Despite the importance of forests in the water and carbon cycles, accurately measuring their contribution remains challenging, especially at night. During clear-sky nights current models and theories fail, as non-turbulent flows and spatial heterogeneity become more important. One of the standing issues is the ‘decoupling’ of the air masses in and above the canopy, where little turbulent exchange takes place, thus preventing proper measurement of atmospheric fluxes. Temperature inversions, where lower air is colder and thus more dense, can be both the cause and result of this decoupling.

With Distributed Temperature Sensing (DTS) it is now possible to detect these temperature inversions, and increase our understanding of the decoupling mechanism. With DTS we detected strong inversions within the canopy of a tall Douglas Fir stand. The inversions formed in on clear-sky nights with low turbulence, and preferentially formed in the open understory. A second inversion regularly occurred above the canopy. Oscillations in this upper inversion transferred vertically through the canopy and induced oscillations in the lower inversion.

We hypothesize that the inversions could form due to a local suppression of turbulent motions along the height of the canopy. This was supported by a 1-D conceptual model, which showed that a local inversion layer would always form within the canopy if the bulk inversion (over the full canopy) was strong enough.

Due to the near-continuous vertical motion and specific height the inversions occur at, a very high measurement density (better than ∼2 m) and measurement frequency (>0.1 Hz) are required to detect them. Consequently, it could be possible that the observed inversions are a regular feature in similarly structured forests, but are generally not directly observed. With DTS it is possible to detect and describe these types of features, which will aid in improving our understanding of atmospheric flows over complex terrain such as forests.

尽管森林在水和碳循环中很重要，但准确测量它们的贡献仍然具有挑战性，尤其是在夜间。在晴朗的夜晚，当前的模型和理论失败了，因为非湍流和空间异质性变得更加重要。长期存在的问题之一是树冠内部和之上的气团“脱钩”，在那里几乎没有发生湍流交换，从而妨碍了对大气通量的正确测量。温度反转，较低的空气更冷，因此更密集，可能是这种脱钩的原因和结果。

借助分布式温度传感 (DTS)，现在可以检测这些温度反转，并增加我们对解耦机制的理解。借助 DTS，我们在一个高大的道格拉斯冷杉林分的树冠中检测到强烈的倒置。逆温形成于晴空低湍流的夜晚，并优先形成于开阔的下层。在树冠上方经常发生第二次倒置。上部逆温中的振荡垂直通过冠层传递，并在下部逆温中引起振荡。

我们假设由于沿树冠高度的湍流运动的局部抑制，可能会形成倒置。一维概念模型支持这一点，该模型表明，如果整体逆温（在整​​个冠层上）足够强，则局部逆温层总是会在冠层内形成。

由于近乎连续的垂直运动和反转发生的特定高度，需要非常高的测量密度（优于~2 m）和测量频率（> 0.1 Hz）来检测它们。因此，观察到的反转可能是类似结构森林中的常规特征，但通常不会直接观察到。借助 DTS，可以检测和描述这些类型的特征，这将有助于提高我们对森林等复杂地形上的大气流动的理解。