Regional Broad-Band OBS Experiments in a Global Context

Regional passive seismic experiments typically concentrate on
the seismic structure within a recording array where quite often 
the analyzed seismic waveforms come from teleseismic sources. Analysis
techniques are then applied that minimize (or correct for) the sensitivity
of the collected seismic dataset to structure outside of the recording 
array. The information the dataset carries about source-receiver structure
is then lost to the data analyst.

On the other hand, numerous attempts have been made to (re-)combine regional
models to obtain the 'big picture' of global or continental-scale 
seismic structure. Up to this day, it appears quite difficult
to embed regional models in global ones due to incompatibilities at
wavelengths that are common to both types of models.
A more direct way of making use of the sensitivity of 'regional datasets'
to global structure is to include them directly in an fine-scale inversion for 
global structure. This has not been feasible in the past but increasing
computer and software resources now encourage us to assess this issue.
Surface wave datasets are particularly desirable to combine because wave
propagation is restricted to two dimensions and hence limiting the model space. 

In this study, we investigate the dataset collected at the 1997/98 SWELL pilot
experiment at the southwestern edge of the Hawaiian Swell. This dataset has been
proven an extremely rich one as we have demonstrated at various
meetings and publications. We were able to image the seismic heterogeneity
and variations in azimuthal anisotropy at the edge of the Hawaiian
Swell. In a global context, this dataset strongly constrains structure
outside of the SWELL array as well. The most prominent change,
compared to pre-existing global surface wave dispersion maps, is that low
velocity anomalies are now concentrated along the Hawaiian island
chain, to the northeast of the SWELL array. We discuss the
great potential of improving the resolution of images in the oceans
with similar passive seismic experiments or proposed permanent
stations of the OSN. We perform realistic resolution tests in which we
adopt the measurement errors gathered at already existing island stations and
at OSN1.