Ba k a e -A aly e 
14 
Hydrographische Nachrichten 
Fteferences 
Föote, kfenneth G; Dezhang 
Chu; Terence R Ham mar; 
kenneth Baldwin;Larry 
A. Mayer; Lawrence C. 
HufnagleJr.; Michael 
Jsch ( 005): Protocols for 
calibrating multibeam 
9anar;TheJournä of the 
Acoustical Sbdety of 
America, \£>l. 7 No. 4, pp. 
03-0 7 
Lamarche,Geo roy; Xavier 
Lurton;Anne LaureVferdler; 
Jean Marie Augustin 
(0 ^Quantitative 
characterisation of sea oor 
substrate and bedforms 
using advanced processing 
of multibeam backscatter- 
Application to Gx>k 3rät, 
New Zealand;Continentä 
Snelf Ffesearch,Vbl.3 ,No. , 
pp.93- 09 
Lan zon I, Jose G; Thomas 
G V\feber( 0 0). High 
resolution cäibratlon of a 
multibeam echo sounder; 
Oceans 0 0,1^^ Seattle, 
pp. -7 
Fig. 6: Decision of the angular 
compensated mean BSIn 
the 5to60degreesangular 
interval for all operational 
Transmission Modes of the 
BVI30 of OOGS»Amundsen« 
the Kcngsberg real time data reduction process 
from the angular compensated BSimage in order 
to obtain an angular uncompensated BSimage. 
9ep consists in modelling the BS angular 
dependence visible in the angular uncompen 
sated BS image. The Generic Sea oor Acoustic 
Backscatter (GSM^ model (Lamarche et al. ) is 
used for that purpose. 
In step , the new »best« modelled estimate of 
the BS angular dependence is applied and the 
Kcngsberg default transmission sector pattern 
compensation is removed. This process isolates 
the sector patterns, which can be »seen« in the 
backscatter image (Rg. a). 
Rnally,step consistsin modelling thetransmis 
sion sector patterns as rst order polynomials, an 
approximation of the mathematically correct sine 
function. Three parameters constitute the vari 
ablesof the polynomial functionsthe source level, 
the transmission angle and the sector opening 
angle. Rg. b shows the corresponding angular 
response of the residual colour coded transmis 
sion sector beam patterns The points represent 
the mean source level residual per ° bin and the 
continuous lines represent the modelling of the 
transmission sector patterns 
Ftesults 
The newly modelled transmisson sector patterns 
can be applied to the survey datato assessthe qual 
ity of the calibration. Rg. ashowsthe angular com 
pensated BS image in ping vs beam geometry for 
the Shallow, Dual, CW Transmission Mode with the 
new transmisson sector patterns and GSAB models 
applied. Rg. b isthe correqconding mean BSangu 
lar reqeonse. Wiile some modulation is still viable in 
the BSimage, espedally at the sector boundaries, the 
overall predsion of the mean BSangular response is 
signi cantly improved. Rg. illustrates the predson 
improvements of the angular compensated mean 
BSinthe to degreesangular interval for all op 
erational Transmisson Modes The latter interval is 
chosen so asto exdude the nadir and grazing angle 
zones, which alwsyscomprise less reliable data 
The modelled transmission sector patterns re 
suiting from the calibration can be utilised in two 
ways Rrst, the valuesof the modelscan be written 
inthebs .x le and injected to the EM echo 
sounder. This will ensure that all newly collected 
data will be exempt from any modulating e ect 
due to improperly calibrated transmission sector 
patterns Second, the models can be applied to 
existing backscatter data in order to mitigate the 
intra line artefacts Rg. a illustrates the identical 
mosaic as in Rg. b with the transmission sector 
patterns properly calibrated. Aside from a BSmini 
mum in the nadirarea, no further along track arte 
facts are visible. 
Several attempts were undertaken to discover 
the reason forthe remaining inter linebiasin Rg. a 
This dB bias between the survey line collected 
on September nd and the one collected on Sep 
tember th could not be explained by changes to 
environmental factors(changesto the seabed type 
or to the water column propertied The only con 
elusive explanation stems from the software insta 
bilitiesexperienced during the surveys These were 
correlated to an incompatibility between hardware 
and software versionswhich necessitated the com 
pletere installation of the data collection software 
after the rst and before the second survey. This 
type of problem cannot be accounted for in post 
processing other than by applying a statistical an 
gular correction whose result isillustrated in Rg. b. 
Although the backscatter mosaic now appearsuni 
form, one cannot speak of absolute BSIevelssince a 
bias has been introduced. 
Discussion and future improvements 
Wiilethe proper calibration of transmission sector 
patterns is a step towards absolute and properly 
referenced backscatter data, transmission sector 
patterns are not the only factor that needs to be 
addressed. Other factors include: 
• proper compensation for absorption of the 
acoustic signal in thewater column; 
• accounting for the topography of the sea 
bed when calculating the ensoni ed area; 
• possible compensation for the receive beam 
patterns; 
• the need for a bathymetry surface except of 
outliersand biases 
Several of these factorscould not beorweresim 
ply not fully addressed during the calibration survey 
and subsequent data analysis Logistical and time 
constraints help explain why the calibration survey 
could not be performed under ideal conditions 
Gven the multi disciplinary nature of the research 
conducted on OOGS »Amundsen«, this is to be ex 
pected. Limitations in the work ow used to transfer 
the raw data collection measurementsto abathym 
etry processing software and nally to SbnarStxipe 
are also a root cause for sub optimal results 
However, the single greatest weakness of the 
calibration procedure isthe reliance on the mod 
elled BSangular response from the empirical data.