| In this article two parts of the work are done for the establishment of a high resolution bio-physical coupled model for regional simulation and forecast of the Bohai Sea, the Yellow Sea, and East China Sea. One is the establishment of a high resolution physical ocean model based on ROMS. As a case test and application of this model, the dynamical mechanism of winter circulation in the Yellow Sea, especially of the winter Yellow Sea Warm Current (YSWC) is discussed. The other is the improvement on MODIS atmospheric correction algorithm based on knowledge about the character of aerosol components in coastal atmosphere of China, which will provide more accurate remote sensing data for data assimilation in the coupled model.First of all, an ocean model of China coastal region with high spatial resolution is setup with usage of the three dimensional, baroclinic, terrain-following, primitive equations ocean model ROMS. The domain of model covers whole area of the Bohai Sea, the Yellow Sea, and East China Sea, also adjacent area of the Northwest Pacific. The model includes momentum and mass transfer with external area from surface and latent boundary and estuary of the Changjiang River. A simulation of climatological winter circulation is processed with this model, and the results of current, temperature and salinity distributions and mass transports through some important channels and sections show good agreements with historic observations and previous works, which suggests the effectiveness of this model.The mechanism of YSWC under control of winter monsoon is studied with surface wind stress and heat- and water- fluxes data during 2003~2004 winter. It's suggested by model results that the seasonal averaged YSWC is a northwestward current controlled by pressure gradient, both barotropic and baroclinic, and Coriolis force. The wind stress plays a role of adjusting the flow perpendicular to the extent direction of the Yellow Sea Trough (YST), but has little influence on the flow parallel to YST. The YSWC has two cores. One is on the slope zone west to YST, while the other one is close to central YST. The former one is mainly driven by baroclinic pressure gradient, while the latter is mainly driven by barotropic pressure gradient.The seasonal averaged current pattern only gives description about the macroscopic state of YSWC, which doesn't tell the dynamics and mechanism. Based on this concept, a further study on the response of YSWC to wind event in synoptic scale is processed. It's found that the YSWC shows totally different patterns and different dynamics during and after wind event. While, the YSWC is composed by both barotropic and baroclinic flows whetherever during wind event or not.During wind blasting, the barotropic current features a wind driven current in shallow water with latent boundary, and the structure of current has significant response to the variation of wind stress. During the first few days of wind event, accumulation of water mass occurs on the western coast attributing to geostrophic effect, which products a sea level gradient and drives a southward flow in the whole area. Along with the lasting of wind, a trend of moving eastward is introduced by the generation of south-north sea level gradient. The result of this antagonism between the two dynamics in zonal direction gives birth to westward mass transport in shallow water and eastward mass transport in deep water. A ridge-trough-ridge-trough system of sea level is generated by this transport, which controls the current pattern in meridional direction. After wind event the barotropic system falls into oscillation. The magnitude of south-north sea level gradient becomes much smaller, while the structure becomes more complicated. The zonal ridge-trough system still exists while weaker and westward-shifted. The barotropic on west flank of YST remains northward, while on east side the flow turns over to southward.The structure of baroclinic geostrophic current doesn't vary much. The current pattern is: northward current exists all over the region except at middle and bottom layers on deep west flank of YST where there is a southward counter flow. The location of the counter flow slightly moves eastward after wind event, which may be caused by the densification effect of cooling. The velocity of baroclinic flow is much smaller than barotropic one during wind event but of similar magnitude after that. The winter YSWC is composed by the above-mentioned barotropic andbaroclinic current, whose current pattern can be describe like this: during wind bursting, there is strong upwind current lying in the central area of YST, whose western extent range is larger than eastern one; during wind relaxation, a weaker northward current exists on the west flank of YST. The structure of YSWC varies between these two patterns under the alternation of wind bursting and relaxing. Whether there is wind or not, a northward current occurs on west flank of YST, although the strength of current varies with wind strength; one east part of YST there is a upwind current during strong wind and southward current during wind relaxation, embodying a characteristic of reciprocating flow.An empirical relationship between aerosol multiple forward scattering factor Ba and the ?ngstr?m exponentαand turbidity coefficientβis put forward based on atmospheric diffuse transmittance, direct transmittance and aerosol optical thickness (AOT) calculated from in situ measurements of Spectroradiometer during 2003~2007. An improvement of diffuse transmittance calculation in MODIS atmospheric correction algorithm is processed according to this empirical relationship.Analyses to measurements show that the MODIS standard algorithm leads to remarkable overestimation of diffuse transmittance, especially in condition that AOT is large. The mean relative error between calculations and measurements can reach as much as 32%, which will lead to evident error in the products of MODIS ocean color remote sensing in area where concentration of aerosol is large. The new algorithm gives good correction to the calculation of diffuse transmittance. Tests show that the relative errors and mean square root errors between new calculations and measurements are only about 1/5 and 1/4 of that of old algorithm, respectively. New algorithm has clear improvement in accuracy.
|
PDF Full Text Request