Machine Learning & Atmospheric Processes Group

Global land surface microwave emissivity estimates are critical for data assimilation, numerical weather prediction, and quantitative precipitation estimation, yet explicit calculations at multiple frequencies remain problematic. We introduce a machine learning model that converts 10.65-GHz H-polarization emissivity retrieved under clear-sky conditions into emissivity for the 12 higher-frequency channels of the global precipitation measurement (GPM) microwave imager (GMI). The model is a feature-tokenizer transformer that ingests ancillary surface variables such as skin temperature, leaf area index, soil texture, and soil moisture. Treating each variable as an individual token lets the transformer direct attention between features and learn nonlinear, spatiotemporally consistent relationships across five years of data [2015-2019]. The network reproduces channel-specific emissivities with R^2 0.80 and mean absolute percentage error < 1%, and it simulates top-of-atmosphere brightness temperatures with R^2 = 0.77-0.95 and mean biases <4 K for every channel up to 89 GHz when compared to collocated GMI observations. Using the independent Tool to Estimate Land Surface Emissivities at Microwave Frequencies Version 2 (TELSEM2) 10.65-GHz H-pol inputs produces similar skill, suggesting that the learned mapping is robust to the input source, and captures transferable physical structures. The model is lightweight and can be embedded in satellite retrieval systems, reanalysis pipelines, and Earth-system models to accelerate microwave forward modeling and improve global precipitation monitoring.

Machine learning is an increasingly popular tool in the geosciences, offering new approaches to numerical weather prediction and complex dataset analysis. However, as reliance on these techniques grows, pressing questions about model transparency, internal biases, and trust emerge. While post-hoc explainability analyses can provide insights on how neural network (NN) outputs are generated, a robust framework for interpreting internal decision-making remains underdeveloped. We address this challenge by exploring a framework to better understand the inner structure of NNs using sparse autoencoders (SAEs). With simplified multilayer perceptrons (MLPs), we demonstrate that hidden layer neurons often exhibit polysemantic behavior where each feature is mapped to a linear combination of neurons, creating an overcomplete representation. This phenomenon, known as superposition, arises when networks encode more features than available neurons, causing neurons to respond to multiple, seemingly unrelated inputs. By introducing a regularized SAE that learns from the original MLP’s activations, we can disentangle these representations, resulting in a 33% reduction in the average number of sensitive inputs per neuron. Applied to a precipitation classification model, this framework reveals evidence of monosemantic behavior in which neurons respond to a single meaningful concept tied to specific physical phenomena such as temperature and fallspeed thresholds for precipitation phase partitioning. We observe similar monosemantic behavior in SAE activations from a snowfall rate regressor related to particle concentration intensity, and vertical radar structures. This framework supports the development of more physically consistent interpretations of hidden neuron activations and improved trust in operational ML models across the geosciences.

High-quality hydrometeor microphysical observations are essential for accurate precipitation estimates and for evaluating weather and climate models. However, analyzing these properties is challenging due to their high variability, complex interactions, and large data volumes. In this study, we examine more than 1.5 million minute-scale rain and snow particle attributes and collocated meteorological variables from seven global measurement sites over 9 years. Applying Uniform Manifold Approximation and Projection (UMAP) for nonlinear dimensionality reduction, we reduce the dataset’s dimensionality by 75%, identifying nine distinct precipitation groups and associated particle evolution pathways. UMAP effectively captures the global structure of precipitation phases such as rain, snow, and mixed-phase types, revealing clear patterns that linear methods struggle to resolve. The resulting UMAP manifold offers a unique perspective on precipitation phase and intensity, advancing our understanding of particle evolutionary processes and offering valuable insights for improving weather and climate models and remote sensing precipitation estimates. Analyzing rain and snow particles reveals hidden patterns, improving precipitation classification for better model predictions.