Zooniverse: August 2017

Zooniverse: August 2017

Citizen Scientists Help Improve Our Understanding of Forest-Snow Interactions

Imagine having hundreds of thousands of images like the ones below and needing to classify the snow in the trees and clouds in the sky. It would surely be great to have thousands of people willing to help, but how would you go about finding them?

 

But before we get to addressing that cliffhanger, we’ll take a step back to what you also might be wondering: Why would you need to classify clouds and snow in thousands of images anyway?

Watersheds can be significantly covered by forest, and snow provides a natural storage of fresh water. Forests can intercept up to 80% of the total annual snowfall. Intercepted snow often either melts (quicker than the snow underneath the canopy) or is sublimated back to the atmosphere. Melted snow in the canopy affects the timing that water is available for agriculture, hydropower and drinking water. Sublimated snow is lost back to the atmosphere, decreasing the amount water storage in our watersheds. Therefore, forest interception plays a vital role in our understanding of how much snow is in forests and ultimately how much water we have available.

In 2014, our UW Mountain Hydrology group began collecting time-lapse camera images for NASA’s OLYMPEX project to monitor snow depth in open areas of the western United States. We got pretty savvy with these time-lapse camera images and more recently began to utilize time-lapse camera images from around the western United States to monitor the amount of snow in trees. With these images, we are able to evaluate how much frozen precipitation was intercepted in trees and how quickly the snow in the forest melts or sublimates – ultimately improving our ability to model these processes.

So now back to the cliffhanger. We were a bit overwhelmed with thousands of images needing to be classified, but luckily Bart Nijssen, an Associate Professor at the University of Washington, suggested we check out the Zooniverse website. Zooniverse is a platform to enable “citizen science”- a term used to describe the engagement of the general public in collecting and analyzing data as part of a collaboration with scientists. This seemed like the perfect way to not only get volunteers to help with processing our images, but also to involve the general public in our research and help educate them about the role of snow and trees in a watershed. We’ve been thrilled by the response from the citizen scientists and have branched out to include even more images from different research projects around the United States.

To date, we have had 4,750+ registered volunteers classify images from over 23 different locations. This corresponds to 20,000+ images and over 350,000 classifications (a single image is classified more than once). We’ve found that the accuracy of these classifications for whether there’s snow in the trees to be ~95%.

We are so thankful for the thousands of citizen scientists that have helped with this project. They have already improved our understanding of forest snow interception and how much fresh water storage we have to last us through the dry season. You too can become a citizen scientist for our project by clicking here.

We hope you enjoy the images as much as we do!

For more information, please contact: Max Mozer: mozerm@uw.edu, or Ryan Currier: currierw@uw.edu

About the Authors:

Max Mozer is an undergraduate research assistant in the Mountain Hydrology Research Group in the Department of Civil & Environmental Engineering, University of Washington.  Max specializes in forest-snow interactions and managing the field office.  His favorite watershed is the Skokomish River Watershed in Olympic National Park, WA.

 

Ryan Currier is a graduate student in the Mountain Hydrology Research Group in the Department of Civil & Environmental Engineering, University of Washington.  Ryan specializes in hydrometeorology and forest-snow interactions. His favorite watershed is the Quinault River Watershed in Olympic National Park, WA .