We're talking about in the mitered corners? If I would lay my head against the wall and look at the edge of the casing, I would see a V at the joint (with the wide point of the V being away from the wall)?

If so, that's kind of interesting. One expects wood to move with changes in humidity, with the greatest movement across it's width. Assuming the miter was accurate, that kind of V would seem to indicate that the surface of the wood contacted at a greater rate than the wood against the wall.

I would be more inclined to think that the change in is in the wall itself, causing the two pieces of casing to somehow move out of the same plane. Is this new construction?

From what I have found on other sites, it appears to be a moisture problem. The heal, short end of the miter, is seperated and the toe, long end of the miter, is still tight. I installed this trim in early spring when it was wet outside. Apparently, the trim had soaked up moisture and when it dried back out in summer the joints opened up. All of the ones that did this are on windows, which are obviously on the outside of the house where climate changes are more likely.

The wood absorbed moisture and expanded.

It is now drying out.

Even glue cannot stop the movement of wood as the moisture content changes.

Chapter 3 of the Wood Handbook (USDA) (linked below) covers wood movement with moisture content.
Figure 3-3 is very good for understanding how shape changes depend on how the wood was cut from the tree.

Here is a link that might be useful: Wood Handbook, Chapter 3

It's definitely a problem with moisture content; the wood is drying out and shrinking. But, it's probably not useful to blame it on the weather. The wood's MC was too high at installation, but that was probably due to improper drying or improper storage, not the fact that it was springtime and rainy. In fact, relative humidity levels tend to be more stable outdoors than in, in most parts of the country.

I thought I'd posted this link, but apparently, I got interrupted. The discussion and diagram explains why this happens -- the nature of solid wood.

Here is a link that might be useful: Why miters open

The FHB explanation is correct, but fails to not that the type of wood cut (flat sawn vs. radial sawn) affects how much movement occurs.

Wood does not shrink the same amount in the two direction (radial and tangential) and this can be used to advantage when making wider molding.

If trim intended for interior use is not stored in controlled conditions it will absorb moisture in storage and come to equilibrium with the humidity in the storage area.

If it is acclimated in its final location in a controlled environment long enough before use it will return to its previous moisture content.

Leaving interior trim sitting on a concrete floor is a great way to promote it absorbing moisture and swelling so things go badly.

The fast growth timber now used does not help.

Miters are inherently unstable "joints" in wood, the degree to which they open up depends on the width, species, cut, and moisture content.
If you browse through a selection of pre-made wood picture frames, you may spot some that have the sides bowed inward. Their joints are fastened in such a way to prevent opening up, so the inevitable wood movement causes the sides to bow instead.
If you fill your open miters with caulk or filler, then you are condemning them to get even worse when their moisture content reaches it's next equilibrium. Then you start over with an even worse-looking joint the next paint job.
Miters suck, but there is little alternative.
Sometimes we biscuit them (see: plate jointer) which allows a positive glue joint at the corners, strong enough to withstand some of the movement, and the glue itself seals the end-grain further limiting the intake of moisture. It helps a little, but it's a very time-consuming step.
Casey

My version of bicuits:

and gravy:

Even biscuits and titebond may not hold in poorly cut (from the log), poorly dried, or wide pieces.

If wood is prevented from moving by cross grain fastening it may just split to relieve the strain.

It will most likely split if it's struck by an asteroid too.

Nothing is guaranteed.

"It will most likely split if it's struck by an asteroid too."

It is not all all uncommon for cross grain construction to result in forces on wood so great it splits, no asteroids required.

This is why cross grain constriction is avoided in anything except veneer work and plywood construction.

If you want to split the side of a furniture carcase, just fasten the drawer runners or shelf supports across the width of the sides.

The splitting is only a matter of time.

Old wood that grew slowly had enough strength and stability that opening of miters even on relatively wide door and window casing was not as much of a problem.

One of the things seen in many antiques that are still in good shape is a relative lack of cross grain construction.
The cheap stuff that had cross grain construction was junked a long time ago.

Wood changes size very little with moisture along the height of the tree, what is usually the length of a board.
It changes significantly in the radial and tangential directions of the growth rings, what is commonly referred to as the 'grain' or figure of the wood.
The radial and tangential changes are not even the same, meaning that a board that is square at one moisture content may become diamond shaped at another moisture content.

Chapter 3 of the USDA 'Wood Handbook' linked below has the nitty-gritty details and factors for many species.
Figure 3-3 is especially telling.

Here is a link that might be useful: Wood Handbook, Chapter 3