DATA  COVERAGE  AND METHODS

Traditional studies of morphological evolution in estuaries have mainly been based upon comparisons of isobaths。 In re- cent years, 3S techniques (remote sensing, global positioning, and geographical information system) have increasingly been employed for morphological studies of estuaries。 Remote sensing is useful for studying changes in marsh areas (Shen et al。, 2006)。 Although it is also used to examine intertidal elevations (Koopmans and Wang, 1994), errors are usually large in intertidal zones with gentle slopes and unclear wa- ter–land boundaries and in areas where tidal corrections are difficult (Shen et al。,  2006)。

Data Specification

In this study, bathymetric maps were collected from the Maritime Safety Administration of China。 The date and scales are shown in Table 1。 The bathymetric data were sur- veyed by echosounder with a depth accuracy of ±1 cm。 Ele- vations in the intertidal wetland were surveyed by the Real- Time Kinematic–Global Positioning System (RTK–GPS) with a vertical accuracy of ±1。5 cm and a planar accuracy of ±1 cm。 Data on the annual water discharge and sediment load at the tidal limit of the Yangtze River were obtained from the Yangtze Water Conservancy Committee。

Data Processing

The ArcGIS software package was employed for processing the topographical data sets。 Bathymetric points and isobaths were digitized and imported into a Geo-database。 To mini- mize manual error, geostatistical analysis was used to modify the outliers resulting from the digitizing process。 Elevations surveyed by RTK–GPS were combined with the  bathymetric

data to generate a whole DEM for JDS in 2005。 Random sub- sets of the bathymetric points were used to estimate the er- rors in the prediction process。 The root-mean–squared (RMS) errors are shown in Table 1。 Isobaths were used to calculate the Fractal Dimensions (FD)。 Fractal theory has been applied in coastal research since its inception (Mandelbrot, 1967, 1983)。 Based on this theory, the box-counting method (Kraft, 1995) was used in this study to calculate the FD of the iso- baths。 Four 1 : 100,000 bathymetric maps were chosen to en- sure that the geometric changes were inspected at a constant map scale。

LONG-TERM  MORPHOLOGICAL  CHANGES

Changes in Planar Geometry

The 0-, 2-, and 5-m isobaths from 1958 to 2005 are shown in Figure 2, whereas Table 2 shows the geometry parameters of the isobaths。 Insight into the changes in planar geometry can be gained by inspecting the variation in these parame- ters。 For example, the long axis of the 5-m isobath increased from 36。9 to 51。1 km, with an expansion rate of 302 m y—1。 The maximum width was seen to have increased from 10。9 to 14。3 km at an expansion rate of 72 m y—1。 The obliquity increased from 21。9° (1958) to 29。6° (1989) but then decreased to 25。9° (2005), whereas the aspect ratio increased from 0。29 (1958) to 0。41 (1971) before decreasing to 0。28 in 2005。 This indicates that on the whole JDS expanded along the river channel with a clockwise rotation, but that the changes in shape and direction swung in most of the years。 Isobaths were more regular during stable periods (i。e。, the box-count- ing dimension was found to be smaller)。 Table 3 shows that the dimensions were smaller for isobaths at a greater depth。 The normalized mean FD declined dramatically from 1。6 in 1958 to —0。33 in 1971 and then increased to 1。47 in 1994 and

2。43 in 2005, suggesting that JDS was expanding but with perturbations。

Changes in Area and Volume

The sediment volume above the 5-m isobath increased from 614 × 106 m3 in 1958 to 1576 × 106 m3 in 2005。 The average annual rate of increase was, therefore, 20。5 × 106 m3  y—1  or 2% y—1。 During the same period, the area enclosed by the 5-m isobath increased from 210 km2 to 413 km2 at an annual rate of 4。31 km2 y—1 or 1。45% y—1。 Meanwhile, the volume above the 0-m isobath increased from 21 × 106 m3 to 129 × 106 m3 (3。9% y—1), and the associated area increased from 38。9 to