Accurate measurements of Sea Ice by Unicode Unicorns 

Accurate measurements of Sea Ice by Unicode Unicorns

The customer

The Unicode Unicorns

The Unicode Unicorns are a data science team founded in October 2018 by Kelby Stockstill and Katelyn Hertel for the NASA Space Apps Challenge. The team, which consists of people with diverse skills from design to programming, set itself to the task of solving real world problems using data. The team won the 2019 NASA Space Apps Challenge NYC with their project For Your Ice Only.

Challenge

Last year during the NASA Space Apps Challenge, the Unicode Unicorns learned that it is impossible to calculate Sea Ice Area during the summer months. Satellite microwave measurements of sea ice are unable to account for water located on top of ice – also known as melt ponds or surface melt – both appear as water, despite ice being present. Without accurate data, researchers must base their calculations on Sea Ice Extent, which does not measure the actual amount of ice present in the Arctic. Given the importance of the Arctic to climate change, it is imperative that cryospheric scientists have as much information about sea ice concentration and distribution as possible. The Unicode Unicorns set out to chase the missing data from the summer months and calculate the true Sea Ice Area in the Arctic Ocean.

Solution

In order to correct for surface melt, the Unicode Unicorns predicted the amount of water that is true open water (rather than surface melt) based on the proportion of water (both open water and surface melt) to ice and the local weather conditions.

Using the Meteomatics Weather API, they were not only able to tailor the data to the regions desired, but also retrieve historical data specific to those areas across the measurement period of several years.

Result

During the 48h NASA Space Apps Challenge, historical data could easily be queried via the Meteomaics Weather API. The historical information on temperature, humidity, wind and radiation was then divided into 1-day, 7-day, and 30-day sets to see which more accurately predicted the open water proportion. A stepwise regression was then used to isolate the most impactful weather metrics while minimizing collinearity between them.

The highest performing model used the 7-day aggregate of temperature and the 30-day aggregate of humidity to predict open water. In the end, the Unicode Unicorns differentiated between open water and surface melt and were able to calculate the true sea ice area in the Arctic Ocean. Using just these measurements, the Unicode Unicorns were able to accurately predict true open water, demonstrating a relatively straight-forward methodology for deriving summer data for climate change scientists.

Kelby Stockstill and Katelyn Hertel, The Unicode Unicorns:

"We would love to thank you Meteomatics not only for your weather data but also for assisting us personally and being there to answer our questions. We would not have been able to do this project and solve this issue in the field of cryospheric science and climate change without Meteomatics!"