Observing the Arctic Water and Energy Cycle: From Snow Cover to Rainfall

Abstract

The Arctic is highly sensitive to climate change, with near-surface air temperatures warming at two to four times faster than the global average. This rapid warming affects key components of the Arctic water and energy cycles, including snow cover and rainfall. Snow cover impacts the surface radiative energy budget by reflecting shortwave radiation and emitting thermal radiation, with its reflectivity and emissivity quantified by surface albedo and spectral emissivity, respectively. Rainfall interacts with various elements of the Arctic water cycle, and can accelerate snowmelt. As climate warms, rainfall is expected to become more frequent and widespread in the Arctic, which is attributable to rising melting layer heights where ice crystals melt into raindrops. Despite the importance of snow cover and rainfall in the Arctic climate system, gaps remain in understanding their dynamics and interactions. These gaps arise from limited observations and challenges in retrieving information from noisy measurements. For instance, the Arctic atmospheric and surface thermal radiation, particularly in the far-infrared spectrum (15 to 100 µm), remains poorly characterized. Additionally, while rainfall’s role in accelerating snow cover melt by reducing surface albedo is acknowledged, quantitative evaluations of albedo changes during such rain-on-snow (ROS) events are lacking. Furthermore, precipitation phase identification poses challenges for understanding rainfall development in the Arctic and its impacts on snow cover. To address these gaps, we developed novel techniques to investigate satellite and ground-based observations of Arctic surface and precipitation processes. First, we develop and assess an optimal estimation-based retrieval algorithm to estimate mid- and far-infrared polar surface spectral emissivity in preparation for the Polar Radiant Energy in the Far-InfraRed Experiment, a satellite mission focused on far-infrared observations. Results indicate that retrievals of far-infrared surface spectral emissivity are largely influenced by the atmospheric moisture content and the choice of a priori constraints. Further analyses using Infrared Interferometer Sounder-D satellite observations highlight the importance of surface spectral emissivity in improving retrievals of Arctic humidity and temperature profiles. Next, we analyze multi-year ground-based observations in northern Alaska and identify ROS events from 2012 to 2022. Results reveal that ROS events distinctly accelerate snow albedo decline in May and June, an effect inadequately represented in models. Notably, ROS events in May are linked to anomalous remote moisture transport, while those in June are associated with local moisture sources. Finally, we develop a machine learning-based framework to improve the detection of melting layers using ground-based radar observations at North Slope of Alaska. Results show that the machine learning algorithm greatly improves the detection accuracy and performs robustly across diverse weather conditions. The machine learning algorithm also provides uncertainty estimates based on ensemble predictions, and can be applied to similar instruments over wider regions. The results of this dissertation offer new insights and adaptable methodologies to address observational challenges in further understanding the Arctic surface conditions and precipitation processes.