Ottawa
Astronomy Weather

The Search for Clear Skies

Sponsored by:

It is a frustrating fact for amateur astronomers that commonly available weather forecasts are overwhelmingly inaccurate or too inappropriate for planning observing sessions. Usually they concentrate on temperature, wind and precipitation and not on the exact hour when the sky will clear. Furthermore common forecasts consider haze or overcast high cirrus to be clear skies. Common forecasts don't work for astronomers because they are designed for the general public.

This page is a collection of links to the best pages I've found for forecasting good astronomical observing weather. Many of the links are good for all of North America. A few are specific to southern Ontario or Ottawa.

Near-Realtime Clouds

A very reliable way to know if it will be cloudy or clear in the next hour or so is to look at current satellite images. The best of these are visible light images because they show the clouds as we see them.

NASA Global Hydrology and Climate Center: 1km Visible Light Image

This is a very useful image because it is usually updates every 15 minutes. Moreover, you can click on any location to zoom in to a 1 kilometer resolution image. Click on the old sample at the left to see full resolution, or here for the latest image:

Not only can we see just where the sucker holes are, but there is so much detail that we can often see if the cloud cover is continuous or broken.


As wonderful as Nasa's visible image is, it doesn't work at night. Urk! The usual solution is to use infrared satellite images. But IR images don't measure reflected light, which is how our eyes work. They measure the temperature of things. Usually they are color-coded to show warm areas as dark and cold areas as white. As temperature usually decreases with height, clouds will be grey or white against the dark, warmer ground. But it can be hard to distinguish the ground, meaning clear skies, from thick haze, fog, or warm low clouds. Temperature inversions, or simply cold winter conditions, further complicate interpretation.

There is a better way to see cloud cover at night.

NOAA Operations Products Development Branch : GOES Nighttime Fog and Low Cloud Imagery

This is one of the best satellite images I've found for showing conditions at night. Unlike a single-band image, it's actually the difference between two images in different infrared wavelengths. This cleverly allows to distinguish between fog and low cloud, as well as highlighting high cirrus (shown as cyan.)

Click here or the sample at the left for the latest image.

The only problem with this image is that is updated rather infrequently.


My own satellite image "product"

This image is my attempt to reproduce the NOAA's Fog/Low cloud image but zoomed in onto south Ontario and updated every 15 minutes. (Or as often as NASA downloads images from GOES8.) I use an algorithm based on NOAA's own.

In this image, black means clear skies. White is that nasty opaque stuff that we don't want. Cyan means fog. No wait, it's really high thin cirrus. Err, no. Maybe its warm low clouds. In any case, it's not transparent skies. I'll be clearer about the explanation when I find some GOES8 calibration data.

Reload this page to see the latest image.

The image is deliberately small so it will load quickly. Anyone is welcome to display this image on their own web page. To do so, insert this html:

<img src=../t/currcld.gif> And reload your page to get the latest image.

An unusual thing about this image is the yellow arrows. They are an attempt to show what is approaching Ottawa, where I do most of my observing. <geekspeak>I am using a least-squares-fit sliding-peephole algorithm to match shifted graphical patterns between two images.</geekspeak> If the arrows are a jumbled mess, my program was unable to determine the direction of movement. If the arrows form a straight line or sweeping curve, they are probably accurate and indicate the motion of clouds.

Writing programs to generate your own satellite pictures is tons of fun. If you'd like to try your own hand at it, the raw data can be found at http://rsd.gsfc.nasa.gov/goese/autogvar/goes8/

Numerical Forecast Models

Close-to-realtime satellite images are excellent, but unless you are a weather expert, they are only good for forecasting the next few hours. Beyond that, the best I've found are numerical weather forecasting models.

The models are big computer programs which create weather maps and even simulated satellite images for times in the future. The satellite-image forecasts are limited to 48 hours.

Canadian Meteorological Center Numerical Forecast Models

Some of CMC's numerical forecasts are unique that they are specifically designed for astronomers. While most other numerical forecasts concentrate on predicting rain, wind or tornados, these try to predict clear skies!

Each of CMC's three forecast products predict up to 48 hours into the future, with synthetic images for every three hours.


Clouds forecast (visible) for astronomical purposes

The image forecasts percentage of cloud cover. It is very much like a visible satellite image except that it works for nighttime hours too. Dark blue areas indicate clear skies. There may, however, still be haze. Read on.

Click on the image to see full resolution.

Sky transparency forecast for astronomical purposes.

Each point on this image shows the total amount of water vapor over a spot on the earth from ground level to space. In other words, it represents just what astronomers call "transparency". Black areas are driest, and also clear of cloud. Dark grey may be also clear, but will be hazier. The lightest areas will almost certainly be cloudy, at least for non-tropical regions. White areas are where other numerical models forecast clouds.

This is an excellent forecast for deciding if you will have good skies for faint fuzzies, or hazy skies where planets and double stars might make better targets.

Click on the image to see full resolution.

The Cloud Cover and Transparency numerical forecast may well be the most radical advance in forecasting clear skies for astronomers. They also seem to be unique.

However, their future is not entirely certain. If you find them useful, please email Allan Rahill of the CMC. Allan needs to show his boss that his astronomy forecasts are actually used.


The Ottawa Clear Sky Chart

CMC's forecast images are excellent. But if you want to find out what hour tonight it will finally clear up, you may have to hunt through 533 images. That is a lot of download time.

To make that task easy, I have written a program to generate, as often as CMC updates their web pages, the image at the left. It is very much quicker to download than any of CMC's images.

It shows the data from all of CMC's astronomical forecast images, but just for one location, in this case, Ottawa. At a glance, one can tell if weather conditions permit observing. With a bit of practice one can tell at a glance whether it's time to do deep sky observing, planetary observing or say indoors and process images.

An explanation of how to read the Chart, including the third "Darkness" line is on the legend page. That page also has a copy of the Chart which is clickable. Clicking on one of the colored blocks there will scoot you to the corresponding CMC image. This makes for a very quick way of navigating CMCs many images.

Clear sky charts for locations other than Ottawa are available and you are welcome to put them in your web pages. See the Clear Sky Chart Home Page


Danko's Homepage Clear Sky Chart Homepage

.