This paper explores the snow-avalanche regime based on tree-ring reconstructions and their triggering weather conditions with classification tree algorithms.

The results show a significant increased frequency of avalanche events on Zăvoaie NE slope for the second half of the 20th century by comparison to the Scărița SW slope. The classification tree models highlight the weather conditions leading to avalanche release with three scenarios in each path. The first scenario underlines the wind's effect as a key weather variable on both slope aspects. The second scenario corresponds to a spring regime, while rain and warm temperatures are the main triggers. The third general weather condition favouring snow avalanche activity are persistent low temperatures and important snowfall throughout the winter season. However, this triggering condition was mainly found on the NE avalanche path, probably related to the pattern of snow accumulation, the prevailing winds, but above all the lower solar radiation which favours a slower and later melting of the snow cover. Finally, the return periods and runout distances calculated from tree ring analysis show a high risk for the location of the infrastructure planned for the ski area expansion, showing once again the usefulness of dendrogeomorphology in natural hazard assessment where historical data are lacking.

本文探讨了基于树木年轮重建的雪崩机制及其触发天气条件的分类树算法。

结果显示，与 Scărişa SW 斜坡相比，20 世纪下半叶 Zăvoaie NE 斜坡上的雪崩事件频率显着增加。分类树模型突出了导致雪崩释放的天气条件，每条路径有三种情景。第一种情况强调了风作为两个斜坡方面的关键天气变量的影响。第二种情况对应于春季，而降雨和温暖的气温是主要触发因素。第三个有利于雪崩活动的一般天气条件是持续的低温和整个冬季的重要降雪。然而，这种触发条件主要出现在 NE 雪崩路径上，可能与积雪模式、盛行风、但最重要的是，较低的太阳辐射有利于积雪较慢和较晚的融化。最后，从树木年轮分析计算的返回周期和跳动距离表明，规划用于滑雪场扩建的基础设施位置存在很高的风险，再次表明树木地貌学在缺乏历史数据的自然灾害评估中的有用性。