This is a long comment,as to answer your question it needs a course in quantum mechanics.
The basic reason atoms form is because the 1/r Coulomb potential that gives macroscopically the attraction between the positive and negative charges, at the quantum level it gives rise to the atomic orbitals. If there were no quantum mechanics the electrons would fall on the nucleus and neutralize it, and no atoms would exist. It is one of the reasons quantum mechanics had to be postulated.
Orbitals are probability loci in (x,y,z,t) where an electron can be found if measusred.
Here are the possible orbitals for the hydrogen atom.
So the negatively charged electron will be in a specific location (x,y,z) at a specific time .This allows the positive charge of the nucleus range to affect other charges, and the orbitals of different molecules can fit , like lego blocks, positive attracting negative. i.e. a new potential can be found, complicated because of the shielding of the electron orbitals, that can attract molecule to molecule, due to the topology of the orbitals that allows positive electric fields windows of long range.
This new (not 1/r) potential would be too complicated to be solved with the Schrodinger equation, and new mathematical models are used in order to fit and predict the bonding of atoms and molecules, as summarized in the link by Roger
Best Answer
No, there has been no change to the bond angle of pure water,1 nor does anyone expect to measure one. The reason is this angle is set by fundamental physical constants - things like the mass of the proton and the speed of light - albeit in a complicated way. Such things are called constants for a reason. Now there are some theories out there, a number of them a bit far-fetched, that speculate that maybe some of our constants have been evolving extremely slowly over billions of years, but certainly there has been no detected variation over the course of human scientific inquiry. Such a discovery would be huge, causing much agitation in all of physics and chemistry.
It should also be noted that the bond angle can both be calculated from theory and measured from experiment, and neither method yields exactly $104.5^\circ$. Now your theoretical calculations might improve if you make fewer simplifying assumptions than your predecessors, and your measurements might improve if you develop better apparatus and technique, but the true underlying value of the bond angle is not changing.
1 Defined here as oxygen-16 bonded to two atoms of hydrogen-1. Now the angle could vary with isotopic abundances, so perhaps someone was referring to changing isotopic abundances of a source of water. You might see this, for instance, in extreme climate change, since as temperatures rise the fraction of "heavy" water ($\mathrm{D}_2\mathrm{O}$, etc.) evaporated from oceans increases relative to "normal" water.