I just noticed the subject line. You are after the zero-sequence impedance, not the footing resistance. In this case, you need some average resistivity to use in Carson's formulas for re and xe. The equivalent depth of return depends on the resistivity and the soil is assumed to be homogenous which it obviously is not a very good assumption in your case. Fortunately, the line zero-sequence impedance is not all that sensitive to the resistivity.
For a typical 115 kV vertical construction Drake conductor line, I calculate the following zero-sequence impedances:
[tt]
rho R0 X0
(ohm-m) (ohm/km) (ohm/km)
0.01 0.250 1.189
0.1 0.257 1.441
1 0.272 1.690
10 0.294 1.935
100 0.323 2.178
[/tt]
Fault values are only partially dependent on the zero-sequence impedance, so this makes the overall result even less sensitive to resistivity. If the entire line is on the causeway, you should probably assume something low if you are concerned mostly with maximum fault levels. The current will not flow entirely in the salt water, so based on the graphs Cuky provided, I'd probably pick a resistivity of 1 ohm-m.