Measuring Snowfall with Rulers...and Satellites

I was born and raised in Michigan’s Upper Peninsula.  If you’re unfamiliar with this extreme northern region of Michigan, it is meteorologically famous for its snow.  Lots of snow.  Insane amounts of snow.  Sled-from-your rooftop piles of snow (see below image). Driving a car in this winter wonderland is difficult since pavement is merely a rumor on most streets from November through April.   While snow provides many anxious driving moments, it is also cherished for creating breathtaking winter landscapes and for providing a wide variety of winter outdoor recreational opportunities.  Because of my fond snow-laden childhood memories, snow still excites me as an adult, both personally and professionally.  As an atmospheric scientist, I’m naturally drawn to the topic of mapping global snowfall using data collected from satellites such as the recently launched NASA-JAXA Global Precipitation Measurement (GPM) mission.

Typical snowfall scene in Michigan's Upper Peninsula from the author's childhood home, circa the mid-1980's.


Because of its abundant snowfall, I use the Upper Peninsula of Michigan as an outdoor laboratory to study snowfall and collect data that will improve our ability to estimate snowfall from satellites.  Why does it snow so much in Michigan’s Upper Peninsula?  The answer lies directly to the north.  Lake Superior is one of the largest freshwater reservoirs in the world, and this large lake’s rugged shoreline defines the Upper Peninsula’s northern border.  In this land of harsh and lengthy winters, Lake Superior serves as a giant snow making machine that helps produce between 150-300 inches of snow during a typical winter season.   Since Lake Superior is so large and deep, it rarely freezes entirely  - although it did fully freeze during the severely cold 2013-2014 winter and is almost completely frozen as of today (February 23, 2015).  When very cold Canadian air flows over the unfrozen waters of Lake Superior, “lake effect” snow results from the turbulent exchange of heat and moisture between the relatively warm lake surface and frigid air directly above it.   Lake effect snow falls from banded cloud structures that are commonly observed over Lake Superior - and all of the Great Lakes - during winter months.

Visible satellite image of a recent lake effect snow event over Lake Superior observed by the MODIS instrument.  The banded cloud structures shown in the image are typically associated with lake effect snow.  Image courtesy of the Cooperative Institute for Meteorological Satellite Studies at the University of Wisconsin-Madison.


Measuring snowfall from the ground, much less from space, is not an easy task.   Scientists rely on a combination of instruments to measure snow accumulations on the ground, including precipitation gauges (cylinders that capture snowfall), snow boards (a white, flat board used to measure how much snow has fallen with a ruler), and automated instruments such as “hot plates” (a hot plate measures the amount of power needed to melt snow to obtain the snowfall rate or intensity).   When snow boards and precipitation gauges are only available, snowfall measurements are complicated by wind that may cause drifting on the snow board and may make it difficult for a precipitation gauge to effectively collect snowflakes.   Snow also settles and compacts if it accumulates over long time periods, so snowfall depth measurements should be taken as frequently as possible to minimize this issue.  Scientists and weather enthusiasts around the world monitor snowfall every day using these basic tools, including GPM-related outreach partners at schools located around the country.   These observations are critical to tell GPM scientists if it snowed, and how much it snowed, in locations that the GPM Core Observatory satellite flies over.   Since we do not have enough humans with rulers to measure snowfall around the globe - especially in areas like Greenland, Antarctica, over oceans and other remote regions that are sparsely populated - we rely on satellites like the GPM Core Observatory and partner satellites to fill in these observational data gaps.

Precipitation gauge (left), snow board (center), and "trail" to the snow measuring equipment (right) at James Madison Memorial High School in Madison, WI.  Images courtesy of Mr. Tyler Spence, a GPM-GLOBE partner educator.


The GPM Core Observatory satellite doesn’t have a ruler to measure snow.  It instead measures snowfall by interpreting radar and microwave radiometer observations that it collects as it orbits the earth.   The relationship between observed GPM radar and radiometer signals and snowfall intensity is complicated, however, by nature’s ability to produce a dizzying array of snowflake types and sizes.   Some snowfall is comprised of very large snowflakes that are light and fluffy.  Other snowfall events may contain smaller snowflakes that are more dense and laden with water.   Understanding the radar and radiometer signatures associated with different types of snowfall is an active area of GPM-related research.

An example of very "fluffy" snow accumulating on a branch in northern Wisconsin.   The GPM Core Observatory's radar and radiometer measurements can be drastically different depending on the type of snow (for example, large fluffy snowflakes versus dense snowflakes).


We recently deployed two ground-based instruments in the Upper Peninsula to help us better understand how different snowflake populations produce different types of radar and radiometer signals.  One instrument is called a Micro Rain Radar (MRR).  The MRR is a radar that points straight up and collects high resolution data in the lowest parts of the atmosphere.   The other instrument – the Precipitation Imaging Package (PIP) – is a high-speed camera system developed by NASA to take pictures of snowflakes.   The PIP automatically measures snowflake sizes, fall speeds (how fast the snowflakes or raindrops fall), and density.   We have observed some very interesting trends over the past year with our combined MRR and PIP measurements.   For instance, Upper Peninsula lake effect snow is generated from very shallow clouds that are usually less than 1.5 kilometers (about 5,000 feet) deep, and lake effect snowflakes are often very large and fluffy.  Other snowfall events associated with deeper cloud structures (generally over 3 kilometers or about 10,000 feet high) contain smaller snowflakes, but a very large number of snowflakes compared to lake effect snow.  These valuable datasets will be directly compared to GPM observations and will improve our ability to detect and estimate snowfall rates from GPM’s instruments.   We look forward to many more snowy days in the upcoming months to benefit GPM snowfall science!

Scientists deploying a Micro Rain Radar (left) and Precipitation Imaging Package (right) at the Marquette, MI National Weather Service Weather Forecast Office in support of the GPM mission.


It must be amazing to study snow- after having experienced it firsthand throughout your childhood. That is one of the things that I think it so wonderful about being a scientist- you wonder about something and then use different types of technology to try to find the answers- like a detective who is seeking to unravel nature's mysteries!