Maybe take a look at some schematic pictures how proper grounding is constructed, as a solid basis for overvoltage protection, in buildings, per modern norms (in the E.U. and in the U.S.).
The essential point is:
on the inside, the building should have a central star ground, called "protective earth" or PE potential, somewhere at its base. If there's no native metal structure to serve as an inherent PE potential (i.e. the building is not a metal tower or some such), the building must be equipped with a wiring system, comprising such a centralized PE system, distributed throughout the building. Those are all the yellow-green conductors in your three-conductor mains wiring/cabling. There will be a large busbar somewhere in the main wiring cabinet (in the basement or at the ground floor) where all those yellow-green wires join, coming from the whole building. There will be similar busbars, just probably smaller, in any wiring cabinets - the central cabinet per floor, or various smaller tributary wiring cabinets. In terms of safety, the yellow-green busbar is always the most important element of a wiring box or cabinet.
the building will also have an outdoor ground, called the lightning collector and down-conductor system. Imagine this like a cage outside the building, with spikey collectors on the roof, and with a "grid" or even just a "ring" buried just under or in the concrete of the building's foundations. When building the concrete foundations of a modern building, the building plans typically mandate that a metal stripe be buried under the foundations of the perimeter walls - preferably a stainless steel stripe, or at least a zinc-galvanized iron stripe, like 30x4 mm cross-section. Plus there may be additional grounding electrodes. Again the point is to bolt the whole outside cage (lightning collector system) to comprise a common potential, that's electrically anchored to the underlying earth. Where I live in central Europe, 99% of the lightning strikes are within 200 kA of peak current. Now imagine if your lightning collector system has a resistance to the earth of 100 milliOhms (I believe the norms are not even as strict as this). Uncle Ohm says that 200 kA * 0.1 Ohm = 20 kV peak voltage. You don't want this to appear across the inner wiring of your building, so the ouside lightning conductor must stay principally separate from your indoor PE yellow-green star (with an important exception, mentioned in the next paragraph). Which has interesting implications if you need outdoor radio antennas or overhead communication wiring come into the building - but that's probably a whole different fairy tale.
Now comes the promised exception: the outdoor lightning conductor earth, and the indoor yellow-green tree, must be interconnected as well - at a precisely mandated spot, at the bottom of the building! I.e., the central inner equipotential busbar of the PE system within the building must be interconnected to the base of the outdoor lightning protection system. Thus, all the grounds across the building "are one", and if a lightning strikes, it gets earthed through an ouside path to the underlying earth - while the inner yellow-green PE system does not receive a direct spanking (inevitably there would be some residual EM-induced noise, from the lightning pulse just sliding through the outdoor down-conductors).
Anything outside gets grounded to the outdoor lightning down-conductor system. Railings on balconies, sheet metal "trimmings" on the roofs, antenna masts, metal things on chimneys, you name it.
Anything inside gets grounded to the yellow-green spider. Such as, any overvoltage arrestors, clipping the voltage on various "nodes" for various purposes. The yellow-green ground circumvents circuit breakers, including the special "balanced" ICCB's that detect "dangerous touch" (leak from the system of live+return).
Incoming mains cabling brings an earth conductor with it. It's a single "return" conductor, shared by the three "phases". It doesn't have a separate yellow-green PE. This "shared return" entering the building gets grounded to the central earth point at the base of the building, too. The separate yellow-green PE splits=stems at the base of the building, heading up the whole house.
Now back to earth electrodes. They are implemented at the base of the building deliberately. If you have some natural/inherent large metal object, buried firmly and deep in the ground, there's hardly any point in not using it. As other have pointed out, make sure that your earth tap does not electro-corrode or some such. There are ways to prevent electro-corrosion.
And, take care to have the rest of your system in good shape inside and outside the building = well interconnected, the way it should be.
If a lightning does strike, or you have a short to GND from your mains, that's bad luck, but you probably wouldn't avoid harm to your neighbors at large by NOT having our PE system grounded locally to the best available earth electrode :-)
EDIT: Regarding the risk of "hitting your neighbors through the ground water": apparently this is a case of a lesser evil :-) And a matter of making the ground as solid as possible, across an area, so that a fault current doesn't "ruffle its feathers" very much.
A lightning strike, or a mains electricity source, all have a limited energy / current capability. The point of grounding is to conduct any fault current (large but still limited) either back to the source (hence the dedicated return conductor) or into the solid geological earth as directly as possible, so that an elevated voltage doesn't spread through the human-made GND/neutral wiring across the neighborhood, causing damage.
I've recently read an analysis written by someone with decades of practice as a project engineer in high-volt electricity, a transmission line designer who has routed maybe thousands of kilometers of thick wires on pylons at 400+ kV and kiloAmperes of nominal working capacity. That's dozens of kA worth of short circuit capability. For the case at hand, he was debating how far from a "spot of accidental short to earth" there's a dangerous level of "stride voltage", when a wire falls off the pylon, and there's no local earth electrode to the center conductor (which is a dangerous design omission). If memory serves, he mentioned that the dangerous zone had a radius of a couple hundred meters - apparently considering relatively dry soil on the surface.
It has to do with the gradient of voltage (and current) through the mass of the earth, the short circuit current available (or the peak current from a lightning strike) - and a key factor is the soil resistivity. The lower the soil resistivity, the lesser the impact (local elevation of voltage potential) caused by a particular "fault current" level. And, your underground water table improves the "resistance" significantly - unless your geology varies wildly, the underground water soaks a couple dozen miles deep. Electrically that's quite a solid "earth node".
Any fault current hits your in-house PE conductors (lightning collector system) first, reaches the local earth electrode, and spreads into the geology. If your earth electrode is good, your local "base of the building equipotential busbar" doesn't jump very much (voltage-wise), so the voltage pulse spreading to your neighbors via the mains return cabling isn't very high, and is less likely to harm anyone, in terms of their local voltage difference between the mains GND and their respective local geological earth potential. Even if your water table happens to be a layer (maybe several meters) on a lake of asphalt or some such, a good connection to the water table will ensure that your neigborhood is "all at the same level" in terms of the geological earth potential, so again not very different from the mains return GND... counter-arguments are welcome :-)
Should you attempt to protect your neighbors by not grounding to the earth, this is what happens: any fault current ("charge from above") will find a path of least resistance to a suitable "near earth potential". A short in the mains is likely gonna end up in the neutral return, flowing back to your neighborhood, resulting in a voltage step at your neighbors' premises. A lightning, worth millions of Volts, may appreciate any natural conductor: any pipes that happen to be around, live mains, metallic telco or CaTV lines, a rainwater drain pipe coming down the wall, a nice soot-lined (carbonated) chimney... and can skip distances to your geological earth, so not grounding your mains return node to a local earth won't prevent lightning from hitting that earth anyway :-)
The local in-house connection of mains neutral to local earth works the other way too: if a "GND wobble" comes from your neighbors via the mains return wire, having a solid local earth electrode will "bolt it to ground". To your local earth - leaving no differential voltage between mains return, PE and the local geological electric earth potential.
Maybe I'd like to mention one other fault scenario: consider a three-phase system. You get an accidental short in your house, one phase to the shared neutral/return. This will move the potential of your neutral/return towards the phase where the short has occured. At the same time, your neutral/return potential will move away from the other two live phases - thus causing overvoltage, as percieved by powered devices (loads) in those two other phases. This gets worse if you have loose joints in your 3-phase ground return wiring, towards your upstream transformer. Or maybe your uplink to the transformer is just relatively long and relatively small cross-section, which makes the shared return "GND potential" more negotiable, dependent on load distribution among the phases (individual things powered in your house). A solid local earth electrode will help you with that problem, too. So that's a partial reason why licenced "proof testing techies" here measure not only the resistence of the neutral ground return towards the trafo, but also the earth electrodes.
Solar means DC, possibly above 100V depending on the particulars of your setup. DC is a whole different can of worms. I'm not an expert, but I'd hazard a guess that a solid grounding system is key to safety.
I've actually seen electric-powered systems operated without a PE, and with a floating inner "neutral/return node". I've seen plastic cabinets wired like this, used in the "railway trackside" segment of industrial process control. The people building those systems called them "IT", though I'm not sure this maps well to the taxonomy of TN/TT/IT that can be found in the interwebs. So it was essentially small-signal railway safety control gear, along with communications/signaling. An interesting partial problem happens to be, how to isolate broadband digital communication lines, entering the cabinet, to the desired isolation level (double digit kiloVolts) - if you want to keep using existing metallic wiring, rather than upgrade to fiber optic. These cabinets contain small, self-contained "islands of no connection to earth". The point is to make the gear resilient against incindents involving loose earth (neutral return) in the railway traction power. An electric railway engine guzzles Megawatts of power at 3 to 25 kV of voltage, i.e. currents up to the kA range... I know - not your problem. Apologies for going off topic.
