A Nonlinear Statistical Model for Extracting a Climatic Signal From Glacier Mass Balance Measurements

Abstract

Understanding changes in glacier mass balances is essential for investigating climate changes. However, glacier-wide mass balances determined from geodetic observations do not provide a relevant climatic signal as they depend on the dynamic response of the glaciers. In situ point mass balance measurements provide a direct signal but show a strong spatial variability that is difficult to assess from heterogeneous in situ measurements over several decades. To address this issue, we propose a nonlinear statistical model that takes into account the spatial and temporal changes in point mass balances. To test this model, we selected four glaciers in different climatic regimes (France, Bolivia, India, and Norway) for which detailed point annual mass balance measurements were available over a large elevation range. The model extracted a robust and consistent signal for each glacier. We obtained explained variances of 87.5, 90.2, 91.3, and 75.5% on Argentière, Zongo, Chhota Shigri, and Nigardsbreen glaciers, respectively. The standard deviations of the model residuals are close to measurement uncertainties. The model can also be used to detect measurement errors. Combined with geodetic data, this method can provide a consistent glacier-wide annual mass balance series from a heterogeneous network. This model, available to the whole community, can be used to assess the impact of climate change in different regions of the world from long-term mass balance series. ©2018. The Authors.