Rapid Quantitative Snow Stratigraphy for Avalanche Forecasting Using Near-Infrared Photography

为了研究积雪加强和减少雪崩危险的方式，我们观察了积雪深度剖面的雪层和烧结。 在过去的几年中，我们使用光学显微镜和扫描电子显微镜来测量雪粒的几何形状和化学成分。 由于这类观测所需的时间，我们只在两个地点取样。 我们计划扩大我们的采样，以了解如何地层和烧结空间变化使用一种新的方法，近红外数字摄影。 这将允许快速定量地层学和雪的稳定性和雪变质率的评价。 从历史上看，雪崩预报集中在积雪中可识别的层，但我们的近红外图像在雪崩的一天内跨越风暴雪崩冠面（最远的上坡拉伸断裂）显示床表面和板之间的光学特性几乎没有明显的差异。 在北美西部的海洋环境中，超过90%的破坏发生在风暴层内，并在非持久性弱层上发生破坏。 我们的假设是雪崩不需要明显的薄弱层。相反，我们建议，新雪的稳定性应被视为一个连续的问题，在任何时候，薄弱点位于应力/强度比最低的地方。覆盖层引起下坡应力，而烧结和压实增加强度。 为了验证这一假设，我们将使用近红外摄影结合机械强度的测量在冠面进行广泛的测量，我们将开发一个连续模型的雪的稳定性，包括下坡应力和连续烧结在风暴层的所有深度。 评估暴风雪期间雪的稳定性是困难的，但在某些环境中，几乎所有的雪崩都发生在暴风雪期间或之后不久。 近红外摄影在雪崩危险性评估中的新应用使雪地层学的快速研究成为可能，其结果可能是数据与雪的机械特性及其在雪崩危险性评估中的应用相关。 虽然在雪表面的粒度的异质性进行了研究与遥感，有很多较少的研究，特别是在斜坡尺度上的雪埋层的属性的异质性。 我们的经验，广泛的仪器，现场可达性，以及大量的雪崩控制记录，使我们的现场位置，猛犸山，一个理想的网站，以测试如何近红外摄影可以用来检查在哪里和何时在雪剖面积雪失败。 我们将开发并提供开放源码软件，以处理近红外雪坑图像并绘制剖面地层图。 我们的软件，结合近红外数字摄影的低成本和速度，可以使高分辨率定量地层学提供给研究界以外的广泛用户群，包括导游，滑雪巡逻，雪崩预报员，并可能娱乐用户。 我们对风暴期间的失败的分析将揭示这一重要的雪崩集，不同于那些在深，弱的老层失败。最终，我们的研究结果和技术可能会减少北美雪崩死亡人数的增加。

To examine the ways a snowpack strengthens and reduces the avalanche hazard, we observe snow stratigraphy and sintering throughout the depth profile of the snowpack. In previous years, we used optical microscopy and scanning electron microscopy to measure snow grain geometry and chemical composition. Because of the time required for those kinds of observations, we sampled only two locations. We plan to extend our sampling to learn how stratigraphy and sintering vary spatially using a new method, near-infrared digital photography. This will allow for rapid quantitative stratigraphy and evaluation of snow stability and rates of snow metamorphism. Historically, snow avalanche forecasting has focused on identifiable layers in the snowpack, but our near-infrared images across in-storm avalanche crown faces (the furthest uphill tensile fracture) within one day of avalanching show little discernible difference in the optical properties between the bed surface and the slab. In the maritime environment of western North America, more than 90% of failures occur within the storm layer and fail on non persistent weak layers. Our hypothesis is that no obvious weak layer is needed for avalanches. Instead, we propose that stability of new snow should be treated as a continuum problem where, at any time, the weakness is located where the ratio of stress/strength is lowest. The overburden causes downslope stress, while sintering and compaction increase strength. To test this hypothesis, we will make extensive measurements at crown faces using near-infrared photography combined with measurements of mechanical strength, and we will develop a continuum model of snow stability that includes downslope stress and continuous sintering at all depths in the storm layer. Evaluating snow stability during storms is difficult, yet in some environments almost all avalanches occur during or shortly after storms. The novel application of near-infrared photography to avalanche hazard evaluation enables the rapid study of snow stratigraphy, with the likely results that the data are correlated with mechanical properties of snow and their application to avalanche hazard evaluation. While the heterogeneity of grain sizes at the snow surface has been investigated with remote sensing, there are many fewer studies of the heterogeneity of snow properties in buried layers, especially at the slope scale. Our experience, extensive instrumentation, site accessibility, and large number of avalanche control records make our field location, Mammoth Mountain, an ideal site to test how near-infrared photography can be used to examine where and when in the snow profile the snowpack fails. We will develop and make available open source software to process near-infrared snowpit images and to map stratigraphy of the profile. Our software, combined with the low cost and speed of near-infrared digital photography, could make high resolution quantitative stratigraphy available to a broad user group beyond the research community, including guides, ski patrollers, avalanche forecasters, and possibly recreational users. Our analysis of failure during storms will shed light on this important set of avalanches, distinct from those that fail on deep, weak older layers. Ultimately, our results and techniques may reduce the increasing number of avalanche fatalities in North America.