WHOTS - ADCP Compass Calibration History - PowerPoint Presentation

Slide1

WHOTS - ADCP Compass Calibration History

Version

4

:

November

9,

2016

By:

Walt

DeppeSlide2

File History

Version 1 – Walt

Deppe

, 9/23/2014

Version 2 – Walt

Deppe

, 10/2/2015

Version

3

– Walt

Deppe

,

9/13/2016

Version

4

– Walt

Deppe

,

11/9/2016Slide3

Outline

Background

Switch from Snug to UH

Magnetic Anomaly Surveys

Things to look for & keep in mind

ADCP S/N & Deployment

List

Individual Instrument Error Analysis

Observations

RecommendationsSlide4

Background

Snug Harbor: initial compass calibration location

Between WHOTS-1 and -2, turnaround & compass calibration done at sea

Some early compass calibration spin data not recorded (and s/n 1825 originally prepared by WHOI)

Switched calibration location from Snug to UH after large magnetic anomalies detected at Snug

Calibration frame changed over years

Trash can

Wooden frame

Aluminum frame (current)Slide5

Switch from Snug to UH

Following WHOTS-7 compass calibration and before WHOTS-6 post-deployment check, compass-spin location switched from Snug Harbor to UH Lower Campus Soccer Field after large magnetic field anomalies detected at Snug

Magnetic Anomaly Surveys (August 2010)

Located on OHU in ‘

C

:\

Users\Craig\Documents\WHOTS\

Compass_Calibs

’

Snug Harbor Analysis (Aug 24

th

):

magnetic_survey.doc

UH Soccer Field Analysis (Aug 27

th

):

magnetic_survey2.doc

“

The manufacturer of the compass used for the survey did not list accuracy specifications but given the compass does not appear to be crafted with precision tools and given the very poor eyesight of the surveyor, it is possible that the +/- 1.5 differences seen could be considered “in the noise”.

“

If/when the Lower Campus site is used for compass calibrations, it would be advisable to perform the calibrations at the midpoint of the 100 point reference line outlined in this survey as it appears to have the smallest mean differences from magnetic north and the smallest standard deviations.

“Slide6

Magnetic Anomaly Surveys

Snug Harbor

UH Lower Campus Soccer FieldSlide7

Things to look for & keep in mind

Differences between calibrations done at Snug (worse) and UH (better)

Tightening of error amplitude after calibration during pre-deployment observations

Comparison between spin

error amplitudes & RDI

“Overall Error” output

Calibration

purpose: reducing error around compass rows due to battery interference

signature

Pre-deployment, pre-calibration spin is using calibration matrix determined for previous battery pack before last deployment

Spin errors currently not used in any data processing, only RDI calibration matrix affects dataSlide8

ADCP S/N & Deployment List

300 kHz

s/n 4891 (WHOTS – 1, 2, 4, 6, 8, 10,12)

s/n 7637 (WHOTS – 3, 5, 7, 9, 11,13)

600 kHz

s/n 1825 (WHOTS – 4, 6, 8, 10,12)

s/n 13917 (WHOTS – 7, 9, 11,13)

1200 kHz

s/n 2530 (WHOTS – 5)Slide9

Individual Instrument Error Analysis

For Each ADCP:

Compass spin error plots by deployment

Compass spin error stats over time

Comparison of spin stats w/ RDI output

ADCPs

s/n 4891

s/n 7637

s/n 1825

s/n 13917

s/n 2350Slide10

S/N 4891Slide11

s/n 4891 – 300 kHz

No spin errors recorded for WHOTS-1 pre-deployment (Snug) or post-deployment (@ sea) or WHOTS-2 pre-deployment (@ sea)

Note

WHOTS-6 post-deployment mean error (performed at UH soccer field) does not reflect large pre-deployment offset (performed at Snug)Slide12

s/n 4891Slide13

S/N 7637Slide14

s/n 7637 – 300 kHz

Instrument

damaged during

WHOTS-5.

No post-deployment check

.

No pre-deployment post-

cal

spin check errors available

Note

WHOTS-7

post-deployment mean error (performed at UH soccer field) does not reflect large pre-deployment offset (performed at Snug

)

A boat

docked 60 m. away

half-way through pre-deployment calibration

for

WHOTS-7Slide15

s/n 7637Slide16

S/N 1825Slide17

s/n 1825 – 600 kHz

WHOTS-4, prepared by WHOI (no cal records available)

Note WHOTS-6 post-deployment mean error (performed at UH soccer field) does not reflect large pre-deployment offset (performed at Snug)Slide18

s/n 1825Slide19

S/N 13917Slide20

s/n 13917 – 600 kHz

Instrument damaged during WHOTS-7. No post-deployment check.Slide21

s/n 13917Slide22

S/N 2530Slide23

s/n 2530 – 1200 kHz

Only one deploymentSlide24

s/n 2530Slide25

Observations

Calibrations done at Snug have mostly larger mean errors (offsets) but similar error amplitudes to UH calibrations

Pre-Deployment, After Calibration consistently smallest error amplitude

Spin error amplitudes & RDI Overall Error very similar, as expected

Not large enough sample size at any calibration location to conclude that offsets based on mean error should be applied at this time

Large mean errors (offsets) could be a product of magnetic anomalies at calibration site or inconsistencies/errors in setup of N-S reference lineSlide26

Recommendations

Continue doing calibrations at UH Soccer field

Find cross-references (E/W) for calibration stand location along N-S reference line in order to keep consistency between calibrations

Most important thing to check for

L

ow pre-deployment after-

cal

“overall error”/spin error amplitude (<2°)

Apply no corrections/offsets based on spin-mean

Calibration purpose: reducing error around compass rows due to battery interference signature

Large spin-mean offsets likely due to magnetic anomalies at

cal

site (Snug) or inconsistencies/inaccuracies in setting up N-S lineSlide27

Recommendations

Run new

Matlab

graphing

scripts

produced by Walt for each instrument after every calibration session (spin & stats)

Path:

/export/aina1/

whots

/

ADCP_cmpscals

/

Create new text files for each spin session according to templates

template_adcpSNNUMcmpss_eval_PRE_whotsDEPNUM.txt

t

emplate_adcpSNNUMcmpss_eval_POST_whotsDEPNUM.txt

plot_adcp_cmpscal.m

– By-deployment/instrument compass spin error plots

stats_adcp_cmpscal.m

– By-instrument compass spin error

stat plots

over time

Use

as sanity check and to follow history for each instrument

These scripts will make for more consistent cataloging of results for future deployments