How do you predict a good sunrise & sunset?

There are many variables that cause a good sunrise & sunset.  Oftentimes it seems like magic and that there’s no way of predicting it.  But we can still give it the old college try, right?

Some of the atmospheric factors to take into consideration for predicting sunrise & sunset:

  • Cloud height & types
  • Humidity
  • Wind direction & speed
  • Atmospheric particulates; pollution, smoke, ash, etc
  • Topography

Can you go to the NOAA website, look at the forecast, and predict how colorful the sky will be?  I’d actually be willing to bet that with a little thought & practice you could get close sometimes.

But isn’t there an easier way?  Wouldn’t it be nice to just check a map and know if it’s worth going out to shoot or not?  Especially on those super early, cold mornings when you wonder if it’s worth getting out of bed.

Reviewing Skyfire and SunsetWx prediction tools

There are two sunrise & sunset prediction services available to photographers.

One is Skyfire, an add-on subscription service for The Photographer’s Ephemeris for iOS, and the other is SunsetWx, a free website.

You’ll see below that results are different when you compare sunset predictions between the two services.  The “ratings” system is very different between the two as well.

The main limitation of both of these: they’re only as accurate (at best) as weather forecasting usually is.


Get The Photographer’s Ephemeris at the App Store

Intro to Skyfire

Skyfire was hatched when award-winning photographer Matthew Kuhns wanted a way to know where to go for good light.  Photographers will spend the majority of their time chasing light.  Matthew wanted to know where to go without wasting any time.


Skyfire’s color key from

Skyfire uses satellite data from NOAA and the NWS to determine cloud cover.  Though those two services may differ, Skyfire will attempt to “average out” what the cloud cover might be.  Then it takes in many other variables, such as those listed at the top of this page, and runs them through algorithms.  These algorithms will eventually tell you how photogenic the clouds will be and how photogenic the lighting will be.

The app is available exclusively as an add-in service in The Photographer’s Ephemeris for iOS, an $8.99 tool every photographer should have on their phone.  If you’re not familiar with it, TPE will tell you exactly where in the sky the sun & moon are throughout the day, where shadows will be, and more.

Features of Skyfire:

  • Coverage of the continental United States, southern Canada, and Europe.
  • Goal to maintain 80% accuracy; currently claims 87% accuracy rate.
  • High-resolution model for very localized forecasts.
  • “Basic” subscription level forecasts two days out and includes Golden Hour light forecasts.
  • “Plus” subscription level forecasts four days out and sends push notifications when forecasts reach a set threshold for your favorite locations set in TPE.
  • Forecast updated continuously throughout the day as NOAA & NWS data changes.
  • Only available through TPE for iOS; coming to Android soon.

You can try Skyfire for free in a 30-day trial period.  The Basic annual plan is $29.99 ($2.49/mo) and Plus is $44.99 ($3.75/mo).  You can also get a three month plan; Basic is $9.99 ($3.33/mo) and Premium for $14.99 ($5/mo).

Intro to SunsetWx


SunsetWx’s global sunset prediction, given at a particular time of day when the sun might not actually be setting. works on much of the same principles but is web-based, and available worldwide.  It too was created by a photographer/meteorologist who wanted to create a “sunset model”.  The major variables taken into account for SunsetWx’s algorithm are cloud cover, moisture, and atmospheric pressure.

It doesn’t have the interactive controls that Skyfire has, thus you can’t “zoom in” like you can in Skyfire.

Also note that the predictions given are independent of when the sun is actually setting – it will tell you how colorful the sunset would be if it were setting at that particular forecast hour you were looking at.  So you must use the controls to go to the forecast hour corresponding to your local sunset.

SunsetWx will give a “vivid” rating to areas it deems have the right mix of clouds, humidity, and pressure, and show these in warm oranges and reds. Dry clear air, or low clouds with precipitation will get a “poor” rating with cool blues on the map.  “Average” in-between areas will be depicted by greens and yellows.

SunsetWx publishes updates a few times per day, at a lower rate than Skyfire.

SunsetWx’s maps do have playback controls, and their time stamps are in Zulu/GMT time.  So it’s important to note that, while watching an animation, the times are not in local time.  The webpage makes an attempt to determine your sunset time based on your location, but the time zone is in Eastern time.  The forecast hour, however, is correct.  See the screenshot below.


SunsetWx telling me when my sunset is in Eastern time, even though I’m in California, and which forecast hour to use (that’s accurate).

How the tests were conducted

I tried to control the tests as best I could.  This was my process:

  • Check the forecast a consistent number of hours prior to sunset.  Eight hours and four hours prior to sunset would be realistic times for my planning purposes.  Checking days in advance doesn’t seem to have much utility for me.
  • Get a good random spread of testing days regardless of weather
  • Shoot a panorama from the same location, panning from sunset azimuth in the west then through north.  I’ve drawn the coverage of the photo on the Skyfire screenshots.
  • Shoot at exactly sunset.

Some days I was unable to get to my normal location, and/or not planning on testing that night but it happened anyways.  I’m including the results from those nights just so that you can get more data points.

I should also add that even though the data presented here only has a sample of ten days, I tried to check it every night even if I could not take any pictures.  I will discuss those nights at the end.

Test Results

Interpreting the results below is fairly straightforward.  You’ll see Skyfire’s prediction next to SunsetWx’s prediction, captured at the same time of day.  I’ve included updated predictions from Skyfire where I remembered to get them.

The area you’re looking at in the photos is drawn on the Skyfire screenshot.

April 18, 2017

Skyfire on the left, SunsetWx on the right, taken 3 hours before sunset.  Skyfire shows 50%-60% chance of photogenic sunset.  SunsetWx rates sunset as “poor”.

skyfire and sunsetwx

tahoe sunset

April 19, 2017

Skyfire on the left, SunsetWx on the right, taken 8 hours before sunset.  Skyfire shows 50%-60% chance of colorful sunset.  SunsetWx rates sunset as “average”.

skyfire and sunsetwx

Updated Skyfire captures 4 hours until sunset, then at sunset.

Panoramic JPG
tahoe sunset

Processed RAW
tahoe sunset

April 29, 2017

Skyfire taken at sunset while out playing with my new drone.  Shows 50%-60% chance of photogenic sunset.


Processed DNG
tahoe sunset

May 1, 2017

Skyfire on the left, SunsetWx on the right, taken 6 hours before sunset.  Skyfire shows clear skies and 0% chance of photogenic light. SunsetWx rates sunset as “average” to “vivid”.

skyfire and sunsetwx

Panoramic JPG
tahoe sunset

May 4, 2017

Skyfire on the left, SunsetWx on the right, taken 5 hours before sunset.  Skyfire shows 50%-60% chance of colorful sunset.  SunsetWx rates sunset as “poor”.

skyfire and sunsetwx

Skyfire updated at sunset, showing even lower chance of colorful sunset.

Panoramic JPG
tahoe sunset

May 5, 2017

Skyfire on the left, SunsetWx on the right, taken at sunset.  Skyfire shows 70%-90% chance of colorful sunset.  SunsetWx rates sunset as “average”.

skyfire and sunsetwx

The color during this sunset never quite “popped”.  Low clouds to the east blocked any further light after this point.
Processed RAW
tahoe sunset

May 8, 2017

Skyfire on the left, SunsetWx on the right, taken 2.5 hours before sunset.  Skyfire shows 60%-80% chance of photogenic sunset.  SunsetWx rates sunset as “average” to “vivid”.

skyfire and sunsetwx

Skyfire updated at sunset, lowering chance of photogenic sunset to 50%-60%.

Processed RAW
tahoe sunset

May 9, 2017

Skyfire on the left, SunsetWx on the right, taken 5.5 hours before sunset.  Skyfire shows 60% chance of photogenic sunset.  SunsetWx rates sunset as “above average”.

skyfire and sunsetwx

Processed RAW
tahoe sunset
tahoe sunset

May 14, 2017

Skyfire on the left, SunsetWx on the right, taken 2.5 hours before sunset.  Skyfire shows a 60%-70% chance of a colorful sunset.  SunsetWx rates sunset as “poor”.

skyfire and sunsetwx

Skyfire updated at sunset, bumped up to 70%-80%.

Panoramic JPG
tahoe sunset

Processed RAW
tahoe sunset


This is where rating these is hard.  Judging a sunset is mostly subjective.  What I think is a good sunset, what I’m hoping the light will do, is different than the next guy.

As I mentioned earlier, I still tried to check this every night even though I could not take photos. I found those nights to be very inconsistent with the results.


I like playing around with Skyfire, especially in conjunction with TPE.  But is it worth the money?

From these studies I found it to be accurate, by my judgement, just over 50% of the time.


The only thing I don’t like about SunsetWx is the scale of their map – the inability to zoom in.  Local terrain can influence sunset factors, especially where I live, and it’s hard to make these out in SunsetWx’s national map.

I also found SunsetWx to be accurate about 50% of the time, again by my own standards for how I rate a sunset.

Overall thoughts on Skyfire vs. SunsetWx

I like SunsetWx’s rating methodology better than Skyfire.  They try to tell you what the “wow” factor will be.  Whereas Skyfire tells you “50% chance of sunset” much like someone will say “50% chance of rain”.  Okay, well, what kind of rain?  And how much?  This is why I prefer Vivid, Average, and Poor ratings more than percentages.

But then again I’m in love with The Photographer’s Ephemeris, and the added functionality of the sunset prediction tool in Skyfire is a big bonus.  Even though it’s correct just over half the time by my standards through these testing days.

I’ll continue to keep SunsetWx bookmarked on my phone and check it often.  I’ll also keep my subscription to Skyfire – it’s only the price of a Starbucks each month (sad how nowadays we relate prices to “how many Starbucks is that”).  However I’ll probably really only trust Skyfire when the overlay is dark orange-red.

Back when I was teaching the next generation of fighter pilots we’d always tell the students, “don’t get so caught up in the gadgetry that you forget to look out the big window.”  I think the same applies to predicting sunsets.  Looking at the sky, knowing the forecast, and predicting the sunset yourself is a skill you should be able to pick up with some practice.

Use the software to augment your own predictions but don’t make them your primary tools. Your primary tool as a photographer is your eyes.

What are your experiences with these services?

Share this:


Get custom newsletters based on your interests!

Send me this newsletter:

You have Successfully Subscribed!