I suspected that I was asking a difficult question, but I was also hopeful that someone actually had a canned solution. There have been a number of thoughtful suggestions. but, so far, I don't believe that we have arrived at a reliable workable solution. I would like to give a summary and commentary on what has been said in this thread. I will place excerpts from responders in bold type, followed by my comments. I invite you to correct, extend or clarify these points.
The “simple ideas” of direct white balance expressed by Paul C and others
“Not to sound overly simple, have you tried shooting a grey or white card at the same time?
Shoot it raw, as then you change the WB per the card. You can leave the camera on AWB, then use your raw converters WB tool to click on the card, then set this to the WB of the series.
Best way I know of.
Paul C.”
are too simple. My experience is that a “click” white balance off the grey card will render the grey card as perfect grey, but it will produce the wrong white balance for the scene that is (to be) dominated by the distant sky.
Don Lashier suggested not to white balance off a reference, like a grey card, but to work with color temperature and tint during raw conversion.
This is what I will call the guess and try method. Success depends critically on the tools proviided in the raw converter, as well as having a really, really well-calibrated monitor and color-managed workflow. Though I have good calibration, I find this approach really difficult to pull off because the color balance and tint controls are too crude.
Asher Kelman gave a mini tutorial on balancing lighting and color temperature. He also recognized implicitly the difficulties of this approach in the field.
I didn't think the table of color temperatures was very useful, except as a very rough guide, because the light is changing very quickly and I think success will require real time measurements. I took away from Asher's discussion that I might try to do two white balances: 1. Maybe something like a diffused reading from the sky but in the direction of sunrise. and 2: Perhaps a "local" white balance to capture the near field ambience. Then to combine two raw conversions in photoshop, perhaps with masks or something like that.
We also had suggestions to measure the color temperature and just to eyeball it.
I don't have a color meter to try this and I am not sure i would know what to do with the data.
So, overall I would rate the thoughts about dealing with mixed color temperature, expressed by Asher, as the most promising, but I repeat that we don't yet have a resipe to solve this technical point.