scieee Science in your language
[en] (orig)

FAIR ADCP data with OSADCP: a workflow to process ocean current data from vessel-mounted ADCPs

Abstract

This paper presents the open-source Python software OSADCP developed for the processing of vessel-mounted Acoustic Doppler Current Profiler (VMADCP) data. At this stage, the toolbox is designed for processing VMADCP measurements from open-ocean applications of Teledyne RDI Ocean Surveyor ADCPs and the data acquisition software VMDAS. Based on the VMDAS ENX binary output format, the software contains implementations for cleaning and vector-averaging of single-ping velocity data, verification of the position data, and applying misalignment and amplitude corrections. The procedures of OSADCP are described in detail to encourage the scientific community to use it for their own purposes. The toolbox is an integral part of a workflow implemented on the German marine research vessels in the framework of the Underway Research Data project of the German Marine Research Alliance (DAM). It aims to ensure standardized data acquisition measures, reliable data transfer from the ADCP to shore both near-real-time and in delayed-mode, processing and quality control, and dissemination of the curated data product in the data repository PANGAEA. From PANGAEA, data sets are forwarded to the European marine data hubs Copernicus Marine Service and EMODnet. The workflow that forms the framework for OSADCP is described here as an example of scientific data management that follows the FAIR data guidelines.

Read accessible full text

FAIR ADCP data with OSADCP: a workflow to process ocean current data from vessel-mounted ADCPs

Author: Kopte, Robert,Becker, Marius,Fischer, Tim,Brandt, Peter,Krahmann, Gerd,Betz, Maximilian,Faber, Claas,Winter, Christian,Karstensen, Johannes,Wiemer, Gauvain
Year: 2024
DOI: 10.3389/fmars.2024.1425086
Source: https://macau.uni-kiel.de/servlets/MCRFileNodeServlet/macau_derivate_00006578/fmars-11-1425086.pdf
FAIR ADCP da a wi h OSADCP:
a wo kflow o p ocess ocean
cu en da a om essel-
moun ed ADCPs
Robe Kop e
1
*, Ma ius Becke
1
, Tim Fische
2
, Pe e B and
2,3
,
Ge d K ahmann
2
, Maximilian Be z
4
, Claas Fabe
5
,
Ch is ian Win e
1
, Johannes Ka s ensen
2
and Gau ain Wieme
6
1
Ins i u e o Geosciences, Kiel Uni e si y, Kiel, Ge many,
2
Depa men o Physical Oceanog aphy,
GEOMAR Helmhol z Cen e o Ocean Resea ch Kiel, Kiel, Ge many,
3
Facul y o Ma hema ics and
Na u al Sciences, Kiel Uni e si y, Kiel, Ge many,
4
Compu ing and Da a Cen e, Al ed-Wegene -
Ins i u e o Pola and Ma ine Resea ch, B eme ha en, Ge many,
5
Da a Managemen , GEOMAR
Helmhol z Cen e o Ocean Resea ch Kiel, Kiel, Ge many,
6
Da a Managemen and Digi aliza ion,
Ge man Ma ine Resea ch Alliance (DAM), Be lin, Ge many
This pape p esen s he open-sou ce Py hon so wa e OSADCP de eloped o
he p ocessing o essel-moun ed Acous ic Dopple Cu en P ofile (VMADCP)
da a. A his s age, he oolbox is designed o p ocessing VMADCP
measu emen s om open-ocean applica ions o Teledyne RDI Ocean Su eyo
ADCPs and he da a acquisi ion so wa e VMDAS. Based on he VMDAS ENX
bina y ou pu o ma , he so wa e con ains implemen a ions o cleaning and
ec o -a e aging o single-ping eloci y da a, e ifica ion o he posi ion da a,
and applying misalignmen and ampli ude co ec ions. The p ocedu es o
OSADCP a e desc ibed in de ail o encou age he scien ific communi y o use
i o hei own pu poses. The oolbox is an in eg al pa o a wo kflow
implemen ed on he Ge man ma ine esea ch essels in he amewo k o he
Unde way Resea ch Da a p ojec o he Ge man Ma ine Resea ch Alliance (DAM).
I aims o ensu e s anda dized da a acquisi ion measu es, eliable da a ans e
om he ADCP o sho e bo h nea - eal- ime and in delayed-mode, p ocessing
and quali y con ol, and dissemina ion o he cu a ed da a p oduc in he da a
eposi o y PANGAEA. F om PANGAEA, da a se s a e o wa ded o he Eu opean
ma ine da a hubs Cope nicus Ma ine Se ice and EMODne . The wo kflow ha
o ms he amewo k o OSADCP is desc ibed he e as an example o scien ific
da a managemen ha ollows he FAIR da a guidelines.
KEYWORDS
ADCP, unde way da a, FAIR da a, py hon, ocean cu en s
F on ie s in Ma ine Science on ie sin.o g01
OPEN ACCESS
EDITED BY
Sa ah T. Gille,
Uni e si y o Cali o nia, San Diego,
Uni ed S a es
REVIEWED BY
E ic Fi ing,
Uni e si y o Hawaii a Manoa, Uni ed S a es
An onio No ellino,
ETT SpA, I aly
*CORRESPONDENCE
Robe Kop e
[email p o ec ed]
RECEIVED 29 Ap il 2024
ACCEPTED 02 Augus 2024
PUBLISHED 26 Augus 2024
CITATION
Kop e R, Becke M, Fische T, B and P,
K ahmann G, Be z M, Fabe C, Win e C,
Ka s ensen J and Wieme G (2024) FAIR
ADCP da a wi h OSADCP: a wo kflow
o p ocess ocean cu en da a
om essel-moun ed ADCPs.
F on . Ma . Sci. 11:1425086.
doi: 10.3389/ ma s.2024.1425086
COPYRIGHT
© 2024 Kop e, Becke , Fische , B and ,
K ahmann, Be z, Fabe , Win e , Ka s ensen and
Wieme .Thisisanopen-accessa icle
dis ibu ed unde he e ms o he C ea i e
Commons A ibu ion License (CC BY). The
use, dis ibu ion o ep oduc ion in o he
o ums is pe mi ed, p o ided he o iginal
au ho (s) and he copy igh owne (s) a e
c edi ed and ha he o iginal publica ion in
his jou nal is ci ed, in acco dance wi h
accep ed academic p ac ice. No use,
dis ibu ion o ep oduc ion is pe mi ed
which does no comply wi h hese e ms.
TYPE Technology and Code
PUBLISHED 26 Augus 2024
DOI 10.3389/ ma s.2024.1425086
1 In oduc ion
Ocean cu en measu emen s a e indispensable o
unde s anding he dynamics and ci cula ion pa e ns o ma ine
en i onmen s. Ocean cu en s edis ibu e hea and impac he
su ace ocean momen um flux and hus influence global clima e
pa e ns ha ha e a - eaching e ec s on wea he sys ems
wo ldwide. Addi ionally, ocean cu en s impac he heal h o
ma ine ecosys ems by dispe sing nu ien s, dilu ing pollu an s,
anspo ing plank on and a ec ing he beha io o mobile
species. Unde s anding and moni o ing ocean cu en s is essen ial
o p edic ing clima e change impac s, imp o ing wea he o ecas s,
and ensu ing he sus ainable managemen o ma ine esou ces.
Obse a ions o ocean cu en s se e as suppo ing a iables o he
es ima ion o so-called Essen ial Ocean Va iables (EOVs,
Linds om e al., 2012), which a e he backbone o obse a ional
con ibu ions o bo h he Global Ocean Obse ing Sys em (GOOS)
and he Global Clima e Obse ing Sys em (GCOS).
Acous ic Dopple Cu en P ofile s (ADCPs) a e among he
mos commonly used oceanog aphic ins umen s, as hey e ficien ly
measu e ocean eloci y p ofiles de i ed indi ec ly om he acous ic
echoes o suspended ma e passi ely ad ec ed by he cu en . A
s anda d applica ion in oceanog aphic esea ch is he use o ADCPs
on a mo ing ship, in he ollowing e e ed o as essel-moun ed
ADCPs (VMADCPs). VMADCPs acili a e unde way and quasi-
s a iona y measu emen s o uppe ocean cu en s, co e ing a
measu emen ange o se e al hund ed me e s below he ship.
Beginning wi h he Wo ld Ocean Ci cula ion Expe imen
(WOCE) in he 1990s (King e al., 2001), VMADCPs ha e
become he s anda d ins umen o measu ing ocean cu en s
du ing esea ch c uises. They a e ou inely used o moni o
changes o he e ical s uc u e o he flow along epea ed
sec ions, e.g., in he opical Pacific(Johnson e al., 2002), in he
opical A lan ic (Johns e al., 2014;B and e al., 2014) o ac oss
D ake Passage (Sp in all e al., 2012;Fi ing e al., 2012). In gene al,
VMADCPs assis esea ch ac oss all ma ine science disciplines,
enhancing in e disciplina y p ocess s udies by p o iding eal- ime
in o ma ion abou he ocean cu en benea h he ship. This
capabili y o example suppo s apid-sampling s a egies ha
allow scien is s o s udy ansien phenomena like mesoscale
eddies, on s, wa es and upwelling e en s (e.g., Menkes e al.,
2002;Shche bina e al., 2015;Hummels e al., 2020;L’Hega e
e al., 2023).
To maximize he alue o VMADCP ocean cu en da a
p oduc s o scien ific and moni o ing use, hei cu a ion,
publica ion, accessibili y and p ese a ion is a key ask. Recen
demands in ma ine da a managemen in gene al a e d i en by he
need o high- esolu ion, e en ually eal- ime da a o add ess
complex global issues such as clima e change, biodi e si y loss
and ocean heal h. In his con ex , he lack o s anda dized da a
o ma s and p o ocols emains a pe sis en issue complica ing da a
in eg a ion and in e ope abili y. A in e na ional le el, a ious
g oups and o ganiza ions add ess hese challenges by de eloping
commonly ag eed s anda ds and bes p ac ices, imp o ing
coo dina ion and dialogue ac oss p ac i ione s and use s.
Embedded in he In e go e nmen al Oceanog aphic Commission
(IOC) o he Uni ed Na ions Educa ional, Scien ific and Cul u al
O ganiza ion (UNESCO) is he Obse a ion Coo dina ion G oup
(OGC) o he GOOS ha assembles a ious global obse a ion
pla o m-specific g oups, such as GO-SHIP, which ocuses on da a
om esea ch essels, and A go, which manages da a om p ofiling
floa s. As gene al guidelines o he managemen o scien ific da a,
he FAIR (Findable, Accessible, In e ope able, Reusable) p inciples
(Wilkinson e al., 2016) we e accep ed as a guiding p inciple o
da a managemen , oge he wi h he CARE (Collec i e benefi ,
Au ho i y o con ol, Responsibili y, E hics) p inciples (Ca oll
e al., 2021). E ficien coo dina ion a na ional le el is a
p e equisi e o he ep esen a ion o indi idual coun ies and
cons uc i e coope a ion o suppo he in e na ional e o s. In
Ge many, wo di e en associa ions o ma ine esea ch ins i u ions
enable s a egic ini ia i es and o e a ching discussions,
encompassing he ins i u ional di e si y. These associa ions a e
he Konso ium Deu sche Mee es o schung (KDM), ounded in
2004, and he Ge man Ma ine Resea ch Alliance (DAM). One
ini ia i e, launched by he DAM in 2019, is he Unde way Resea ch
Da a p ojec . The p ojec pools he da a managemen expe ise o
a ious ma ine esea ch ins i u ions o demons a e he sys ema ic
p o ision o FAIR obse a ional da a om he Ge man ma ine
esea ch essels. Following he app oach aken in o he coun ies,
such as he Rolling deck o Reposi o y in he U.S. (R2R, Ca bo e
e al., 2022), wo kflows o ship bo ne scien ific senso s ha e been
es ablished ha add ess ship-specific equi emen s o
configu a ion, (nea -) eal- ime moni o ing, da a access and
quali y assu ance. Wi h his p ojec , DAM also con ibu es o he
ecen ly es ablished Na ional Resea ch Da a In as uc u e (NFDI,
Ha l e al., 2021) as a case s udy on sys ema ic FAIR and open
p o ision o he e ogeneous and ansdisciplina y ma ine da a.
VMADCPs a e among he senso s co e ed by he DAM
Unde way Resea ch Da a p ojec . A e a sho summa y o he
s a e o he a o VMADCP echnology (Sec ion 2), his a icle
desc ibes he wo kflow implemen ed on he la ge ships o he
Ge man Ma ine esea ch flee as an example o scien ific da a
managemen ha ollows FAIR da a guidelines (Sec ion 3). An
essen ial componen o he wo kflow is he pos -p ocessing so wa e
OSADCP (Kop e e al., 2024), w i en in Py hon and designed o
p ocess VMADCP aw da a acqui ed using Teledyne RDI’s Vessel
Moun Da a Acquisi ion So wa e (VMDAS). The handling o
OSADCP is desc ibed in Sec ion 3.3, as a basic amewo k o he
so wa e is made a ailable o he scien ific communi y along wi h
his a icle.
2 Backg ound
2.1 (VM)ADCP echnology
ADCPs in e wa e eloci ies indi ec ly om he acous ic echoes
o suspended ma e , e.g., plank on o suspended sedimen ,
assuming passi e ad ec ion o he sca e e s wi h he cu en .
Hyd oacous ic pulses a e emi ed along na ow, slan ed beams.
The ene gy is abso bed, sca e ed and pa ly eflec ed by he
sca e e s o be ecei ed by he ADCP. As o he Dopple e ec , a
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g02
mo emen o sca e e s along he beam axis causes a change o
equency o he ecei ed echo compa ed o he equency o he
emi ed pulse. Wi h indi idual beams o ien ed in di e en
di ec ions, each beam measu es a p ojec ion o he local eloci y
ec o on i s espec i e beam axis. Assuming ho izon ally
homogeneous flow ac oss he a ea co e ed by he slan ed beams,
he beam geome y is aken in o accoun o calcula e he
componen s o he cu en eloci y in ca esian coo dina es (RD
Ins umen s, 2010).
In o iginal ADCP echnology, a single homogeneous acous ic
pulse is emi ed (na owband ping ype). La e , he b oadband ping
ype was de eloped, which sends a con inuous sequence o sub-
pings coded by phase e e sals (Pinkel and Smi h, 1992). The eby
he in o ma ion con en o he e u ning echo is inc eased,
imp o ing he p ecision o he eloci y measu emen s
subs an ially, a he expense o some educ ion in ange
and obus ness.
To de e mine a eloci y p ofile om one acous ic pulse, he
ange along he acous ic pa h is spli in o successi e empo al
segmen s, which co espond o dep h cells a successi ely g ea e
dis ances om he de ice. This p ocess is e e ed o as ange-ga ing
(RD Ins umen s, 2011). In deep-wa e applica ions, VMADCPs
ypically ope a e a equencies be ween 150 kHz and 38 kHz,
esul ing in maximum p ofiling anges be ween 400 m and mo e
han 1000 m, espec i ely. To achie e highe e ical esolu ion a
he expense o p ofiling ange, ADCPs wi h highe equencies, e.g.,
600 kHz o 300 kHz, may be used simul aneously wi h long-
ange ADCPs.
2.2 VMADCP da a acquisi ion sys ems
The mos common acquisi ion sys ems used by he
oceanog aphic communi y a e VMDAS and UHDAS, as desc ibed
and compa ed by Fi ing and Hummon, 2010.VMDAS is a
Windows so wa e p o ided by Teledyne RD Ins umen s
(TRDI), one o he manu ac u e s o ADCPs. UHDAS is a Linux
so wa e de eloped and main ained by he Uni e si y o Hawaii,
U.S (Fi ing e al., 2012;Hummon e al., 2023). UHDAS handles
auxilia y da a s eams mo e obus ly han VMDAS. I p o ides
au oma ic p e-p ocessing and he expo o diagnos ic da a, and can
p ocess da a om se e al VMADCPs a he same ime.
2.3 VMADCP pos -p ocessing solu ions
Se e al s eps o pos -p ocessing a e equi ed o de e mine ocean
cu en s om aw VMADCP da a (Fi ing and Hummon, 2010).
These include coo dina e ans o ma ions, he cleaning o single-
ping da a, me ging o ADCP wi h auxilia y da a, e ifica ion o
posi ioning in o ma ion, ec o -a e aging o single-ping eloci y
da a, and he applica ion o heading misalignmen and eloci y
ampli ude co ec ions. Pos -p ocessing also includes he gene al
assessmen o da a quali y, which may be influenced by
en i onmen al condi ions, e.g., he sea s a e o he occu ence and
he na u e o sca e e s in he wa e column.
Scien ific use s de eloped a numbe o pos -p ocessing ools o
VMADCP da a, ollowing di e en app oaches and enabling
di e en ways o quali y con ol. Mos o hese ools ead ADCP
aw da a in pd0 o ma , he s anda d bina y ou pu o ma o TRDI.
Basic p ocessing solu ions consis o sc ip s, which implemen
essen ial p ocessing s eps and a e easily adap able o specific
equi emen s o use cases. One example is he OSSI MATLAB
oolbox, de eloped a he depa men o Physical Oceanog aphy o
he Ins i u ü Mee eskunde, now GEOMAR Helmhol z Cen e o
Ocean Resea ch Kiel, Ge many (Fische e al., 2003). The OSSI
oolbox was no eleased o ficially, bu is widely used in he Ge man
ADCP use communi y. Ano he example o a sc ip -based
implemen a ion o p ocessing and isualiza ion o essel-moun ed
ADCP da a is con ained in he R package o oceanog aphic
analysis oce (Kelley, 2018). Mo e in eg a ed and au oma ed
me hods a e also a ailable. CASCADE is an open-sou ce
MATLAB p og am de eloped a he F ench Labo a o y o
Physical Oceanog aphy (LOPS, Le Bo e al., 2011). CASCADE
uses a g aphical in e ace and allows he selec ion o h esholds and
c i e ia o da a quali y checks. A widely-used p ocessing
amewo k is CODAS p ocessing, de eloped and main ained by
he UH Cu en s G oup a he Uni e si y o Hawaii, U.S (Fi ing and
Hummon, 2010;Hummon e al., 2023). CODAS p ocessing is used
by he UHDAS acquisi ion so wa e o p o ide online p e-
p ocessing and emo e moni o ing du ing da a acquisi ion.
2.4 VMADCP da a dissemina ion
Following acquisi ion and pos -p ocessing, he alue chain o
VMADCP da a is supposed o yield quali y-con olled and openly
accessible ocean eloci y da a p oduc s, published acco ding o
FAIR p inciples (Wilkinson e al., 2016). VMADCP da a om U.S
na ional c uises bu also om a ious in e na ional p og ams a e
published by he Join A chi e o Shipboa d ADCP (JASADCP),
main ained by he Na ional Cen e s o En i onmen al
In o ma ion(NCEI)incollabo a ion wi h he Uni e si y o
Hawaii. In he E.U., he Cope nicus Ma ine Se ice (CMEMS)
and EMODne a e he main in eg a o s o pla o m-independen
ma ine in-si u da a. They in eg a e da a om na ional
oceanog aphic eposi o ies, e.g., om PANGAEA, he Ge man
node o ea h sys em esea ch da a (Felden e al., 2023).
None heless, due o he he e ogenei y o he da a se s ollowing
di e en me a da a s anda ds and ocabula ies, in eg a ion o en
emains a challenge and e o s a e needed o achie e a highe le el
o ha moniza ion.
3 Wo kflow o essel-moun ed ADCP
da a managemen
The wo kflow (Figu e 1) p esen ed he e aims o map he en i e
alue chain o VMADCP da a and co e s he a eas o da a
acquisi ion (Sec ion 3.1), da a ans e (Sec ion 3.2), da a
p ocessing and quali y con ol (Sec ion 3.3) and dissemina ion o
he final da a p oduc (Sec ion 3.4).
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g03
3.1 Da a acquisi ion
The Ge man ma ine esea ch essels R/V Ma ia S. Me ian, R/V
Me eo , and R/V Sonne a e equipped wi h 75-kHz and 38-kHz
Teledyne RDI Ocean Su eyo ADCPs o p ofiling he uppe wa e
column, up o 1200 m dep h. High-quali y, ime-synch onous
posi ion and a i ude da a a e p o ided by Kongsbe g Seapa h
sys ems. Each VMADCP is con olled by a VMDAS ins ance
ins alled on a dedica ed senso PC ha bundles he VMADCP
and na iga ion da a s eams. All aw da a, ha accumula es on he
senso PC is con inuously backed up o he mass da a managemen
(MDM) sys em implemen ed in he IT in as uc u e o each ship in
he amewo k o he Unde way Resea ch Da a p ojec . The MDM
sys em collec s all eco ded esea ch da a in a uni o m, campaign-
based manne and con inuously mi o s i on o sui able anspo
ha d d i es.
FIGURE 1
ADCP da a wo kflow as implemen ed and es ablished on he Ge man esea ch essels R/V Ma ia S. Me ian, R/V Me eo , and R/V Sonne (see Sec ion
3 o de ails).
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g04
In he wo kflow’s s anda d se ing o ou ine unde way deep-
wa e p ofiling, VMDAS is configu ed o use na owband single-
ping eco ding wi h bo om- acking disabled. Na owband
measu emen s a e less sensi i e o dis u bances and noise o low
abundances o sca e e s in he wa e column. The p ofile
pa ame e s (numbe o cells/cell size/blanking dis ance) o se he
sampling dep hs a e chosen in he ange o he ecommended
de aul alues p o ided by he manu ac u e o 75 kHz (100/8 m/8
m) and 38 kHz (50/32 m/16 m) sys ems.
VMDAS gene a es h ee, pa ly edundan aw da a o ma s
con aining single-ping da a in pd0 o ma , each ep esen ing a
di e en s a us o in e nally applied coo dina e ans o ma ion and
da a s eam me ging. ENR files con ain eloci ies in beam coo dina es
bu no ex e nal na iga ion da a. ENS files addi ionally con ain
na iga ion da a o each single ping. ENX files con ain eloci y da a
in ea h coo dina es and na iga ion da a. Using ENR o ENS files o
s a pos -p ocessing, he equi ed coo dina e ans o ma ions (RD
Ins umen s, 2010) a e o be conduc ed by he use . In case o ENR
files, he use also me ges he ancilla y da a s eams wi h he ADCP
single-ping da a. S a ing wi h ENR o ENS files allows edi ing o
single-ping da a o indi idual beams and gi es con ol o he use on
how o pe o m he coo dina e ans o ma ion (e.g., using a h ee-
beam-solu ion i one beam is ound mal unc ioning). S a ing wi h
ENX files, he use needs o speci y he senso s in he da a acquisi ion
se up ha a e o be conside ed as a i ude da a sou ces by he
VMDAS in e nal coo dina e ans o ma ion.
In addi ion o being inse ed in o he ENS and ENX files, he
na iga ion da a is logged in sepa a e files. Bina y o ma na iga ion
da a is collec ed in NMS files, while N[1,2,3]R files con ain aw
NMEA messages as configu ed in VMDAS.In hewo kflow’s
s anda d se ing, NMEA posi ion da a (GGA) and NMEA ship
speed da a (VTG) a e collec ed in N1R files, and NMEA a i ude
(heading, pi ch, oll) da a (PRDID) a e collec ed in N2R files.
Addi ionally, PADCP messages wi h he ADCP ensemble numbe
and a imes amp a e con ained in he N[1,2,3]R files. These can be
used o align na iga ion da a wi h he ADCP ime se ies, i he e was
a p oblem wi h eco ding o he na iga ion da a in he ENS o ENX
files. VMDAS elies on he ime se ings o he eco ding PC when
c ea ing imes amps o pings, so i is ad isable o keep he PC ime
synch onized wi h UTC ime. None heless, PADCP messages
p o ided by VMDAS con ain in o ma ion on he di e ence
be ween UTC and PC ime, ha can be used o co ec ping
imes. VMDAS p io i izes inse ing PADCP messages in o he N
[1,2,3]R files, which migh occasionally co up o he NMEA
messages and complica es pa sing du ing pos -p ocessing. Also,
in he pas , whole sec ions o na iga ion da a could be co up ed by
se ial da a o e flows, bu wi h mode n IT in as uc u e his is
no mally no longe a p oblem.
3.2 Da a ans e
3.2.1 Nea eal- ime
Fo sys em diagnos ics and da a quali y checks, incoming
VMADCP aw da a is egula ly sen asho e. A c onjob is
wa ching he di ec o y o VMADCP aw da a on he MDM.
E e y hou , newly c ea ed ENX files (usually 10 MB in size) a e
iden ified, comp essed and sen as a sepa a e e-mail including me a
in o ma ion and a checksum. Asho e, da a a e ecei ed by he O2A
inges se ices o aw da a (Koppe e al., 2015) hos ed a Al ed-
Wegene -Ins i u e (AWI) in B eme ha en, Ge many and s o ed in
he p ojec ’s cen al wo kspace.
3.2.2 Pos -campaign ans e
In he des ina ion po be ween wo campaigns, he dedica ed
anspo ha d disks wi h a comple e copy o he expedi ion da a a e
emo ed om he MDM and sen o AWI as an insu ed shipmen .
In e u n, he ship ecei es emp y ha d d i es o he nex
campaign. A AWI, he en i e da a se is checked o in eg i y
and ed in o he cen al wo kspace using he O2A inges (Koppe
e al., 2015) and is made accessible o delayed-mode p ocessing and
final quali y con ol ia he p ojec ’s cen al wo kspace.
3.3 P ocessing and quali y con ol
using OSADCP
The Py hon-based so wa e OSADCP (Kop e e al., 2024)is
de eloped bo h o he nea - eal- ime moni o ing and o he
delayed-mode p ocessing o he VMADCP aw da a. OSADCP
con ains modules ha include he essen ial p ocessing s eps
(Figu e 2) o coo dina e ans o ma ion, posi ion da a e ifica ion,
eloci y da a cleaning, bo om in e e ence de ec ion, ensemble
a e aging and wa e - ack calib a ion. I configu ed in VMDAS,
also bo om- ack eloci ies can be used o he calcula ion o he
eloci y p ofiles. Cus omized add-on modules ha e been de eloped
o he wo kflow p esen ed he e o au oma e o example he
gene a ion o da a p ocessing epo s, da a isualiza ions, o da a
o ma con e sions o dissemina ion p ocedu es.
3.3.1 Abou OSADCP
OSADCP eme ged om he MATLAB-based OSSI oolbox,
which was de eloped o e many yea s a GEOMAR, Helmhol z
Cen e o Ocean Resea ch Kiel. As pa o he DAM Unde way
Resea ch Da a p ojec , i was edesigned and u he de eloped o
mee he equi emen s o s anda dized and, as a as possible,
au oma ed p ocessing o VMADCP da a om Ge man esea ch
essels. To acili a e access o he whole scien ific communi y,
OSADCP is w i en in Py hon. The so wa e is designed o p ocess
VMADCP aw da a acqui ed wi h Teledyne RDI’s da a acquisi ion
so wa e VMDAS. A basic amewo k o he so wa e is made
a ailable o he scien ific communi y along wi h his a icle (see
Appendix o ins uc ions on he ins alla ion and se up). By de aul ,
OSADCP p ocesses ENX files, which equi es bo h high-quali y and
comple e ADCP aw da a and na iga ion da a. F om ou
expe ience, wo king wi h ENX files is gene ally p e e ed due o
he simplici y o he handling. In he e en ha p oblems a ise wi h
ei he ADCP o na iga ion da a, ha equi e o example edi ing o
beam-wise ADCP eloci ies o he manual me ging o na iga ion
da a, OSADCP has add-on modules p ocessing ei he ENR o ENS
files ha a e no ye made publicly a ailable.
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g05

The gene al p ocedu e o wo king wi h OSADCP is shown in
Figu e 3. To illus a e key aspec s o he p ocessing we use an
example da a se om a 75 kHz Ocean Su eyo ADCP o he R/V
Me eo campaign M189 (Dengle e al., 2023) ha was acqui ed
acco ding o he wo kflow p esen ed he e.
3.3.2 os_se ings.py
A cen al poin o he oolbox is he module os_se ings.py,
which uns a basic GUI (Figu e 4). This GUI is used o c ea e and
modi y a configu a ion json file o he p ocessing. I makes use o
he json dic iona y json/ essel_adcp_lib a y.json, whe e me a-
in o ma ion on essel-specific ADCPs a e s o ed (see Appendix).
The configu a ion file is sa ed o json/OSCONFIG.json. A lis o
files o be p ocessed is sa ed o he chosen p ocessing di ec o y,
along wi h a copy o he configu a ion file o documen a ion
pu pose. Bo h he configu a ion file json/OSCONFIG.json and he
lis o files will be used by all o he OSADCP p ocessing modules
(see Figu e 3).
I is possible o load an exis ing configu a ion ei he om he
wo king di ec o y json/OSCONFIG.json o om any p ocessing
FIGURE 3
Wo kflow o he OSADCP p ocessing oolbox. The main Py hon modules a e depic ed by da k blue boxes. Files s o ing majo p ocessing in o ma ion
a e depic ed in g een boxes, ADCP da a (bina y pd0, in e media e ne CDF files and final ne CDF files) a e ep esen ed by yellow boxes, while files
documen ing he p ocessing a e shown by o ange boxes.
FIGURE 2
Gene al p ocessing chain o essel-moun ed ADCP da a. O ange (g een) colo s ep esen s eps ca ied ou on he single-ping (ensemble-a e age)
da a (see Sec ion 3.3 o de ails).
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g06
di ec o y [p oc_di ]/OSCONFIG_[adcp]_[ essel]_[leg].json, which
is a common s ep when efining a da a se o esuming wo k on an
olde da ase .
An al e na i e way is o manually edi he file json/
OSCONFIG.json di ec ly and hen upda e he lis o files and he
configu a ion copy in he p ocessing di ec o y by execu ing he
Py hon module os_upda e_filelis .py.
3.3.3 os_ ead_enx.py
The Py hon module os_ ead_enx.py is execu ed o loop o e all
files lis ed in [p oc_di ]/OSFILELIST_*. x and ex ac all VMADCP
aw da a and na iga ion da a om he bina y ENX files. Fo each
bina y aw da a file, a co esponding ne CDF file (Unida a, 2021)is
sa ed o he p ocessing di ec o y, con aining all da a on he single-
ping le el (Figu e 3). These files can be accessed o any use -specific
edi ing on single-ping le el no co e ed by he OSADCP s anda d
ea men be o e con inuing on he ensemble-a e age le el.
A majo ask ca ied ou by he module is he e ifica ion o he
na iga ion da a me ged in o he ENX files. Common p oblems include
he occu ence o ze o/ze o posi ions when no eal da a is a ailable due
o p oblems o he na iga ion senso and i egula i ies in he ime
alloca ion such as ime s ops, backwa d ime jumps, ime shi s e c.
A ec ed pings a e flagged acco dingly and igno ed in u he p ocessing.
In he case o e y poo na iga ion da a, he manual inse ion o ex e nal
na iga ion da a on single-ping le el should be conside ed.
3.3.4 os_edi _bo om.py
An impo an aspec o he p ocessing conce ns he scanning o
echo eedback om he g ound, which in oduces spu ious eloci ies
in he a ec ed cell ange. This ask can be au oma ed only o a limi ed
ex en since he echo in ensi y p ofile used o he iden ifica ion is
a ely unambiguous. I can be a ec ed by acous ic in e e ence in he
wa e column o he occu ence o s ong sca e ing laye s, o
example, ha may lead o alse bo om de ec ion. Fu he mo e,
nea he sea bed eloci ies a e po en ially co up ed by sidelobe
in e e ence. Depending on he beam angle qand he dis ance z o
he ADCP ansduce (igno ing pi ch and oll il s), eloci ies in he
ange zshadow =z(1−cos q) abo e he bo om, he so-called shadow
zone, a e hen con amina ed by he sidelobe echo e u ning ea lie o
he ansduce han he slan ed main lobe. The de ec ion and emo al
o spu ious eloci ies benea h he shadow zone is a p e equisi e o
he subsequen wa e - ack calib a ion, especially i he shadow zone
dep h o e laps wi h he e e ence laye chosen o he calib a ion (see
os_wa e ack.py below).
OSADCP p o ides he module os_edi _bo om.py ha loops
h ough all da a files and le s he use decide whe he a bo om
signal is p esen and should be ea ed based on a isual inspec ion
o he ime se ies o he beam-a e aged echo in ensi y (Figu e 5, le
panel). As he co esponding eloci y bias in he shadow zone is
mos ob ious in he along- ack eloci y while he ship is unde way,
i is shown as assis ing in o ma ion in he igh panel o he
in e ac i e plo p o ided by os_edi _bo om.py. To isualize a
po en ial eloci y bias, he median along- ack eloci y is emo ed
o each ping. Using he in o ma ion p o ided in Figu e 5, a line
co esponding o he dep h o bo om influence on he measu ed
eloci ies is picked manually, adjus ed as equi ed and finally sa ed
as a sepa a e ne CDF file wi h a *_bo .nc ex ension added o he file
naming con en ion (Figu e 2).
FIGURE 4
GUI con olling ADCP da a p ocessing in OSADCP. The GUI is execu ed ia os_se ings.py.
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g07
3.3.5 os_wa e ack.py
The p ima y module o he so wa e o de i ing deep-wa e
eloci y p ofiles is called os_wa e ack.py. I ca ies ou impo an
s eps conce ning he de i a ion o ship speed, he cleaning and
a e aging o single-ping da a, and he es ima ion and applica ion o
he eloci y calib a ion (Figu e 2). Fo he ea men o single-ping
da a, he module loops o e all da a files c ea ed by os_ ead_enx.py
du ing he bina y da a con e sion (Figu e 3). A e a e aging each
da a se sepa a ely, he module ca ies ou he wa e - ack
calib a ion and calcula ion o wa e eloci ies based on he
ensemble-a e aged da a o all da a se s as specified
by os_se ings.py.
The ship speed is calcula ed ia cen al di e ences based on he
GNSS posi ion da a, igno ing only pings wi h ques ionable
na iga ion da a as flagged by os_ ead_enx.py. I he configu a ion
file json/OSCONFIG.json con ains le e a m in o ma ion o he
gi en se up o ADCP and GNSS senso , he GNSS posi ions a e
co ec ed p io o he calcula ion o ship speed as ollows:
lonadd = 1
1:11x105mcos (la )( cos (H)(xADCP −xGPS)
+  sin (H)(yADCP −yGPS))
la add = 1
1:11x105m(−sin (H)(xADCP −xGPS) +  cos (H)(yADCP
−yGPS))
wi h Hbeing he ship’s heading, and xADCPjGPS  and yADCPjGPS
being he posi ions o he ADCP ansduce and GNSS an enna
ela i e o he ship’s cen e poin along he mino and majo axes o
he ship, espec i ely (x: s a boa d, y: o wa d). This co ec ion
imp o es cu en eloci ies du ing ship maneu e s when he
heading changes apidly. In he example (Figu e 6), he dis ance
be ween ADCP and GNSS an enna was abou 15 m. Single-ping
da a we e combined in o 1-minu e ensembles and he
app oxima ely 3-hou exce p con ains h ee sec ions wi h ab up
changes o di ec ion o he ship a a ound 06:30, 07:50, and 08:55
UTC (Figu e 6A). Wi hou le e a m co ec ion, he de i ed
di ec ion o he ocean cu en p ofiles du ing hese sec ions is
no ably di e en om he emaining ime, whe e he ship’s heading
is ei he cons an o changes smoo hly (Figu e 6B). Geome ic
co ec ion o he le e a ms hen imp o es he measu ed
di ec ion o ocean cu en s, he e elimina ing bands o e oneous
cu en di ec ion o he h ee sec ions (Figu e 6C).
os_wa e ack.py applies a numbe o au oma ed cleaning
c i e ia o he single-ping eloci y da a:
1. E o eloci y: La ge e o eloci ies indica e a non-
homogeneous flow field o la ge eflec o s like fish, ha
canno be assumed o mo e passi ely wi h he cu en .
Du ing he cleaning all cells a e masked, whe e he de i ed
cell e o eloci y exceeds wice he s anda d de ia ion o
he e o eloci y.
2. Doub ul heading da a: This c i e ion deals wi h un ealis ic
ship heading da a by masking en i e eloci y p ofiles, whe e
he ping- o-ping heading change exceeds 10°.
3. Doub ul cu en eloci ies: Fo his c i e ion, he single-
ping cu en eloci y is app oxima ed by sub ac ing he
ship eloci y componen s om he measu ed eloci ies in
he e e ence laye . Subsequen ly, en i e eloci y p ofiles a e
masked, ei he whe e he ping- o-ping change in ei he
eloci y componen exceeds 2 m s
-1
,o whe eei he
eloci y componen exceeds 5 m s
-1
.
4. Acous ic in e e ence: This is an op ional cleaning c i e ion
ha can be enabled ia os_se ings.py, wi h addi ional
con ol on he dep h ange conside ed o he cleaning
(Figu e 4). To de ec spikes associa ed wi h acous ic
in e e ence, os_wa e ack.py fi s cons uc s an echo
FIGURE 5
In e ac i e plo execu ed by os_edi _bo om.py, o pick he a ea a ec ed by bo om in e e ence. The le panel shows he a e age o he echo
in ensi y measu ed by he ou beams. The igh panel shows he along- ack eloci y componen wi h he median emo ed o each ping. The
along- ack componen is displayed, which mos p ominen ly shows a po en ial eloci y bias caused by bo om in e e ence, in case o da a
collec ed unde way. The echo in ensi y local maximum in he le panel co esponds o he bo om dep h, while he ansi ion owa ds high-biased
along- ack eloci ies in he igh panel ma ks he dep h o sidelobe in e e ence by he bo om e u n echo. The ed line is picked based on he
in o ma ion p o ided by he wo panels. In subsequen p ocessing s eps, cells below he ed line a e igno ed.
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g08
in ensi y anomaly field by sub ac ing he median echo
in ensi y p ofile om each echo in ensi y p ofile con ained
in he da a file. Cells po en ially a ec ed by in e e ence a e
hen de e mined by iden i ying ping- o-ping di e ences
ha ei he exceed o all below +0.7 o -0.7 o he o al
s anda d de ia ion o ping- o-ping di e ences, espec i ely.
The candida e cells a e hen checked i hey a e o single-
ping du a ion and whe he he associa ed in ensi y spikes
ex end o e se e al dep h cells. All cells ha a e iden ified
asbeinga ec edbyacous icin e e encea e
subsequen ly masked.
5. Bo om and sidelobe in e e ence: To mask cells a ec ed by
he bo om e u n echo, os_wa e ack.py eads he
co esponding *_bo .nc files c ea ed by os_edi _bo om.py
and applies he gene a ed bo om in e e ence mask o he
single-ping eloci y da a.
Subsequen ly, single-ping da a along he wa e column is
a e aged in o so-called ensembles, wi h he eloci ies being
ec o -a e aged. Using ensemble a e ages educes he sp ead o
single-ping cu en es ima es, inc easing he p ecision o he
measu emen . In os_wa e ack.py he de aul a e age in e al is
1 minu e.
The wa e - ack calib a ion implemen ed in os_wa e ack.py
add esses wo di e en e o s (Joyce, 1989;Fi ing and
Hummon, 2010):
1. Misalignmen e o : A de ia ion o he ansduce
alignmen wi h espec o he heading e e ence o he
ship in oduces a bias wi h i s main e ec being a spu ious
c oss- ack eloci y componen p opo ional o ship speed.
2. Scaling e o : Small e o s in he beam geome y o a non-
ze o im o he ansduce o ship can cause a sys ema ic
bias a ec ing mos ly he along- ack eloci y componen ,
p opo ional o ship speed.
The calib a ion aims a minimizing hese e o s by de e mining
a misalignmen angle co ec ion aand a scale ac o co ec ion b,
FIGURE 6
E ec o le e a m co ec ion on he cu en di ec ion. The da a se was gene a ed by a 75 kHz Ocean Su eyo ADCP du ing he M189 campaign o
R/V Me eo (Dengle e al., 2023). (A) Ship heading on May 9, 2023. (B) Di ec ion o he uppe -ocean cu en s as 1-minu e ensemble a e ages
wi hou le e a m co ec ion applied, (C) same as (B), bu wi h le e a m co ec ion applied du ing pos -p ocessing. The dis ance be ween he ADCP
ansduce and he GNSS an enna was app oxima ely 15 m along he main axis o he ship.
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g09
Appendix - OSADCP ins alla ion
and se up
OSADCP equi es a wo king and easonably cu en Py hon3
ins alla ion. Depending on he ope a ing sys em, an ins alla ion o
QT5 migh be equi ed. Al hough i is pe ec ly possible o ins all
OSADCP wi hin he sys em’s main Py hon ins alla ion, some kind
o isola ion, such as a i ual en i onmen , is ecommended.
OSADCP is buil p ima ily wi h he ollowing open-sou ce
packages: numpy, scipy, PyQ 5, ma plo lib, ne CDF4 (in o de :
Ha is e al., 2020;Vi anen e al., 2020,Ri e bank Compu ing,
Hun e , 2007;Unida a, 2021).
The OSADCP oolbox is published ia a Gi eposi o y a
h ps://gi .geoma .de/dam/osadcp_ oolbox. I can be downloaded
as zip a chi e and ex ac ed locally. Al e na i ely, i he use has Gi
ins alled, he ollowing command can be used o download he
oolbox: gi clone h ps://gi .geoma .de/dam/osadcp_ oolbox.gi .
Subsequen ly, OSADCP’sdependenciesasspecified in he
equi emen file need o be ins alled ia pip ins all -
equi emen s. x . Please e e o he gi eposi o y o he la es
e sion o he oolbox and o upda ed ins alla ion ins uc ions and
use manuals.
A e se ing up he oolbox, me a-in o ma ion on ADCPs and
essels, such as se ial numbe s, le e a ms e c., can be added o he
json-dic iona y json/ essel_adcp_lib a y.json. This dic iona y
manages ADCP and essel-specific me ada a cen ally and is used
by he p ocessing managemen module (Sec ion 3.3) o e ficien
alloca ion and o wa ding.
Kop e e al. 10.3389/ ma s.2024.1425086
F on ie s in Ma ine Science on ie sin.o g16