Results

Site hydrology

Both seasonal and monthly lunar influences were important for the hydrological conditions at the site (Fig. 2). The highest monthly mean stages at the Shark River were in September and October (-0.23 and -0.19 m respectively), typical for this drainage. The high river stage was a result of the maximum discharge of accumulated water from the wet season (June to September, Fig. 2). Groundwater piezometric head pressure was also high during September and October (0.06 and 0.12 m respectively) due to hydrological recharge from the wet season. Daily river stage was a reflection of monthly lunar tidal flooding, wet season river discharge and annual sea level variability (thermal expansion, Provost 1973).

There was moderate correlation between the two hydrological metrics used in the multiple regression with an r = 0.72 for Shark River stage to groundwater piezometric head pressure. Tolerance values were above 0.547 and Variance Inflation Factors were less than 1.829, suggesting that despite some correlation between predictor variables, collinearity was not a serious issue for these data (Quinn and Keough 2002; Neter et al. 1996).

Accretion

The feldspar marker horizons did not become completely covered until 172 days after installation (September 10, 2002). The marker horizons were covered with mineral, organic and root matter. The annual accretion rate was 6.64 ± 0.56 mm yr^-1 (± 1 SE). Sediment deposition values were intermittent in nature with high rates in October 2002 and March 2003 (Fig. 3a). Slight erosion was evident during the November to December 2002 period (-1.8 mm) and the December 2002 to January 2003 (-0.8 mm) sampling.

Figure 3. Mean absolute soil surface elevation (±1SD) for (A) Accretion, (B) Shallow-RSET, (C) Original-SET, and (D) Deep-RSET. [click on images above for larger version]

Soil elevation

Changes in absolute soil surface elevation for both the Deep-rod and Original- SETs followed a similar pattern (Fig. 3c and d). Both devices recorded the highest mean soil elevations at the end of the wet season (8.89 mm on October 10, 2002 for the Original-SET and 15.14 mm on November 9, 2002 for the Deep-RSET) and the lowest mean elevations during the dry season (January 09, 2003; -2.24 mm and -0.06 mm respectively). The Shallow-RSETs had a distinctly different pattern of soil surface elevation, with the highest elevation at the end of the wet season (6.83 mm on November 9, 2002) and the lowest early in the wet season (-0.66 mm on June 03, 2002, Fig. 3b).

Relationships between soil elevation and hydrology

The daily rate of soil elevation change of the Shallow-RSET was partially explained (Adjusted R² = 0.16) by a negative relationship with the daily rate of change of the river stage at the site (Table 2). That is, as river stage increased, the soil elevation that was influenced by the shallow soil zone decreased (Fig. 4a). The rate of soil elevation change of the Original SET was positively related with the daily rate of change of the groundwater head pressure (Adjusted R² = 0.61; Fig. 4b; Table 2). This model was run with a reduced data set (n=28) due to a one time sampling error of Original SET 2. The daily rate of change of soil elevation for the Deep-RSET had a strong positive relationship to the daily rate of change of the groundwater head pressure (Adjusted R² = 0.90; Fig. 4c; Table 2). When groundwater head pressure increased the soil elevation increased for both the Original SET and the Deep-RSET.

Contribution of each zone to expansion/contraction of the entire profile

I calculated the variation in thickness of each of the four constituent soil zones (Eq. 2) and for the entire soil profile. I determined how much each of these soil zones contributed to absolute change of the entire profile by using a stepwise multiple regression model in which absolute change in the thickness of the entire profile was the dependent variable and the absolute change in thickness for each soil zone were independent variables.

The contribution of each soil zone was not equivalent to the relative proportion of soil profile it comprised (Fig. 5). The bottom zone (4-6 m) accounted for 63% of the variation in the absolute change in thickness of the complete profile whereas the middle zone (0.35 - 4 m) accounted for only 22% (Table 3, Fig. 5). However, the bottom zone comprises only 31% of the entire profile whereas the middle zone comprises 63%. Accretion and the shallow zone were not significant contributors to the overall absolute change in thickness of the entire profile (Table 3).