The alarm doesn’t go off at 4am because I don’t set one. I’m already awake, already calculating whether the cloud cover from the night before has broken, already thinking about whether the light I drove toward is still worth chasing. Last October I was parked at a trailhead outside Sisters, Oregon, headlamp on, boots laced, staring at a sky that had gone completely wrong. The forecast had called for a clear sunrise window. What I got was a flat grey ceiling sitting right on the ridgeline. I sat in the car for twenty minutes, drank my coffee, and decided to hike anyway.

That decision, and the patience behind it, is most of what mountain photography actually is.

But patience only works if your technical foundation is solid before the light arrives. Because when it does arrive, you have minutes, sometimes seconds, and the mountain does not care about your schedule.

Why Your Meter Gets Confused Above Treeline

Mountain environments break evaluative metering in ways that flat landscapes don’t. The problem is tonal range. When you point a lens at a snow-covered peak backlit by early morning sun, your camera sees a scene that might span 12 to 14 stops of dynamic range. Your sensor, on a modern mirrorless body, can capture around 14 stops under ideal conditions. But your meter doesn’t know which stops matter. It averages the scene toward middle grey, which means the bright snowfield gets pulled down and the shadowed foreground rock face gets pushed up. The result is an image that looks exposed for neither.

Add thin atmosphere at elevation and the problem compounds. Air at 10,000 feet scatters less light, which means contrast is harsher and color shifts faster. What looks warm and soft at sea level looks sharp and almost clinical up high. Magenta and cyan casts sneak into shadows. Your white balance preset, locked to “Daylight” or “Cloudy,” isn’t going to hold.

The Exposure Approach I Use at Every Alpenglow Window

I don’t use auto mode. Not for exposure, not for white balance, not for focus. This isn’t stubbornness, it’s control. In fast-changing mountain light, automation makes decisions you can’t reverse in the moment.

For sunrise and sunset alpenglow, I start with spot metering off the brightest part of the scene that I want to retain detail in, typically the illuminated face of the peak or the lit edge of a cloud. I then overexpose from that spot reading by 2/3 to 1 full stop to push the exposure right without blowing the highlight. On my Sony A7R V, I’m watching the histogram live and keeping the rightmost data within one column of the right edge. Highlight blinkies are on. If they fire on the snow, I back off.

My base starting point for alpine dawn in low wind: ISO 400, f/8, shutter speed adjusted until the histogram sits where I want it. F/8 gives me enough depth of field that foreground rocks at 6 feet and a peak at 2 miles are both sharp, assuming I’ve focused a third of the way into the frame. I shift to f/11 only if I need more depth or if I’m using a longer focal length, typically my 100-400mm GM, where focus plane gets thinner fast.

For white balance, I shoot a manual Kelvin value. At pre-dawn blue hour I’m around 4500K. As alpenglow hits I nudge to 5200K to 5600K and shoot a bracket at each. I fix the rest in Lightroom, but getting the Kelvin close in-camera means my RAW files need less correction and the color shift between frames in a bracket is more predictable to blend.

Blending Brackets Without Losing the Feeling of the Scene

Mountain scenes almost always require exposure blending. I shoot three-frame brackets at 2-stop intervals for most high-contrast situations, five frames if I’m shooting into direct sun with deep shadow foreground in the same composition.

In Lightroom, I merge to HDR with deghost set to “High” if there’s any wind moving foliage or cloud edges. Then I bring the merged DNG into Photoshop and mask the exposures by hand using luminosity selections rather than relying on the auto-blend. The HDR merge from Lightroom gives me a 16-bit DNG that’s roughly 80 to 120 MB depending on the body. That file holds up through heavy local adjustments without banding, which matters when I’m printing at 24x36 or larger.

The goal of blending isn’t to make the scene look technically correct. It’s to make it look the way my eyes adapted to it standing there. That’s a meaningful distinction.

The Fog That Became My Best-Selling Print

I once drove six hours to the North Cascades chasing a forecast that promised a three-day clear window on a peak I’d been circling for two years. I got fog. Thick, unmoving, total fog. I sat in it for two days, hiked every morning anyway, shot almost nothing usable.

On the second afternoon, the fog thinned for about eleven minutes. A single shaft of light dropped through the cloud layer onto a ridgeline of larches in full autumn gold. I had my camera on the tripod already because I’d been waiting, watching the edges of the clouds all day. I made six frames. One of them became the best-selling print I’ve ever made.

The image works because the fog is still there. The mood of those two days of waiting is inside it. You can’t manufacture that by showing up when the weather is easy.

Reading the Mountain Before You Set Up

Weather apps are starting points, not answers. I cross-reference three sources: the National Weather Service point forecast for the specific elevation I’m targeting, Windy.com for the actual cloud layer altitude, and the Mountain Forecast app which breaks conditions into 3-hour windows at multiple elevations. If those three sources disagree significantly, I assume the worst and plan around it.

I also scout on bad weather days deliberately. I want to know where the drainage channels are, where the wind hits hardest and might shake my tripod, where the first light lands and at what angle before I’m standing there at 5am trying to solve geometry problems I should have already answered.

The single most important thing I can tell you about mountain photography is this: the shot you don’t get because the light didn’t cooperate still teaches you something the easy shot never will. Go anyway.