Abstract:
Wave climate statistics are fundamental for navigation, designing near/offshore structures
and assessment of coastal hazards. Together with satellite altimeters and buoys, wave hindcasts
are the typical source of wave climate data due to their temporal and spatial availability.
However, most hindcasts provide wave data as an unimodal set of integrated parameters,
even when the sea state is composed of two or more wave systems – which is often the case
in New Zealand waters. Furthermore, in the context of climate change, statistics based on
current wave data might not represent the future wave climate.
This PhD project consisted of generating and studying state-of-the-art multimodal wave climate
databases of New Zealand. In order to achieve this, we also developed novel techniques
of multimodal wave height correction and wave climate analysis among others.
The calibration technique here developed outperformed the standard methods of wave height
correction and provided insightful information on the systematic errors present in the wave
simulations around New Zealand. This technique is of great use for the wave modelling
community as it allows wave height correction along areas where buoy measurements are
unavailable.
The databases produced are a wave hindcast and past/projected wave data from global
circulation models (GCM) ensembles. The hindcast consists of downscaled partitioned
WAVEWATCH-III results from a 20-year (1993–2012) global wave model. The boundary
forcings extracted from the global simulation were calibrated using the aforementioned wave
height correction method. The multimodal wave spectra of each boundary were reconstructed
from the partitions and used as wave forcings of a SWAN grid encompassing New Zealand.
Waves were downscaled in non-stationary mode and had both partitioned and integrated
parameters stored. The partitioned hindcast of New Zealand allowed for the development of a wave climate
analysis framework that exploits and displays characteristics of wind-sea and swell waves
unseen in previous studies. Such insightful information provided a better understanding
of the wave climate of this complex area, and also allowed to identify how atmospheric
anomalies modulate the wind-sea and swell waves around New Zealand.
The GCM simulated wave climate consists of one 13-year (1993–2006) and two 20-year
(2026–2047, 2080–2101) downscales of past and projected wave climate from three GCMs
under two representative concentration pathways (RCP4.5, RCP8.5). The boundaries of
each downscale grid were also reconstructed from bimodal wave data, which required the
development of a data-driven technique based on the “k-nearest neighbour” algorithm to
generate directional spread data.
The assessment of the anomalies between the past and future simulations of GCMs ensembles
produced thoughtful information about the potential changes in the future wave climate, its
main drivers and how it may affect New Zealand’s coasts. Such projection databases, together
with the hindcast, are crucial for coastal management, risk assessments and climate-change
adaptation.