Systematic errors estimation in repeated MBES surveys · 2020. 4. 17. · HBMC: a challenging context Sparse and irregular time distribution of MBES surveys Hydrodynamic complexity

1

Systematic errors estimation in repeated MBES surveys N. Debese J. J. Jacq K. Degrendele M. Roche R. Moitié

CHC 2020, Québec, Wednesday February 26

2

Data description DBM Registration Error analysis Submarine Sandbank monitoring

Belgium continental shelf

HBMC monitoring area is localized in the S4c sector of the belgian continental shelf

Oost hinder bank

S4c sector HBMC monitoring area

HBMC

Sand extraction

3


Estimation of the extracted volumes

Electronic Monitoring System (EMS) records the GNSS positions and pumps activity levels of dredgers

Time (days)

-1,5

-1,3

-1,1

-0,9

-0,7

-0,5

-0,3

-0,1

0 400 800 1200 1600 2000 2400

Equ

ival

ent t

hick

ness

cha

nge

(m)

2012 2013 2015 2016 2017 2018 2014 2019

No sand extraction No sand extraction

EMS data

4

Data description DBM Registration Error analysis

13 MultiBeam Echo Sounder mapping surveys Bathymetric data acquired using the RV Belgica MBES : SIMRAD/EM3002D

April 2012

Index Date1 Apr. 2, 12

2 May 9, 12

3 Mar. 14, 13

4 Oct. 3, 13

5 Mar. 12, 14

6 May 6, 14

7 Nov 24, 14

8 May 7, 15

9 Dec 15, 15

10 Sep 21, 17

11 Mar 14, 18

12 Sep 27, 18

13 Jul 3, 19

At various time sampling Intervals (1 to 22 months between two surveys)

In various environmental contexts (Tide, storms...)

Using various data corrections (Tide gauge or GNSS tide corrections)

July 2019

depo

sit

Ero

sion

DTM of differences between two successive DBM

Sand dunes dynamics Dredging impact

time

Submarine Sandbank monitoring

5

MBES Survey’s index

Data description DBM Registration Error analysis Submarine Sandbank monitoring E

quiv

alen

t thi

ckne

ss c

hang

e (m

)

-1,5

-1,3

-1,1

-0,9

-0,7

-0,5

-0,3

-0,1

0 400 800 1200 1600 2000 2400

Time (days)

2012 2013 2015 2016 2017 2018 2014 2019

Analysis of discrepancies between volumes estimated using EMS and MBES data

While taking the first MBES (#1) as the reference survey

EMS data

MBES data

6


HBMC: a challenging context

Sparse and irregular time distribution of MBES surveys

Hydrodynamic complexity of dune migration Irregular anthropogenic process

Oost hinder bank stability

Last survey (#13) corrected using GNSS tide

2012 2013 2015 2016 2017 2018 2014 2019

0

0,2

0,4

0,6

0,8

1

1,2

1,4

1,6

# 13

Time (days)

Taking as the reference survey

0.39 m

Systematic errors highlighting

Equ

ival

ent t

hick

ness

cha

nge

(m)

EMS data

MBES data

7

Data description DBM Registration Error analysis Automatic spatio temporal analysis

Sandbank Sand dunes dyamics

Sand extraction Systematic errors

OSC

Step 1: Getting rid of dune footprint

Through bottom-osculatory surfaces

DBM #1 DBM #2 DBM #i DBM #j DBM #13

OSC #1 OSC #2 OSC #i OSC #j OSC #13

Sand dunes dynamics

Sandbank Sand extraction

Systematic errors

8


Step 2: Taking sandbank stability into account

OSC #1 OSC #2 OSC #i OSC #j OSC #13 Sandbank

Sand extraction Systematic errors

OSC(t) OSC(t+1)

Sandbank

OSC j – OSC i Sand extraction Systematic errors

Through differences of two successive bottom-osculatory surfaces

9


Step 3: Robust fitting of a planar top-osculatory surface

OSC j – OSC i Sand extraction Systematic errors

Through planar top-osculatory surface applied to differences of two successive bottom-osculatory surfaces

PlanOSCsup(OSC j – OSC i)

Sand extraction

Planar top-osculatory surfaces, obtained using a global robust registration approach, represent:

Vertical bias: Heave Tide Drafts (dynamic and static)

10 2012 2013 2015 2016 2017 2018 2014 2019


MBES Survey indices 13 12 11 10 9 8 7 6 5 4 3 2 1

Mean correction (m) 0.00 0.11 0.40 0.36 0.01 0.33 0.27 0.33 0.42 0.42 0.29 0.44 0.42

Systematic error estimation through a global and robust registration of osculatory surfaces

Results

0

0,2

0,4

0,6

0,8

1

1,2

1,4

1,6

Equ

ival

ent t

hick

ness

cha

nge

(m)

Time (days)

EMS data

MBES data

Robust registration applied to MBES data

11

Data description DBM Registration Error analysis A deep analysis of bathymetric raw data

Heave

Index Mean (m) Std(m)

9 -0.18 0.10 12 -0.11 0.070 3 -0.10 0.19 6 -0.06 0.05 1 -0.06 0.23

4 -0.04 0.09 2 -0.04 0.05 5 -0.01 0.09 8 -0.0 0.10 7 0.01 0.12

13 0.03 0.18 11 0.17 0.14 10 0.22 0.08

Analysis of heave data Analysis of tide and dynamic draft dataMean heave values not centered on zero for some surveys

Hea

ve (m

)

Survey indices

Z GPS data acquired for surveys #5 to #13 but not used (too many discontinuities and loss of RTK signal)

Example of survey #12

Mean of differences between DBM

Mean correction (Osculatory approach)

Offset explained by heave analysis

Offset explained by heave and tide analysis

GNSS tide correction GNSS tide correction (heave

offset on zGPS)

12


0,00

0,20

0,40

0,60

0,80

1,00

1,20

1,40

1,60

2012 2013 2015 2016 2017 2018 2014 2019

Equ

ival

ent t

hick

ness

cha

nge

(m)

Time (days)

EMS data

MBES data

Robust registration applied to MBES data

MBES data corrected for systematic errors

z GPS not acquired

A deep analysis of bathymetric raw data

Confirms the estimation of errors deduced from the robust registration of osculatory surfaces time series

13

Conclusions and perpectives

Osculatory surfaces are robust asymmetric trend surfaces

can be applied to detect and estimate systematic errors in bathymetric time series

can be used as reference surfaces (sand extraction monitoring)

are useful tools for investigating the dynamics of sand dunes

Upcoming studies Processing of other areas (BRMC ...)

Estimating volume of available sand (mobile sand)

Generalization of the approach to the detection and analysis of other underwater structures

Systematic errors estimation in repeated MBES surveys · 2020. 4. 17. · HBMC: a challenging context Sparse and irregular time distribution of MBES surveys Hydrodynamic complexity

Documents