共查询到20条相似文献，搜索用时 671 毫秒
1.
GWENAëLE JAN YVES MÉNARD MATHILDE FAILLOT FLORENT LYARD ERIC JEANSOU PASCAL BONNEFOND 《Marine Geodesy》2013,36(3-4):615-629
Absolute calibration of sea level measurements collected from space-borne radar altimeters is usually performed with respect to collocated sea level in situ records from tide gauges or GPS buoys (Ménard et al. 1994; Haines et al. 1996; Bonnefond et al. 2003; Haines et al. 2003; Schum et al. 2003; Watson et al. 2003; Watson et al. 2004). Such a method allows regular and long-term control of altimetric systems with independent records. However, this approach is based on a single, geographically dependent point. In order to obtain more significant and accurate bias and drift estimates, there is a strong interest in multiplying the number of calibration opportunities. This article describes a method, called the “offshore method” that was developed to extend the single-point approach to a wider regional scale. The principle is to compare altimeter and tide gauge sea level data not only at the point of closest approach of an overflying pass, but also at distant points along adjacent satellite passes. However, connecting sea level satellite measurements with more distant in situ data requires a more accurate determination of the geoid and mean ocean dynamic topography slopes, and also of the ocean dynamical changes. In this demonstration experiment, 10 years of averaged TOPEX/Poseidon mean sea level profiles are used to precisely determine the geoid and the mean ocean circulation slope. The Mog2d barotropic ocean model (Carerre et Lyard 2003) is used to improve our estimate of the ocean dynamics term. The method is first validated with Jason-1 data, off Corsica, where the dedicated calibration site of Senetosa provides independent reference data. The method is then applied to TOPEX/Poseidon on its new orbit and to Geosat Follow On. The results demonstrate that it is feasible to make altimeter calibrations a few tens to hundreds of kilometers away from a dedicated site, as long as accurate mean sea level altimeter profiles can be used to ensure the connection with reference tide gauges. 相似文献
2.
A. Eslami M. Alimirzaei E. Aflaki H. Molaabasi 《Marine Georesources & Geotechnology》2017,35(1):62-79
For the past four decades, the CPT has played a key role in onshore and offshore soil investigations. One of the main applications of cone penetration test (CPT) is the soil behavioral classification. Most of the developed methods for soil identification using CPT and CPTu (piezocone) data are well categorized for common soils, such as clays, silts, and sands. Soils with low resistance or more compressibility generally involve problems in geotechnical engineering practice and construction projects. Consequently, these unusual deposits require further evaluation and more detailed data. Five major groups of problematic soils including: liquefiable, sensitive, peaty, collapsible, and expansive soils have been considered in this study. One hundred and forty CPT and CPTu test records were collected from fifteen countries. Sixty-one of the records are related to difficult soils. A brief comparison is performed for currently used soil behavioral classification charts, such as by Campanella et al. (1985), Robertson (1990), Jefferies and Davies (1991) and Eslami and Fellenius (1997). Analysis based on CPT data indicates that a few commonly used charts recognize relatively well problematic deposits. However, further studies are needed to increase the accuracy and capability of methods. Existing charts have some problems due to the limitations of the nature of rectangular charts based on two axes. A new format of classification chart, i.e., triangular form containing cone tip resistance (qc), sleeve friction (fs), and pore pressure (u2) is proposed for soil identification which can be realized in practice. The proposed chart with more accuracy and less scattering of data than the previous charts is able to identify soil types particularly for deltaic soils. 相似文献
3.
The FEMWATER model is a three-dimensional finite element model that can simulate both groundwater flow and mass transport at the same time, and so by using this FEMWATER model seawater intrusion was simulated and the results were compared to observation data. By using the observation data, a conceptual model was made of 3 layers based on the type of soil, and the hydrogeological parameter was selected. There are two types of dispersivities: longitudinal dispersivity and transverse dispersivity. Longitudinal dispersivity was calculated by Xu and Eckstein's method (1995) and the transverse dispersivity was put at 3/10 of the longitudinal dispersivity for the simulation. The electrical conductivity suggested by Freeze and Cherry (1979) and the correlation with total dissolved solids, TDS, was used to classify the state of the water. Upon analyzing the depths at 15 m, 30 m, 60 m, and 90 m, occurrence of seawater intrusion near the ground surface was limited to the vicinity of the coastline, but as the depths became deeper, seawater intrusion expanded inland. From these results we can come to the conclusion that in order to prevent seawater intrusion there needs to be sufficient preliminary examination. 相似文献
4.
It has been well documented that natural marine Ariake clays are sensitive clays. In this study, extensive data of marine Ariake clays are obtained to investigate the gravitational compression behavior for sensitive clays. Analysis results indicate that the compression behavior of remolded Ariake clays is not different from that of other remolded/reconstituted soils. But natural Ariake clays do not follow the gravitational compression pattern reported by Skempton (1970) for natural sedimentary soils. At a given value of effective overburden pressure, the void ratios of natural Ariake clays are almost independent of liquid limits. Most natural Ariake clays lie above the sedimentation compression line proposed by Burland (1990). When the liquid limit is larger than 90% and the ratio of natural water content over liquid limit ranges 0.8–1.1, the natural Ariake clays lie around the sedimentation compression line. In addition, the natural Ariake clay with higher value of the ratio of natural water content over liquid limit lies above the natural Ariake clay with lower value of the ratio of natural water content over liquid limit. Salt removal is the most probable cause for such a phenomenon. 相似文献
5.
M. Murat Monkul 《Marine Georesources & Geotechnology》2013,31(3):145-159
Earthquakes usually cause both inertial and kinematic loading of pile foundations. Inertial loading and lateral pile response can be predicted to a certain extent by lateral pile load tests for important engineering projects. However, there is no accepted in situ assessment method for pile behavior due to kinematic loading. Therefore, it is important for geotechnical engineering practice to validate the practice oriented models and examine the influence of soil strength parameters on the performance of kinematically loaded piles due to lateral spreading. For this purpose, a well-documented case of full-scale lateral spreading test conducted at Port of Tokachi in Japan by Ashford et al. (2006) is selected as a primary reference in terms of test data and site conditions. Four p-y analyses are performed using a special purpose software based on finite difference technique. The first two investigate the employment of different p-y curves suggested in the literature. The third one investigates the influence of variations in internal friction angle and undrained shear strength. In the fourth analysis, liquefied shear strength and residual shear strength approaches are utilized for the liquefiable layers and their effects are investigated. In all analyses, mobilized lateral pressures in non-liquefied layers are also scrutinized. The results of the analyses are discussed and compared with the measured values in the field. Finally, conclusions are given based on the analyses and discussions. 相似文献
6.
Fei Tong 《Marine Georesources & Geotechnology》2013,31(4):346-363
Due to the viscous nature, especially the swelling property, bentonite is selected as buffer and backfill material for marine radioactive waste disposal. This article presents experimental results from one-dimensional oedometric compression tests with multi-stage loading and unloading/reloading on specimens of bentonite mixed with different sand contents. The sand contents are 50%, 60%, 70%, 80%, and 90%, respectively, in terms of the dry mass of sand over the dry mass of the mixture. Influences of sand contents on time-dependent stress-strain behavior of the bentonite-sand mixture under 1-D straining are examined with special attention to the nonlinear creep and swelling behavior. It is found that the bentonite-sand mixtures exhibit distinct nonlinear creep and swelling behavior that cannot be neglected. Moreover, a nonlinear creep function proposed by Yin (1999) for Hong Kong Marine Deposits has been modified to describe the nonlinear creep behavior of the bentonite-sand mixture. 相似文献
7.
Zhen-Feng Qiu 《Marine Georesources & Geotechnology》2018,36(6):640-642
A closure is presented to the discussion of Anisotropic hydraulic conductivity and critical hydraulic gradient of a crushed sandstone–mudstone particle mixture by Salahou, Jiao, and Yousif (2017). The relationship between the critical hydraulic gradient and the uniformity coefficient can be limited to two curves. The internal erosion mechanism of soils with different particle grade curves is different. The maximum diameter of fine particles of different particle grade curves can be calculated by a lot of methods. The density of the original paper is the average density. The anisotropic hydraulic conductivity is caused by uneven density distribution. The hydraulic conductivities are tested on the saturated soils. The saturated hydraulic conductivity can be regard as a constant value for a saturated soil. 相似文献
8.
Isabel López L. Aragonés Y. Villacampa P. Compañ 《Marine Georesources & Geotechnology》2018,36(6):698-708
The paper describes the training, validation, testing, and application of models of artificial neural networks (ANN) for computing the cross-shore beach profile of the sand beaches of the province of Valencia (Spain). Sixty ANN models were generated by modifying both the input variables as the number of neurons in the hidden layer. The input variables consist of wave data and sedimentological data. To select and evaluate the performance of the optimal model, the following parameters were used: R2, absolute error, mean absolute percentage error, and percentage relative error. Finally, the results are compared with the numerical model proposed by Aragonés et al. (2016b) for the equilibrium profile in the study area. The results show a mean absolute error of 0.21?m compared to 0.33?m Aragones’ model, significantly improving the results of the numerical model in the bar area around de Valencia Port. In addition, when comparing the results with other methods currently used (Dean’s or Vellinga formulation), the errors of these compared to ANN are of the order of 167 and 1538% higher, respectively. 相似文献
9.
ABSTRACT The genus Xenostrobus consists of small, marine and estuarine mussels that all appear similar externally. One of its estuarine species, Xenostrobus securis, with a native range in New Zealand and Australia, has become invasive in the Northern Hemisphere. No genetic data are available to determine if X. securis populations from the two countries are conspecific. Additionally, marine Xenostrobus from New Zealand have often been regarded as a species, X. neozelanicus, distinct fromthe marine Australian species X. pulex. We combined new DNA sequences with published data to assess the taxonomic status of New Zealand Xenostrobus. The data comprised 658 aligned bases of mitochondrial cytochrome c oxidase subunit I (COI) gene and 331 bases of nuclear histone H3. There was no evidence that X. securis populations from Australia and New Zealand are specifically distinct. Northern Hemisphere specimens of X. securis belonged to Australian, not New Zealand, clades in phylogenetic analyses of COI data, suggesting the former country as their more likely original source. The results confirm that X. neozelanicus and X. pulex are distinct species and for nomenclatural completeness for this taxonomic decision a lectotype is designated for Mytilus ater Zelebor in Dunker and Zelebor, 1866 [?=?Modiolus neozelanicus Iredale, 1915]. 相似文献
10.
A large number of tar globules with sessile organisms were collected from the surface tows taken with larval nets in the waters around the Ryukyu Islands during November and December, 1973. Bryozoans (one species), tubeworms (Serpulidae,Janua (Dexiospira) foraminosa (Moore &Bush)) and goose barnacles (Lepas pectinata Spengler,L. anatifera Linné) were the most important species of sessile animals found on these tar globules. Sinking of tar globules byLepas of middle or large size was suggested from the differences in their specific gravity. 相似文献
11.
AbstractThe coefficient of lateral pressure at rest (k0) is a key parameter required to determine the horizontal stress distribution on subsoils. In this paper, k0 was measured from consolidation test for calcareous and siliceous sand, and the effect of water contents on k0 of calcareous sand was also discussed. Comparison of k0 values measured and calculated based on empirical formulas were carried on. Results indicated that (1) the reduction rate of k0 for calcareous sand with an increase in relative density was less than that of siliceous sand. The k0 at low water content was generally smaller than that of siliceous sand, and it was similar to that of siliceous sand for high-water content; (2) the k0 of calcareous sand increased with increasing initial water content initially, and then decreasing. However, the magnitude of variation was small. So the impact of water content on the k0 of calcareous sand was limited; (3) the formulas proposed by Abdelhamid and Krizek (1976) for estimation of the k0 was highly applicable to calcareous fine sand.
- Highlights
The k0 of calcareous fine sand was investigated.
The k0 of calcareous and siliceous sand were compared.
Shape parameters of sand particles for calcareous and siliceous sand were analyzed.
Effects of related density and initial water content on the k0 were discussed, respectively.
The applicability of various empirical formulas for calculating k0 of calcareous fine sand was evaluated.
12.
R. L. Allen 《新西兰海洋与淡水研究杂志》2013,47(4):687-692
The method of comparing growth curves that is described in this paper consists of describing growth for the ith group by the equation Li(t) = ai + bi rt, which is linear in the two parameters ai and bi, and subsequently using the analysis of variance to compare the curves. The advantages of the linear growth equation over the von Bertalanffy equation, l i (t) = ai + bi rt i, are easier analysis and a much simpler interpretation of the differences in the parameters in terms of growth rate and maximum size. An example comparing the growth of male and female red gurnard Chelidonichthys kumu from four areas within the Hauraki Gulf is given. 相似文献
13.
Mouctar Kamara Rashila Deshar Sahadev Sharma Md. Kamruzzaman Akio Hagihara 《Journal of Oceanography》2012,68(6):851-856
Weller??s allometric model assumes that the allometric relationships of mean area occupied by a tree $ \bar{s} $ , i.e., the reciprocal of population density $ \rho $ , $ \bar{s}\left( { = {1 \mathord{\left/ {\vphantom {1 {\rho = g_{\varphi } \cdot \bar{w}^{\varphi } }}} \right. \kern-0em} {\rho = g_{\varphi } \cdot \bar{w}^{\varphi } }}} \right) $ , mean tree height $ \bar{H}\left( { = g_{\theta } \cdot \bar{w}^{\theta } } \right) $ , and mean aboveground mass density $ \bar{d}\left( { = g_{\delta } \cdot \bar{w}^{\delta } } \right) $ to mean aboveground mass $ \bar{w} $ hold. Using the model, the self-thinning line $ \left( {\bar{w} = K \cdot \rho^{ - \alpha } } \right) $ of overcrowded Kandelia obovata stands in Okinawa, Japan, was studied over 8?years. Mean tree height increased with increasing $ \bar{w} $ . The values of the allometric constant $ \theta $ and the multiplying factor $ g_{\theta } $ are 0.3857 and 2.157?m?kg???, respectively. The allometric constant $ \delta $ and the multiplying factor $ g_{\delta } $ are ?0.01673 and 2.685?m?3?kg1???, respectively. The $ \delta $ value was not significantly different from zero, showing that $ \bar{d} $ remains constant regardless of any increase in $ \bar{w} $ . The average of $ \bar{d} $ , i.e., biomass density $ \left( {{{\bar{w} \cdot \rho } \mathord{\left/ {\vphantom {{\bar{w} \cdot \rho } {\bar{H}}}} \right. \kern-0em} {\bar{H}}}} \right) $ , was 2.641?±?0.022?kg?m?3, which was considerably higher than 1.3?C1.5?kg?m?3 of most terrestrial forests. The self-thinning exponent $ \alpha \left( { = {1 \mathord{\left/ {\vphantom {1 {\varphi = }}} \right. \kern-0em} {\varphi = }}{1 \mathord{\left/ {\vphantom {1 {\left\{ {1 - \left( {\theta + \delta } \right)} \right\}}}} \right. \kern-0em} {\left\{ {1 - \left( {\theta + \delta } \right)} \right\}}}} \right) $ and the multiplying factor $ K\left( { = \left( {g_{\theta } \cdot g_{\delta } } \right)^{\alpha } } \right) $ were estimated to be 1.585 and 16.18?kg?m?2??, respectively. The estimators $ \theta $ and $ \delta $ are dependent on each other. Therefore, the observed value of $ \theta + \delta $ cannot be used for the test of the hypothesis that the expectation of the estimator $ \theta + \delta $ equals 1/3, i.e., $ \alpha = {3 \mathord{\left/ {\vphantom {3 2}} \right. \kern-0em} 2} $ , or 1/4, i.e., $ \alpha = {4 \mathord{\left/ {\vphantom {4 3}} \right. \kern-0em} 3} $ . The $ \varphi $ value was 0.6310, which is the same as the reciprocal of the self-thinning exponent of 1.585, and was not significantly different from 2/3 (t?=?1.860, df?=?191, p?=?0.06429), i.e., $ \alpha = {3 \mathord{\left/ {\vphantom {3 2}} \right. \kern-0em} 2} $ . Thus the self-thinning exponent is not significantly different from 3/2 based on the simple geometric model. On the other hand, the self-thinning exponent was significantly different from 3/4 (t?=?6.213, df?=?191, p?=?3.182?×?10?9), i.e., $ \alpha = {4 \mathord{\left/ {\vphantom {4 3}} \right. \kern-0em} 3} $ . Therefore, the self-thinning exponent is significantly different from 4/3 based on the metabolic model. 相似文献
14.
A formulation for the aerodynamic roughness length of air flow over wind waves $$z_0 = \gamma {\text{ }}u_* /\sigma p$$ which was proposed by Toba (1979) and Toba and Koga (1986) from dimensional considerations with some data analysis, is shown to correspond with a formulation for irregular solid surfaces $$(z_0 /h) = a(h/l)^{1 + \beta } $$ which resulted from work by Woodinget al. (1973) and Kustas and Brutsaert (1986);u * is the friction velocity,σ p the spectral peak frequency of wind waves,h the mean height of the solid obstacles,l the mean distance between their crests, andα,Β, andγ are constants. This correspondence is reached by the existence of a statistical 3/2-power law and an effective dispersion relationship for wind waves. Because both approaches of parameterizingz 0 were arrived at independently, they provide each other mutual reinforcement. 相似文献
15.
The breakup of western margin of Australia from Greater India started around 155 Ma and progressed southwards. After the separation, the interceding intraplate region experienced large volumes of submarine volcanism, extending over 100 Myrs. The Christmas Island Seamount Province (CHRISP, as it has been dubbed) lies south of the Java-Sunda Trench, and contains numerous submerged volcanic seamounts, and two sub-aerially exposed island groups—Cocos (Keeling) Islands, and Christmas Island. While recent geochronological investigations have shed light on the diverse eruption ages of the volcanics of this region, some islands/seamounts have demonstrated protracted volcanic histories, and it is not clear how the volcanic loading, tectonic subsidence, and subsequent emergence history of the islands relates to these discrete volcanic episodes. This study utilises a number of geophysical techniques to determine the crustal structure, loading and subsidence history, and last sub-aerial exposure age for the CHRISP. The study shows that flexural and subsidence modelling are reliable techniques in constraining the age of the seamounts when geochronological techniques are not possible. Utilising regional gravity signatures, we model the crustal structure underneath the Cocos (Keeling) Island, and constrain the thickness of the limestone cover between 900 and 2,100 m. Using age-depth subsidence curves for oceanic lithosphere the time since these seamounts were exposed above sea-level was determined, and a trend in exposure ages that youngs towards the west is observed. Two episodes of volcanism have been recorded at Christmas Island and they are of different origin. The younger phase in the Pliocene is a manifestation of flexure induced cracks produced in the lithosphere as it rides the subduction fore-bulge, whereas a low velocity seismic zone rising from the lower mantle, and tectonic reorganization, may be associated with the older Eocene volcanic phase, as well as much of the rest of the province. Our modelling also supports the existence of an older, undated volcanic core to Christmas Island, based on the loading ages from flexural modelling. 相似文献
16.
This paper is focused on the hydrophysical state of the western basin of the Large Aral Sea in its present conditions of ecological crisis. The investigation is based on the direct observations carried out during a field survey in the Aral Sea in October of 2013. The analysis of the distribution of the temperature values across the western basin is given. Time series of the water currents variability were recorded at four mooring stations. The pattern of the basin-scale circulation response to the wind forcing was described. Based on the in situ measurements of the time variability of the temperature fields, the principal characteristics of the internal waves were specified for the present-day Aral Sea for the first time. 相似文献
17.
Severe red tides due toChattonella antiqua occur sporadically during summer in the Seto Inland Sea, Japan, and cause significant damage to the fishing industry. In order to assess the chemical environment with respect to the outbreak ofC. antiqua, environmental factors that affect the growth ofC. antiqua were monitored around the Ie-shima Islands, the Seto Inland Sea, in the summer of 1986. In addition, a growth bioassay of the seawater usingC. antiqua was conducted under a semicontinuous culture system. Although temperature, salinity and light intensity were optimum for the growth ofC. antiqua, red tides by this species did not occur. Concentrations of NH 4 + , NO 3 ? and PO 4 3? were low (<0.4, <0.2, <0.06 µM, respectively) above the thermocline (8–12 m) and high below it (0.6–2, 4–8, 0.4–0.8 µM, respectively). Vitamin B12 concentrations did not change significantly between the surface (0 m) and below the thermocline (25 m) in the level of 2–4 ng·l?1. The growth bioassay revealed that in the surface waters, concentrations of N- as well as P- nutrients were too low to support a rapid growth ofC. antiqua. At the depth of 25 m, neither N, P nor B12 limited the growth rate. In order to obtain more quantitative information on the growth rate as a function of the concentrations of N- and P- nutrients,C. antiqua was grown in a semicontinuous culture system by changing nutrient concentrations systematically. The observed growth rate (μ) can be approximated as follows: $$\mu = \mu _{\max } .\frac{{S_N }}{{K_g ^N + S_N }}.\frac{{S_{PO4} }}{{K_g ^P + S_{PO4} }},$$ whereS N is the concentration of NO 3 ? plus NH 4 + (0–6 µM),S PO, the concentration of PO 4 3? (0–0.6 µM), μmax (0.97 d?1) the maximal growth rate,K 0 N (1.0 µM) andK 0 P (0.11 µM) the half saturation constants for NO 3 ? and PO 4 3? , respectively. Using the above equation with nutrient concentrations measured, the rate at which seawater supports the growth ofC. antiqua can be estimated and this can be used for the assessment of chemical environments with respect to the outbreak ofC. antiqua. 相似文献
18.
Errol Wiles Andrew Green Mike Watkeys Wilfried Jokat Ralf Krocker 《Geo-Marine Letters》2014,34(6):525-540
Although global thermohaline circulation pathways are fairly well known, the same cannot be said for local circulation pathways. Within the southwest Indian Ocean specifically there is little consensus regarding the finer point of thermohaline circulation. We present recently collected multibeam bathymetry and PARASOUND data from the northern Natal Valley and Mozambique Ridge, southwest Indian Ocean. These data show the Ariel Graben, a prominent feature in this region, creates a deep saddle across the Mozambique Ridge at ca. 28°S connecting the northern Natal Valley with the Mozambique Basin. Results show a west to east change in bathymetric and echo character across the northern flank of the Ariel Graben. Whereby eroded plastered sediment drifts in the west give way to aggrading plastered sediment drift in the midgraben, terminating in a field of seafloor undulations in the east. In contrast, the southern flank of the Ariel Graben exhibits an overall rugged character with sediments ponding in bathymetric depressions in between rugged sub/outcrop. It is postulated that this change in sea-floor character is the manifestation of deep water flow through the Ariel Graben. Current flow stripping, due to increased curvature of the graben axis, results in preferential deposition of suspended load in an area of limited accommodation space consequently developing an over-steepened plastered drift. These deposited sediments overcome the necessary shear stresses, resulting in soft sediment deformation in the form of down-slope growth faulting (creep) and generation of undulating sea-floor morphology. Contrary to previous views, our works suggests that water flows from west to east across the Mozambique Ridge via the Ariel Graben. 相似文献
19.
Masanori Ito Yutaka W. Watanabe Masahito Shigemitsu Shinichi S. Tanaka Jun Nishioka 《Journal of Oceanography》2014,70(5):415-424
To estimate benthic denitrification in a marginal sea, we assessed the usefulness of \({\text{N}}_{2}^{*}\) , a new tracer to measure the excess nitrogen gas (N2) using dissolved N2 and argon (Ar) with N* in the intermediate layer (26.6–27.4σ θ ) of the Okhotsk Sea. The examined parameters capable of affecting \({\text{N}}_{2}^{*}\) are denitrification, air injection and rapid cooling. We investigated the relative proportions of these effects on \({\text{N}}_{2}^{*}\) using multiple linear regression analysis. The best model included two examined parameters of denitrification and air injection based on the Akaike information criterion as a measure of the model fit to data. More than 80 % of \({\text{N}}_{2}^{*}\) was derived from the denitrification, followed by air injection. Denitrification over the Okhotsk Sea shelf region was estimated to be 5.6 ± 2.4 μmol kg?1. The distribution of \({\text{N}}_{2}^{*}\) was correlated with potential temperature (θ) between 26.6 and 27.4σ θ (r = ?0.55). Therefore, we concluded that \({\text{N}}_{2}^{*}\) and N* can act complementarily as a quasi-conservative tracer of benthic denitrification in the Okhotsk Sea. Our findings suggest that \({\text{N}}_{2}^{*}\) in combination with N* is a useful chemical tracer to estimate benthic denitrification in a marginal sea. 相似文献
20.
The mixed layer depth (MLD) front and subduction under seasonal variability are investigated using an idealized ocean general circulation model (OGCM) with simple seasonal forcings. A sharp MLD front develops and subduction occurs at the front from late winter to early spring. The position of the MLD front agrees with the curve where \({\rm D}T_{\rm s}/{\rm D}t = \partial T_{\rm s} /\partial t + {\user2{u}}_{\rm g} \cdot \nabla T_{\rm s} = 0\) is satisfied (t is time, \({\user2{u}}_{\rm g}\) is the upper-ocean geostrophic velocity, \(T_{\rm s}\) is the sea surface temperature (SST), and \(\nabla\) is the horizontal gradient operator), indicating that thick mixed-layer water is subducted there parallel to the SST contour. This is a generalization of the past result that the MLD front coincides with the curve \({\user2{u}}_{\rm g} \cdot \nabla T_{\rm s} = 0\) when the forcing is steady. Irreversible subduction at the MLD front is limited to about 1 month, where the beginning of the irreversible subduction period agrees with the first coincidence of the MLD front and \({\rm D}T_{\rm s}/{\rm D}t =0\) in late winter, and the end of the period roughly corresponds to the disappearance of the MLD front in early spring. Subduction volume at the MLD front during this period is similar to that during 1 year in the steady-forcing model. Since the cooling of the deep mixed-layer water occurs only in winter and SST can not fully catch up with the seasonally varying reference temperature of restoring, the cooling rate of SST is reduced and the zonal gradient of the SST in the northwestern subtropical gyre is a little altered in the seasonal-forcing case. These effects result in slightly lower densities of subducted water and the eastward shift of the MLD front. 相似文献