXP20

We redid experiment 19 and this time everything worked. dt is now 50ms and the exposure time 20ms. We had no corner. The measured flow rate was 93.5 l/min.

Observations: Most of the current is entering the trough and a small part passes the trough, following the slope.

Experiment EXP21

For this experiment, we wanted to increase the flow rate to Q = 110 l/min, with a diaphragme of 17 mm. The measured flow rate was 120 l/min. The experiment was again without corner and dt = 50 ms, E = 20 ms.

Observation: The whole current went into the trough.

Experiment EXP22

This experiment had the same setup as EXP21 but with the corner. The measured flow rate this time was only 100 l/min.

Note: The flow rate is difficult to measure for high values, as the time interval is shorter and therefore produces a larger error. However, in the end we can plot a diameter - flow rate curve fitted with all the values from our experiment. The diaphragme should determine the flow rate quite clearly and with the fitted curve, we can get the actual flow rate.

Observations: Most of the current went into the trough, only a small part passed by. There were many particles in the beginning of the experiment.

7.10 Wednesday 20 September

Experiment EXP23

After swirling up the particles in the tank, we started early at 08:13 with the first experiment. It is the same experiment as EXP22 with a flow rate of Q=80l/min. The calculated flow rate is 94 l/min, which is very close to the previous flow rate (100l/min).

Observation: The flow was similar to the previous experiment, with the main current entering the trough. A large rotation occurred on the shelf.

Experiment EXP24

We increased the water level to 70 cm and kept the corner. The plan was to do one experiment with Q=110 l/min, but because that was so similar to 80 l/min before, we decided to increase it to 150 l/min (diaphragme 19.5mm). Measurements of the flow rate gave a value of Q = 150 l/min. We kept the small dt = 50 ms and the E = 20 ms.

Observations: The current entered the trough in a very large radius and builds an anti-cyclone on the shelf. Only a very small part of the current went straight, passing the trough.

Note: During this and the following experiments with a water level of 70 cm, we didn't convert the flow rate relative to the source area below the water level! So, all these experiments give a smaller velocity compared to the same experiments with the lower water level.

Experiment EXP25

The desired flow rate was reduced to 80 l/min (diaphragme 16mm), which gave a flow rate of 94 l/min.

Observations: At the slope current towards the first corner, large vortices were created. The inflow into the trough was not very strong and more turbulent. Half of the flow passed the trough along the slope.

Experiment EXP26

The desired flow rate was reduced to 50 l/min (diaphrame 12.6mm), which gave a flow rate of 54 l/m. The images were started after the flow started, because it takes a long time for the flow to reach the view of the first camera.

Note: The images for the first HS were taken for too short time (10 min), so that the flow was not established yet. We restarted and took more images for 5 min (saved in EXP26_B).

Note: We added larger particles (60 micron) into the inflow water (instead of the 30 micron), because Joel suggested that it may be better for the PIV. The particles in the ambient water were still the same.

Observations: The current enters the trough in a very large radius.

Experiment EXP27

This is a repetition of EXP26 without the corner. The measured flow rate was 53.2 l/min.

Observations: Compared to EXP26 with the corner, the current enters the trough in a much smaller radius, following neatly the topography. After waiting, also a large part flow straight.

Note: In the beginning of the experiment, there were still a lot of particles in the ambient water.

Experiment EXP28

The desired flow rate was increased to 80 l/min (16mm), which gave a measured flow rate of 92.4 l/min.

We had to do this experiment twice (first one was overwritten), because the honeycomb in the source loosened in the beginning of the first time and released a lot of clustered particles/bubbles that spread over the whole flow area. The source had to be fixed and we didn't have the time to completely wait until all the clusters escaped the tank. During the second round of this experiment, there were therefore still some of the clusters moving into the sight of the cameras.

Observations: The largest part of the current continued straight, but there was quite a strong return flow across the trough from the western corner of the trough to the southeastern corner of the trough.

7.11 Thursday 21 September

During the night, the tank was emptied. This day was used to fix the 5 probes for the vertical salinity profiles, they were set up and tested. The lower part of the source was covered with tape to only get a squared inflow. Then, the tank was filled up with salt water (density of 11.0 kg/m3) and turned into solid body rotation with T=50s.

7.12 Friday 22 September

Today, the baroclinic experiments started. The density of the inflow was 1.2 kg/m3 to get a density difference of 9.8 km/m3. We were told that it is easier to increase reduce the density difference by adding salt in the inflow than to decrease it (However, after all we can switch between the two quite easily). Due to the higher density, we used 60 micron particles in the ambient and the inflow water.

Experiment EXP29

During this experiment, we started with a HS (20 min), run the profiles with the probes as soon as the inflow was established (while doing the HS) and did the vertical slice (2 min) and the scan afterwards.

The flow rate of 20 l/min (8.5mm diameter) gave a measured flow rate of 20 l/min.

Probes: downward velocity: 2 cm/s; upward velocity: 5 cm/s; break at the surface: 10s. In total 10 repetitions. The height of the probes are adjusted to follow the slope. They start about 20.5 cm above the slope so that only the probe furthest away from the coast touches the water, the others stay above the surface. During the profiles they get lowered by 19 cm.

Scans: 11 levels (55:-2:35) with 5 images at each level. dt = 100ms and dt1 = 1000ms. Exposure time was 30ms.

Observations: The flow was very fast and a nice clear front developed. In the beginning the water formed a vortex just after the first corner, and then most of the flow entered on the shelf between the first and the second corner. On the vertical slice, the front was clearly visible, but it was very steep, so that it may be above the laser sheets.

Note1: The vertical camera is too far up for the experiments with density difference, as the free surface is not visible. It will be put further down for the following experiments.

Note2: The second probe from the coast is not working properly!!

Note3: We have to remember that the source area is smaller and the flow therefore higher!

Experiment EXP30

The flow rate was increased to 50 l/min (12.6mm diameter), which gave a measured flow rate of 53 l/min. The flow was very high and the dt therefore reduced to 50ms and Exposure time was reduced to 20ms after a few minutes (saved in Exp_B).

Probes: We kept the same setup as EXP29, but did now 20 profiles.

Observations: The flow turn onto the shelf at the second corner and followed the topography also on southern side of the topography. After a while, a vortex developed in the trough and parts of the main flow entered this vortex.

Note: In the beginning of the experiment, the water was still in motion from the previous experiment due to the density difference.

Experiment EXP31

The flow rate was increased even more to 80 l/min (14mm diameter), which gave a measured flow rate of 75 l/min. As in the previous experiment, the we always got too high flowrates for the 80 l/min - experiments with the 16mm diameter, we decided to now chose a smaller diameter. We were also concerned that it was not possible to do PIV with an ever faster flow, because it is not possible to use a dt < 50 ms and E<20!

Observations: All water turned onto the shelf at the 2nd corner and formed a vortex over the shelf. Some of the flow turned within the trough in a very wavy way (didn't follow the topography out of the trough again).

Experiment EXP32

Now, we removed the corner and did an experiment with 50 l/min flow rate (12.6 mm), which gave 52 l/min. dt = 50ms, E=20ms.

Observations: The flow went straight around the 1st corner and followed the topography.

Notes: The water was still in motion in the trough from the previous experiment and due to the density difference. It looked like the particles started settling down, which may create vertical motion. It may be, that there was already a fresh water layer at the surface that influenced the flow, but also the laser height.

Experiment EXP33

This experiment was with a low flow rate of 20 l/min and had to be stopped after a while as there was a too strong stratification. The laser then got deflected downward at the interface between the two densities and was only about 1-0 cm above the topography on the shelf.

7.13 Monday 25 September

Experiment EXP34

We did NOT redo experiment EXP33! Instead, we increased the inflow density to get a difference of about 5 kg/m3 (measured: 4.8 kg/m3). The measured water level was 58.4 cm in the centre and the laser sheet was at 55 cm in the centre. We started with an inflow of 20 l/min (8.5mm) which measured 22.3 l/min.

Experiment EXP35

The inflow was increased to 50 l/min.

Note: At the beginning of the experiment, the water was not completely at steady-state in the trough. The scans were supposed to do 550 images, but only took 500. And it only did 10 levels, instead of 11, so the last level is missing!

' 6 - Table of Experiments:

  • will be updated from Excel sheet!
Exp No.NameValve opened$H_{water}$$T_{rot}$$Q$$R_{curvature}$$\Delta \rho$Diaphragme Flowrate estimated Type of photoHeight of HSHeights of Scan and #Time photo DyeComments Notes
$(m)$$(s)$$(L min^{-1})$$(m)$$(kg m^{-3}$$(mm)$$H_1,H_2,\Delta t $VS/HS/Scan$cm$$cm$t1/t2/t3
$\Delta t$
yes / no
0test11092017 hh:mm0.6045058.50.50013170,150,41sVS--?yesDye did not show up in photos :-(Elin
0.5845040.670.50012.6170,150,59sHS,VS,Scan5555:[-5:30,4 -5:30;4]500ms
(reality: 1000ms)
noExpl. time = 50 ms too little. Only every second photo was takenElin
(reality: 400ms)
noOnly every second photo was takenElin
3exp031309201750500.50012.6HS,Scan200msnoProblem with photos fixed; particles in ambientElin
4exp0413092017500.5006.8HS,Scan200msnoLaser started vibrating for the scansElin

Attachments (18)

Download all attachments as: .zip