Using Photo Ephemeris Web, Part 8: Visual Search

This tutorial applies to Photo Ephemeris Web.

Visual Search is a tool within Photo Ephemeris Web that allows you to determine when you can see the sun, moon or galactic center in a specific place in the sky.

Why might you want to do this? Here's one reason:

Moonwalk from Reel Water Productions on Vimeo.

This spectacular sequence was planned with the help of Photo Ephemeris.

The key to getting these sorts of shots is identifying the windows of opportunity that put the moon into the frame from your shooting location. Oftentimes you have some latitude as to where you shoot from, but sometimes that's not the case - particularly when the sun or moon must align with specific features in the landscape or skyline.

Starting out with Visual Search

Availability: Visual Search is available to all users of Photo Ephemeris Web. A PRO subscription is required to use it in conjunction with the moon or galactic center.

Navigate to the 3D Sphere page to use Visual Search.

You'll see the controls on the right hand side. You can try your first search with a single click, using the default values. The controls are displayed to the right of the 3D sphere:

By default, the form is set up to search for the position of the center of the Sun over a period of 6 months from the current selected date. The target area is set to +2 to +10° degrees above the horizon and an azimuth of 86° to 94° (i.e. roughly due east). 

Note that the Range checkbox is checked. This is usually - but not always - a good option. It tends to find the broadest set of results. See the discussion below for when to use position target mode  instead.

In this example, the primary map pin is placed on Timbuktu. Click Search to execute the search and see the results:

Each result shows the following details:

• A sparkline indicating the path of the sun through the search area for the given date
• The date and the start and end times of the passage of the sun through the search area
• The rough duration of the passage through the search area

You can click on any result to set the date and time. There are three options:

1. Click the start time for when the sun enters the search area
2. Click the end time for when the sun exits the search area
3. Click anywhere else on the result to set the time to midway between start and end times (but see below for some exceptions to this rule)

By default, clicking the search result will simply update the selected date/time so you can visualize the position of the sun in the sphere. If you hold down the Command/Windows key, the app will navigate back to the map page upon clicking a result.

Let's click the first result (21 Sep 2020) and look at what we see on the sphere (click without holding the Command/Windows key):

In this example, I've zoomed in and rotated the sphere so we can see the position of the sun in the target area. See here for instructions on adjusting the view of the sphere. The time is set to 6:28am - the mid-way point of the sun's transit across the target area. If we click the 6:11am and 6:45am blue links respectively, you will see the sun entering and exiting the area:

Position Target Mode

Click back onto the Visual Search tab and then uncheck the Range mode checkbox. The displayed fields change as shown:

Instead of targeting an area of the sky, we're now targeting a specific azimuth and altitude. In addition, you will see two fields, Tolerance and Altitude priority:

• Tolerance: tolerance is the allowable difference (plus or minus) in degrees between the calculated position and input position, used to determine if a result is acceptable or not
• Altitude priority: when checked, the tolerance is applied to the azimuth; when unchecked, tolerance is applied to the altitude (i.e. azimuth has priority)

We'll discuss how to use these fields in more detail below, but first, let's re-run the search (click Search) and look at the results:

There are some immediate differences apparent. Firstly, the result is a moment in time, rather than a time span. For Sep 21, 2020, you may have noticed that the time is 6:28am - the same time as the mid-point of the range result we clicked previously.

The 'sparkline' is no longer a line, per se, but instead indicates the position of the sun relative to the target position. You may have noticed that for all results the sun lies directly on the horizontal line. This gives a clue as to what altitude priority does: between azimuth and altitude, altitude is given priority: the app allows some slack in azimuth (within the given tolerance) but finds results that (pretty much) exactly match the specified altitude.

Let's re-run the search now with Altitude priority unchecked:

Two things to notice:

1. There are only two results instead of 15
2. Both results match azimuth exactly

That there are fewer results is not evidence of any problem or malfunction: it is simply that the sun passes through 90° within the allowable range of altitude on only two occasions for the search duration and start date. Looking at the 'sparklines' showing the sun always below the target altitude, you might think 'I bet we'd get more results for 90° if we reduced the altitude target to, say, +3°'. Let's try it - click back to the Visual Search form and change the altitude field to +3°, then run the search again:

Indeed, we see five results instead of two. The final result, March 22, 2021, shows one more feature: if the result matches both azimuth and altitude to within 0.25° (roughly the semidiameter of the sun and moon), the position indicator is shown in green - i.e. a good match.

Why is altitude priority recommended?

Altitude priority performs better in general than azimuth priority. The app uses a variety of interpolation algorithms under the hood to obtain results efficiently (rather than brute force tabulation of the position of the sun and moon for every minute of the day for months or years!).

In the Arctic and Antarctic circles and in the Tropics, azimuth can vary over time in ways that make it unsuitable for standard interpolation techniques (think higher order polynomials - or rather, don't, it's just brain ache!), and so for some locations and dates, the interpolation may fail to produce acceptable results. This is particularly the case when searching azimuths near to the meridian (i.e. when azimuth equals 0° or 180°).

Instead, we recommend using altitude priority or enabling range mode.

Azimuth priority can be useful when the body in question moves predictably across the sky with only a slow change in altitude - this is the case for the Galactic Center in many locations. Azimuth priority works well for the Galactic Center - but you might be better off simply using Range mode instead.

Should I use Range or Position Targeting?

For general landscape use, where you need the sun or moon in a particular area of the sky with some latitude allowed, Range mode is the better choice.

However, if you are planning a shot where, for example, the sun should appear sat on top of a mountain, e.g. the so-called 'Diamond Mt. Fuji' shot, then position targeting is probably a better choice.

Use with Geodetics

Let's switch over to the moon. We're going to find dates when we can see the moon perched atop Mount Sanitas as viewed from the Community Gardens here in Boulder, Colorado. You can get to the right location using this link. Switch to one of the topographic map options for this part of the tutorial.

You'll need to be familiar with using Geodetics - if you haven't used it before, check out Part 3 of this series.

Enable Geodetics, and then drop the secondary pin on the summit of Mount Sanitas, due west from the primary pin (or use this link):

Clicking back to the 3D Sphere, the secondary pin is now shown and a button is available to set the target position for Visual Search to the geodetics bearing and elevation angle:

Click 'Use geodetics' to update the position target. While we're there, we'll switch the Body to the Moon and the duration to 2 years:

You can see that the 3D sphere shows the target area in blue (for the moon), and that it is centered behind the secondary pin. Running the search and choosing the first result, we see the following (with a little adjustment of the 3D sphere viewpoint for illustrative purposes):

The moon lies in the target area, slightly below the grey pin - we might do better on Oct 31 when it lies above the summit.

As you've seen, using Geodetics together with Visual Search provides a very easy way to find dates when you can find the moon exactly where you need it.

Filters

Typically you need other conditions in place as well to make a shot of the moon or galactic center. For example, you may want a full moon, or you may want to shoot the Milky Way on a moonless night.

Filters allow you to find the dates that match your required conditions. Let's use filters to refine our moon search results. Click the Filters tab:

There are two ways to use filters: the easy way or the slightly harder way. Let's use the easy way first - it'll make it much clearer how the 'harder' way works. Click "Full Moon", then Apply:

The Suggested filters (available for the Moon or Galactic Center) are simply shorthand ways to set filter conditions. In the example above, clicking Full Moon sets a condition that the Moon illumination fraction should be between 0.95 and 1.0. On applying the filter, only 10 results match - the count is shown in the Results tab title.

You can apply multiple filters either by selecting additional suggested filters or by adding filters manually. To add filters manually, click '+'; a new filter for Sun altitude will appear. Change this to the criterion you want using the drop-down menus.

For suggested filters, the only rule is one suggestion per body - i.e. if you choose Civil Twilight, it will set the sun altitude filter to -6° to 0°; if you then choose Nautical Twilight, it will replace the existing filter for the sun to require altitudes between -12° and -6°, rather than extend the range. Any filter conditions for the moon will remain in place.

Click 'Nautical Twilight' then Apply - we have two matches:

• Suggested filters for the Moon include:
  • 'Supermoon' - a perigee full moon
  • Full moon - 95-100% illuminated
  • New moon - 1-5% illuminated and waxing
  • Cresent moon - 1-25% illuminated
  • Near sunrise/sunset - sun between +2° and -2° altitude (good times for shooting the moon)
  • Civil/Nautical twilight
• Suggested filters for the Galactic Center include:
  • Astronomical twilight - sky is almost dark and stars are visible
  • Darkness - after/before astronomical twilight
  • Moonless night - darkness, plus the moon below -6°
  • Crescent moon - 1-25% illuminated (enough moonlight to illuminate a landscape, but not too much to overly impair the stars)

Manually creating filter conditions allows even more flexibility, but the suggestions should cover most planning situations.

Range Results & Optimistic vs. Pessimistic Filters

If you're using Range mode to conduct your search in conjunction with filters, it's possible that the filter conditions are met for only part of a potential result. In this case, filters are applied 'optimistically', i.e. if the conditions are met at start OR end of the range, then the result is kept. It is only filtered out if the filter conditions are not met at both extremes of the range result.

For example, when searching for the Galactic Center on a moonless night, the moon may rise above it's max altitude threshold of -6° degrees during the duration of the result, but if it is below the threshold at the start of the time period, the result is retained - after all, your filter conditions were met, just not for the entire duration.

(Why can't we limit the range to the time when the filter conditions are met? That would be possible, but it would require additional interpolation or brute force calculations on multiple variables - we'll leave that as something for the future...)

This consideration does not arise in target mode, as each result reflects a moment in time for which the filters can be evaluated just once.

A note on search duration

The default search duration is 6 months from the selected date, but you can increase it to up to 5 years. Generally speaking, you don't need to search for longer than 1 year if you only care about either the Sun or the Galactic Center. The positions of these vary only by a very small amount year on year, so generally a 12 month search will suffice.

The same does not apply to the moon. If your search is for the moon - or if you plan to apply filters that include the moon - you may want to increase the search duration to 2 or 5 years.

The moon follows roughly a 19.5 year cycle, meaning that its position varies year by year. You can't simply look out 12 months ahead and expect to cover all the opportunities where the moon may show up in your search range or target.

Disc Alignment

One setting we've skipped past so far is Disc Alignment. For the sun or moon, you can specify the preferred alignment for the body (center - default, upper limb - top, lower limb - bottom).

For example, if you want the moon to appear sat atop a mountain, use lower limb, so the moon is fully visible. If you want a sun star as it clears a ridgeline, choose upper limb, so you catch the sun just as it appears/disappears - this would be useful for the 'Diamond Mt Fuji' shot, for example.

Edge Cases

If you've made it this far, you might be interested to learn about the exceptions to range mode results that I hinted at earlier. These can arise if the azimuth range you specify spans the meridian. In these cases, it is possible for the body (sun/moon/galactic center) to rise through the target range and then set back through it once again. Here's an example for Iceland in 2022 as we approach the major lunar standstill:

With a wide enough azimuth range, you may see 'compound' results with two crossings through the target range.

In these circumstances, clicking the result once will set the time to the mid-way through the first crossing. Clicking it again will set it to the middle of the second crossing.

It's also possible that the interpolation algorithm will obtain incomplete results in some circumstances and that only one of the two crossing may be selected by clicking the result listing.

Conclusion

Congratulations on making it through to the end! We'd love to hear how you find Visual Search - tell us your successes and let us know how we can make it better.

Next up: Using Photo Ephemeris Web, Part 9: Maps and Terrain in the 3D Celestial Sphere