Snow, a defining feature of New England winters, is as complex as it is fleeting. Its distribution and behavior—how much falls, where it accumulates, and when it melts—are shaped by a web of factors ranging from topography to vegetation. But these patterns are changing, and so are the tools used to study them. Two recent studies from the Hubbard Brook Experimental Forest in New Hampshire showcase how drones equipped with advanced sensors are transforming our understanding of snow and its role in the region’s ecosystems.

Snow isn’t just a uniform blanket covering the ground; it’s a patchwork of conditions shaped by local factors. These “snow microclimates” are the focus of the study Modeling Snow Microclimates within Montane Forests and Ephemeral Snow Environments, led by Jeremy Johnston and colleagues. The researchers note that “snow properties such as depth and persistence vary widely at small scales” and that these variations “frequently develop across distances < 1 km.” Such differences matter because they influence hydrology, local energy balances, and even transportation.

To explore these microclimates, the team used drones to map snow depth and persistence at high spatial resolutions of less than 50 centimeters across two sites: one in Durham, NH, characterized by “ephemeral shallow snow and mixed forest-agricultural land cover,” and another at Hubbard Brook, a colder, forested mountain basin with a “deeper transitional Montane Forest snowpack.” The researchers paired these observations with clustering algorithms and random-forest-based feature metrics to uncover the drivers of snow variability, including “wind attenuation coefficients, temperature lapse rates, terrain, and forest structural variables.” This effort produced not just detailed maps but also a framework for predicting snow microclimates using both physical and proxy variables like canopy density and distance to the canopy edge.

These insights could improve computational models of snow distribution. Johnston’s team used a method they call “Paint-By-Numbers” modeling, which assumes that areas within the same microclimate classification have similar snow properties. This approach simplifies the complexities of snow distribution, providing a practical tool for resource managers and scientists alike.

While Johnston’s work focuses on the texture of snow, Ryan Reed and colleagues at the University of New Hampshire zeroed in on its weight—specifically, the snow water equivalent (SWE). SWE is critical for understanding how much water is stored in snow and when it might be released. Traditionally, SWE is measured using snow courses: points where snow is collected and analyzed manually. But these snapshots often fail to capture the spatial variability crucial for understanding water balance and flood risks.

In their study, Basin Scale Snow Water Equivalent Observations from Uncrewed Aerial Systems (UAS) Lidar at Hubbard Brook Experimental Forest, New Hampshire, USA, Reed and his team used drones equipped with lidar to address this limitation. They flew over a 0.42 km² watershed at Hubbard Brook, combining high-resolution snow depth data with on-the-ground snow density measurements to create detailed SWE maps. These maps revealed spatial patterns that traditional methods missed, especially when it came to “wind-driven variations in snowpack.” The researchers note that “spatially distributed UAS SWE estimates allowed for more accurate water balance calculations than traditional point observations,” underscoring the value of drone technology for capturing fine-scale variability.

To interpret and model their findings, Reed’s team turned to HEC-HMS, the Hydrologic Engineering Center’s Hydrologic Modeling System, a tool developed by the U.S. Army Corps of Engineers to simulate precipitation-runoff processes. HEC-HMS allows researchers to model snow accumulation and melt at both a lumped (basin-wide) scale and a distributed scale with spatial resolutions as fine as 1 meter. By comparing these approaches, the study showed that traditional point observations performed well under certain conditions but fell short in capturing variability driven by wind or terrain. The distributed models, enriched by drone data, provided a much clearer picture of SWE dynamics and enabled more precise water balance calculations.

Scientific modeling, like the HEC-HMS system used by Reed’s team, exemplifies the broader landscape of tools used to study environmental processes. Models like this fall into categories ranging from empirical (relying on observed relationships) to physically based (simulating processes using physical laws) to hybrid approaches that combine the two. Physically based models such as HEC-HMS excel in capturing dynamic, real-world conditions by incorporating detailed input data, but they also demand significant computational power. The integration of drone-generated data enhances these models, bridging the gap between broad-scale predictions and the small-scale details critical to understanding snow’s role in hydrology.

Both Johnston’s and Reed’s studies highlight how drones are transforming snow science. At Hubbard Brook, where decades of research have documented declining snowpack, these methods offer a way to bridge past data with new tools. Accurate measurements of SWE and snow microclimates are vital for managing water resources, predicting floods, and conserving habitats. They also have practical applications in industries like forestry and tourism, which depend on understanding snow behavior.

Snow is more than a seasonal hallmark; it is a linchpin of New England’s environment and economy. Snow acts as a natural reservoir, storing water in the winter and releasing it slowly in the spring. It insulates soils, regulates ecosystems, and serves as a marker of climate health. Yet across New England, snow cover is declining—both in depth and duration—due to rising temperatures and shifting precipitation patterns. These changes disrupt everything from forest dynamics to flood management, making it essential to understand snow’s behavior at increasingly fine scales.

As snow becomes less predictable, the tools we use to study it are becoming more precise. The work by Johnston, Reed, and their teams represents a new frontier in environmental science, one where drones and advanced modeling systems enable us to see snow not as a static layer but as a dynamic, intricate system. By revealing its hidden complexities, these researchers are helping us prepare for a future where snow is both an enduring mystery and a fleeting resource.

Basin Scale Snow Water Equivalent Observations from Uncrewed Aerial Systems (UAS) Lidar at Hubbard Brook Experimental Forest, New Hampshire, USA

  • Ryan Reed (University of New Hampshire)
  • Adam Hunsaker (University of New Hampshire)
  • Jeremy M. Johnston (University of New Hampshire)
  • Jennifer M. Jacobs (University of New Hampshire)

Modeling Snow Microclimates within Montane Forests and Ephemeral Snow Environments

  • Jeremy M. Johnston (University of New Hampshire)
  • Adam Hunsaker (University of New Hampshire)
  • Jennifer M. Jacobs (University of New Hampshire)
  • Ryan Reed (University of New Hampshire)
  • Mahsa Moradi (University of New Hampshire)