Published: March 23, 2018

A new study from CU Boulder uses a historical data set to analyze the air temperature threshold for rain-snow events. Research led by INSTAAR graduate student Keith Jennings shows that the air temperature threshold differentiating rain from snow events varies significantly across the Northern Hemisphere.  This in effect highlights the inaccuracies of many land surface models that use 32 degrees Fahrenheit as a uniform threshold to predict precipitation phase.

Spatial variation of rain vs snow events

The observed 50% rain–snow Ts threshold over the Northern Hemisphere for 6883 land stations from 1978 to 2007. Each point represents one station and only stations with a sufficient number of snowfall events were analyzed. a Thresholds mapped by station location. b Thresholds plotted by station longitude. The horizontal dashed line represents the Northern Hemisphere mean threshold (1.0 °C), the shaded gray box covers thresholds within ±2 standard deviations of the mean, and the blue line is a generalized additive model fit to the threshold data by longitude. Regions of interest are denoted by text within vertical dashed lines. Map provided by Co-author Molotch.

By using nearly 18 million precipitation, humidity, and temperature observations across the Northern Hemisphere, the study presents the first continuous Northern Hemisphere map of precipitation phase partitioning. Continental climates show a warmer rain-snow threshold, while coastal climates show cooler rain-snow threshold. Additional findings show that incorporating humidity in model simulations improves model performance. Former INSTAAR M.S. student Taylor Winchell, CEAE assistant professor Ben Livneh and CWEST Director Noah Molotch co-authored the paper published this week in Nature Communications.

Snow versus rain events plays a significant role in both the global hydrologic cycle and the climate system, where snowfall and rainfall have opposite effects on land surface water and energy fluxes. Importantly, snow accumulation increases surface reflectance of solar radiation and stores water that over a one billion people rely on over the world. As the climate warms, rain events are occurring at a higher proportion than snow events. With less snowfall, snow water equivalent decreases, snowmelt occurs earlier, and streamflow reduces. Further, experts predict more rainfall relative to snowfall and more rain-on-snow events in the future, which increase flood risks. Thus, it is with increasing importance that meteorologists improve forecasting for rain versus snow events.

The results of this study suggest that using humidity and air temperature in large-scale land surface model runs may improve precipitation phase predictions. With more research, these findings will contribute to improved water management and flood mitigation planning.

From a local perspective, Denver’s 9News interviewed Jennings and Livneh to help Coloradans better understand how and why they are receiving snow. Colorado has one of the warmest snow-rain thresholds, where we may get snow precipitation around 40 degree Fahrenheit. While the snow forms high in the atmosphere below 32 degrees Fahrenheit, the low relative humidity keeps the snowflake cold through evaporative cooling as it falls through variable temperatures in the atmosphere. CU Boulder Today covers more about this study in their article, Rain or Snow? Humidity, location can make the difference.