Is there snow?

In Season Spotter Questions we ask you about snow. Is there a lot or a little? Is it on the trees or just on the ground? These might seem like strange questions to ask. After all, snow doesn’t tell us that much about phenology. Trees and shrubs can have leaves or not have them when there is snow. Flowers can appear before all the snow has melted in the spring.

snowThe main reason we ask about snow is that it affects our automated processing of image greenness — sometimes in weird ways. You’ll recall that we calculate something called GCC, which is a measure of the greenness of an image. When there are lots of green leaves in an image, we get a high GCC value.

What happens when there aren’t any green leaves in the image? Well, it depends on what the landscape looks like. If the landscape looks yellow or brown because it’s full of dead, dry vegetation it will have a particular GCC value. And if it is covered by white snow, it will have a different GCC value.

Snow is typically white, and so for snowy pictures GCC tends to be about 0.33, because there are equal amounts of red, green, and blue in “white”. But dead vegetation can have a GCC higher than this. Or, it can have a GCC a bit lower. And that’s where the trouble comes in. In some locations, snow on the ground can actually make an image look “greener” than when the vegetation is dead and there is no snow.

As a result, our GCC time series sometimes looks like this for snowy sites:

merbleue-GCCThis particular time series is for the Mer Bleue Conservation Area in Ottawa, Canada. In the summer, the view is rather green.

merbleue_2013_07_18_120104But in the fall, the foliage turns reddish. This makes the fraction of green (GCC) go particularly low.

merbleue_2013_11_05_120103And in the winter, the snow covers the reddish ground, bringing GCC up!

merbleue_2014_03_18_120105If we were just looking at the GCC time series, we might mistakenly think that this site has two green periods — one big one in the summer and a smaller one in the winter. Patterns like this can be seen in tropical and subtropical regions where rainfall controls vegetation greenness rather than temperature. But it would be very odd to see two green periods each year in Canada!

Another problematic thing that snow does is that it adds a lot of noise to some GCC time series. In many places in the U.S., it snows and then the snow melts. Then it snows and the snow melts. And so sometimes there’s a lot of snow in the images, sometimes no snow, and sometimes just patches of snow. And this goes on all winter long. As a result, the GCC values jump around a lot.

At this site at Niwot Ridge in Colorado, we can see that there’s a lot more variablity in our GCC time series in the winter than in the summer:

niwot2-snow-rawBut if we remove the days that have snow or bad images (blurry, etc.) from the graph, we get a much cleaner time series and a better sense of what is happening with the vegetation at that site.

niwot2-snow-fixed

Advertisements

About Margaret Kosmala

I am an ecologist exploring the complex dynamics of plant and animal systems. I am especially interested in understanding how species communities change over time and how humans impact them.
This entry was posted in Why do we ask that? and tagged , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s