Glacier ice thickness is crucial to quantifying water resources in mountain regions, and is an essential input for ice-flow models. Using a surface velocity inversion method, we combine ice thickness measurements with detailed surface elevation and velocity data, and derive ice thickness and volume estimates for the Monte Tronador glaciers, North Patagonian Andes. We test the dependence of the inversion model on surface slope by resampling glacier slopes using variable smoothing filter sizes of 16–720 m. While total glacier volumes do not differ considerably, ice thickness estimates show higher variability depending on filter size. Smaller (larger) smoothing scales give thinner (thicker) ice and higher (lower) noise in ice thickness distribution. A filter size of 300 m, equivalent to four times the mean ice thickness, produces a noise-free thickness distribution with an accuracy of 35 m. We estimate the volume of the Monte Tronador glaciers at 4.8 ± 2 km3 with a mean ice thickness of 75 m. Comparison of our results with earlier regional and global assessments shows that the quality of glacier inventories is a significant source of discrepancy. We show that including surface slope as an input parameter increases the accuracy of ice thickness distribution estimates.

中文翻译：

---

坡度估计对冰厚反演模型的影响：以北巴塔哥尼亚安第斯山脉蒙特罗纳多冰川为例

冰川冰的厚度对于量化山区水资源至关重要，并且是冰流模型的重要输入。使用表面速度反演方法，我们将冰厚度测量值与详细的表面高程和速度数据相结合，并得出北巴塔哥尼亚安第斯山脉的 Monte Tronador 冰川的冰厚度和体积估计值。我们通过使用 16-720 m 的可变平滑滤波器大小对冰川坡度重新采样来测试反演模型对地表坡度的依赖性。虽然冰川总体积差异不大，但冰厚度估计显示出更大的可变性，具体取决于过滤器的大小。较小（较大）的平滑尺度在冰厚度分布中产生较薄（较厚）的冰和较高（较低）的噪声。过滤器尺寸为 300 m，相当于平均冰厚度的四倍，产生精度为 35 m 的无噪声厚度分布。我们估计蒙特罗纳多冰川的体积为 4.8 ± 2 公里3平均冰厚 75 m。将我们的结果与早期的区域和全球评估结果进行比较表明，冰川清单的质量是差异的重要来源。我们表明，将表面坡度作为输入参数提高了冰厚度分布估计的准确性。