DS, I agree that the tank should be bonded, but it is a pressure vessel and you just can't cadweld to the vessel itself. This is most likely an ASME certified vessel and any welding to the vessel must be made by an ASME certified welder using specific procedures. I highly doubt that the cadwelding process is certified for a pressure vessel. A bolt or cadweld to the support would be fine.
In our 1.9 million gallon pool we bond everything with a crimp on lug and thread the metal if possible and use 316 stainless bolts, 1/4" or 3/8". This allows us to remove the bond to make repairs to the equipment or the bonding strap when needed. This makes maintenance a snap... well, as snappy as it can be for over 1000 documented bonding points
Others' comments are on the money- this is not an NEC problem, but a violation of the ASME Boiler & Pressure vessel code (and you thought the NEC was expensive!) From an engineering viewpoint, the cadweld process is hot enough to affect the temper/strength of the metal. The depth of the weld also introduces a stress riser/weak spot in the tank wall. It is also located in possibly the most vulnerable spot of the tank. Last but not least, if the tank was filled with anything -even water- at the time, the cadweld isn't even electrically sound. Another fine Darwin candidate.
(Webmaster note: This may duplicate an earlier post) This is a good example of why you should know something about other trades. This is not an NEC violation; rather, it violates the ASME "Boiler & Pressure Vessel Code." This job violates training, technique, materials, and certification requirements. From an engineering viewpoint, the cadweld is hot enough to change the temper of the metal, and introduce a stress riser. It does so in one of the most vulnerable areas- on the flat, near a seam. Even if the tank was filled with water at the time, there are penetration issues. The ground should have been attached to the tank's foot. The installer is a Darwin candidate.