Search

Autoencoders are powerful machine learning models used to compress information from multiple data sources. However, autoencoders, like all artificial neural networks, are often unidentifiable and uninterpretable. This research focuses on creating an identifiable and interpretable autoencoder that can be used to meld and combine climate data products. The proposed autoencoder utilizes a Bayesian statistical framework, allowing for probabilistic interpretations while also varying spatially to capture useful spatial patterns across the various data products. Constraints are placed on the autoencoder as it learns patterns in the data, creating an interpretable consensus that includes the important features from each input. We demonstrate the utility of the autoencoder by combining information from multiple precipitation products in High Mountain Asia., Comment: 13 pages, 7 figures

Snow density estimates as a function of depth are used for understanding climate processes, evaluating water accumulation trends in polar regions, and estimating glacier mass balances. The common and interpretable physically-derived differential equation models for snow density are piecewise linear as a function of depth (on a transformed scale); thus, they can fail to capture important data features. Moreover, the differential equation parameters show strong spatial autocorrelation. To address these issues, we allow the parameters of the physical model, including random change points over depth, to vary spatially. We also develop a framework for functionally smoothing the physically-motivated model. To preserve inference on the interpretable physical model, we project the smoothing function into the physical model's spatially varying null space. The proposed spatially and functionally smoothed snow density model better fits the data while preserving inference on physical parameters. Using this model, we find significant spatial variation in the parameters that govern snow densification.

Snow density estimates below the surface, used with airplane-acquired ice-penetrating radar measurements, give a site-specific history of snow water accumulation. Because it is infeasible to drill snow cores across all of Antarctica to measure snow density and because it is critical to understand how climatic changes are affecting the world's largest freshwater reservoir, we develop methods that enable snow density estimation with uncertainty in regions where snow cores have not been drilled. In inland West Antarctica, snow density increases monotonically as a function of depth, except for possible micro-scale variability or measurement error, and it cannot exceed the density of ice. We present a novel class of integrated spatial process models that allow interpolation of monotone snow density curves. For computational feasibility, we construct the space-depth process through kernel convolutions of log-Gaussian spatial processes. We discuss model comparison, model fitting, and prediction. Using this model, we extend estimates of snow density beyond the depth of the original core and estimate snow density curves where snow cores have not been drilled. Along flight lines with ice-penetrating radar, we use interpolated snow density curves to estimate recent water accumulation and find predominantly decreasing water accumulation over recent decades.

Surface mass balance (SMB) is an important factor in the estimation of sea level change, and data are collected to estimate models for prediction of SMB over the Antarctic ice sheets. Using a quality-controlled aggregate dataset of SMB field measurements with significantly more observations than previous analyses, a fully Bayesian nearest-neighbor Gaussian process model is posed to estimate Antarctic SMB and propose new field measurement locations. A corresponding Antarctic SMB map is rendered using this model and is compared with previous estimates. A prediction uncertainty map is created to identify regions of high SMB uncertainty. The model estimates net SMB to be 2345 Gton $\text{yr}^{-1}$, with 95% credible interval (2273,2413) Gton $\text{yr}^{-1}$. Overall, these results suggest lower Antarctic SMB than previously reported. Using the model's uncertainty quantification, we propose 25 new measurement sites for field study utilizing a design to minimize integrated mean squared error.

Changes in atmospheric circulation over the past five decades have enhanced the wind-driven inflow of warm ocean water onto the Antarctic continental shelf, where it melts ice shelves from below1-3. Atmospheric circulation changes have also caused rapid warming4 over the West Antarctic Ice Sheet, and contributed to declining sea-ice cover in the adjacent Amundsen-Bellingshausen seas5. It is unknown whether these changes are part of a longer-term trend. Here, we use waterisotope (Delta O-18) data from an array of ice-core records to place recent West Antarctic climate changes in the context of the past two millennia. We find that the δ18O of West Antarctic precipitation has increased significantly in the past 50 years, in parallel with the trend in temperature, and was probably more elevated during the 1990s than at any other time during the past 200 years. However, Delta O-18 anomalies comparable to those of recent decades occur about 1% of the time over the past 2,000 years. General circulation model simulations suggest that recent trends in Delta O-18 and climate in West Antarctica cannot be distinguished from decadal variability that originates in the tropics. We conclude that the uncertain trajectory of tropical climate variability represents a significant source of uncertainty in projections of West Antarctic climate and ice-sheet change.

Glacierized change in the Himalayas affects river-discharge, hydro-energy and agricultural production, and Glacial Lake Outburst Flood potential, but its quantification and extent of impacts remains highly uncertain. Here we present conservative, comprehensive and quantitative predictions for glacier area and meltwater flux changes in Bhutan, monsoonal Himalayas. In particular, we quantify the uncertainties associated with the glacier area and meltwater flux changes due to uncertainty in climate data, a critical problem for much of High Asia. Based on a suite of gridded climate data and a robust glacier melt model, our results show that glacier area and meltwater change projections can vary by an order of magnitude for different climate datasets. However, the most conservative results indicate that, even if climate were to remain at the present-day mean values, almost 10% of Bhutan s glacierized area would vanish and the meltwater flux would drop by as much as 30%. Under the conservative scenario of an additional 1 C regional warming, glacier retreat is going to continue until about 25% of Bhutan s glacierized area will have disappeared and the annual meltwater flux, after an initial spike, would drop by as much as 65%. Citation