Deep-Ocean Tides in the South-West Indian Ocean . . .
159
(panels d, e, i and n) these measurements confirm the dominance of surface tides
in the time-varying pressure field. At other locations, however, mid-water column
pressure signals frequently suffer from severe blow-down. This is a common feature
for pressure measurements within the water column, and for this reason often only
BPRs are used in the validation of altimetry-derived harmonic constants [34]. In
section “Coherent Surface and Internal Tides”, we will therefore first look at BPRs,
but, since we also want to analyse pressure records higher up in the water column
(section “Deep Versus Mid-depth Pressure-Measurements of Tides”), we subject all
our measurements to the following procedure.
The amplitudes and phases resulting from MT_TIDE are assigned to the central
time within each one-year time interval. Those of the most important semidiurnal and
diurnal components are discussed in the next sections. The tidal variance estimate,
obtained by averaging the variance of subsequent T_TIDE analyses, and its relative
contribution to the total HP-variance, are shown in Table 4. The total and tidal variance are all O (1 m 2 ). For all moorings, a weak increase of the total HP-variance with
height above the bottom is noticed. Except for the most Eastern mooring, lmc8, the
tidal variance also increases with height above the bottom. This may be related to
the stronger presence of coherent internal tides higher up in the water column, while
being weaker or absent in the bottom boundary layer. Yet, the relative contribution
of the tides to the total variance always decreases towards the surface. This may be
attributed to the stronger (oceanic and atmospheric) mesoscale activity near the surface, introducing more phenomena influencing the pressure variability. Note that the
tidal variance in Table 4 that T_TIDE computes for each one year period is based on
many more (High-Passed) tidal components than those listed in Table 1, including
higher harmonics, the so-called shallow water tides [28]. But in the following we
will focus on just the major eight (semi)diurnal components.
East Madagascar
East of Madagascar, tides are much weaker than in the Mozambique Channel, at
least near the bottom where they measure free surface displacements of O(10 cm)
for the main tidal component, M2. Figure 5a, c show BPR (black line), tidal backprediction (grey) and residual (red) at moorings EMC1 and EMC3, respectively (see
Table 3). Notice that at these two instruments the total apparent surface displacement
is restricted to about 1 m. The other three instruments in Fig. 5b, d, e, all far above
the bottom, indicate unrealistically large vertical excursions that are a hundred times
larger (notice the difference in scale). These are obviously due to mooring motions,
resulting in occasional blow-down.
Harmonic Analysis of the BPRs at moorings EMC1 and EMC3 shows that the
tides give a modest, yet genuinely tidal contribution of O(8–32%) to the total variance (Table 5). Indeed, this estimate of tidal variance looks reliable in Fig. 5, when
comparing the tidal string of pearls (grey) to the variance carried by the residual
(red). As to the question why these values are so low (compared to those in MC),
159
(panels d, e, i and n) these measurements confirm the dominance of surface tides
in the time-varying pressure field. At other locations, however, mid-water column
pressure signals frequently suffer from severe blow-down. This is a common feature
for pressure measurements within the water column, and for this reason often only
BPRs are used in the validation of altimetry-derived harmonic constants [34]. In
section “Coherent Surface and Internal Tides”, we will therefore first look at BPRs,
but, since we also want to analyse pressure records higher up in the water column
(section “Deep Versus Mid-depth Pressure-Measurements of Tides”), we subject all
our measurements to the following procedure.
The amplitudes and phases resulting from MT_TIDE are assigned to the central
time within each one-year time interval. Those of the most important semidiurnal and
diurnal components are discussed in the next sections. The tidal variance estimate,
obtained by averaging the variance of subsequent T_TIDE analyses, and its relative
contribution to the total HP-variance, are shown in Table 4. The total and tidal variance are all O (1 m 2 ). For all moorings, a weak increase of the total HP-variance with
height above the bottom is noticed. Except for the most Eastern mooring, lmc8, the
tidal variance also increases with height above the bottom. This may be related to
the stronger presence of coherent internal tides higher up in the water column, while
being weaker or absent in the bottom boundary layer. Yet, the relative contribution
of the tides to the total variance always decreases towards the surface. This may be
attributed to the stronger (oceanic and atmospheric) mesoscale activity near the surface, introducing more phenomena influencing the pressure variability. Note that the
tidal variance in Table 4 that T_TIDE computes for each one year period is based on
many more (High-Passed) tidal components than those listed in Table 1, including
higher harmonics, the so-called shallow water tides [28]. But in the following we
will focus on just the major eight (semi)diurnal components.
East Madagascar
East of Madagascar, tides are much weaker than in the Mozambique Channel, at
least near the bottom where they measure free surface displacements of O(10 cm)
for the main tidal component, M2. Figure 5a, c show BPR (black line), tidal backprediction (grey) and residual (red) at moorings EMC1 and EMC3, respectively (see
Table 3). Notice that at these two instruments the total apparent surface displacement
is restricted to about 1 m. The other three instruments in Fig. 5b, d, e, all far above
the bottom, indicate unrealistically large vertical excursions that are a hundred times
larger (notice the difference in scale). These are obviously due to mooring motions,
resulting in occasional blow-down.
Harmonic Analysis of the BPRs at moorings EMC1 and EMC3 shows that the
tides give a modest, yet genuinely tidal contribution of O(8–32%) to the total variance (Table 5). Indeed, this estimate of tidal variance looks reliable in Fig. 5, when
comparing the tidal string of pearls (grey) to the variance carried by the residual
(red). As to the question why these values are so low (compared to those in MC),