Just looking at typical midday values:

Rsum = 500 W/m2

H = 100 W/m2

LE = 200 W/m2

G = 50 W/m2

Thus, there is an imbalance of 250 W/m2 (which is bigger than any of the measured terms).  The above values are before WPL corrections, etc., but I would be surprised if the corrections could double these terms.  The logical conclusion is that energy advection by the surface water is huge.   I'll try to see if this explanation is validated by the tidal cycle.

P.S. It could be useful to have a dat() function of the tide level...

P.P.S. I'll repeat the suggestion to mount a Wetness sensor close to the surface to detect water.