Stratocumulus As a Gravity Wave Observatory

大气中充满了内部重力波，就像你在水面上看到的波浪一样，但在天空的每个高度都会出现。这些无数重叠的内波有许多波长，许多深度，许多传播速度，这取决于它们的许多来源。来源包括风切变、云和风暴、从山上吹过的风，以及喷射气流和其他风的特征所产生的、鲜为人知的大规模震动。在波峰处，空气可能饱和并形成可见的云。你经常可以在天空中看到短波（波长小于几公里）的条纹云图案。较长的波（100公里或以上）在时间上更多地作为周期性的天空覆盖脉冲，但它们可以从卫星图像上看到云的增强或减少的平行条纹或带或弧。这些条纹与运动波相协调，将我们可以在动画中跟踪的单个云特征聚集在一起或分散开来。在低云（层积云）甲板上，这在海洋的凉爽部分特别常见，这些波浪特别明显，而且很容易测量，跟踪，并仅从卫星图像中收集统计数据。这些非常直接的相机测量的统计数据将告诉科学家很多关于通常神秘的波源的本质，波的传播途径，以及波如何调制单个云粒子的本质，以及云的特征或细胞。该项目还将照亮整个云甲板上的波浪效果。例如，一个非常深的云层增厚事件可以导致降水，从而导致长时间的天空晴朗，而即使是中等振幅的波浪也会在阴天或晴朗的天空中使云量达到50%。这些甲板尺度的净效应可能对地球的热量收支很重要，因为云层会反射阳光。随着计算机档案中大量卫星图像的出现，该项目将能够阐明波浪的季节性，甚至可以寻找几十年来波浪或波浪特征的趋势。为了测量这些波，该项目将使用层积云区域的图像对，使用粒子图像测速（PIV）软件跟踪特征或纹理元素，同时计算时间上的亮度差异。PIV测量的水平速度主要是平流（下风运动），但速度的空间梯度可以提供信息。在云的特征收敛和亮度增加的地方，特别是在可以在许多卫星图像帧中跟踪的周期性细长区域，可以推断出上升的波峰，并且可以估计其波列的性质。在某些情况下，可以确定一个清晰的源（如飓风），但在许多情况下，源可能很远或不清楚，假设包括风场中微妙的平流非线性。该项目将寻求自动化和优化波识别算法（PIV，光流），并将技术扩展到红外图像（具有较低的分辨率和动态范围），用于夜间识别，以最大限度地减少虚假识别等。为了吸引专业人士的参与，该项目最初将侧重于充分研究的最新情况，如实地活动。随着算法的改进，它们将被用于获取数十年的卫星图像档案。结果将是波浪度的数据集，共享给社区和项目人员，以描述和理解这些神秘但无处不在的云调制波现象的来源、性质和影响。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The atmosphere is full of internal gravity waves, something like the waves you can see on a water surface but occurring at every altitude in the sky. These myriad overlapping internal waves have many wavelengths, and many depths, and many propagation speeds, depending on what are their many sources. Sources include wind shear, clouds and storms, wind blowing over mountains, and little-understood larger-scale shudders that can emanate from jet streams and other wind features. At the crest of a wave, air may saturate and form a visible cloud. You can often see short waves (less than a few km wavelength) in the sky as striped cloud patterns. Longer waves (100 km or more) are experienced more as periodic pulses of sky cover in time, but they can be seen from satellite images as parallel stripes or belts or arcs of enhanced and reduced cloudiness. These stripes are coordinated with waves of motion, drawing together and spreading apart the individual cloud features which we can track in animations. In low cloud (stratocumulus) decks, which are especially common over the cool parts of the ocean, these waves are especially obvious, and easy to measure, track, and gather statistics about from satellite imagery alone. These statistics of very direct camera measurements will tell scientists a lot about the nature of the often-mysterious wave sources, wave propagation pathways, and the nature of how waves modulate individual cloud particles, and cloud features or cells. The project will also illuminate wave effects on the cloud deck as a whole. For instance, a very deep cloud thickening event can lead to precipitation and thus a long-lasting sky clearing, while even modest amplitude waves drive the cloud cover toward 50% in otherwise overcast or clear skies. These deck-scale net effects may be important to Earth’s heat budget, as clouds reflect sunlight. With the massive quantities of satellite imagery now available in computer archives, this project will be able to elucidate wave seasonality and even look for trends in waviness or wave characteristics across the decades.To measure these waves, the project will work with image pairs over stratocumulus areas, tracking features or texture elements with Particle Image Velocimetry (PIV) software while simultaneously computing brightness differences in time. The horizontal velocity measured by PIV is mostly advective (motion downwind), but spatial gradients in velocity are informative. Where cloud features converge and brightness increases, especially in periodic elongated zones that can be tracked across many satellite image frames, a rising wave crest can be inferred, and the properties of its wave train can be estimated. In some cases, a clear source can be identified (like a hurricane), but in many cases the source may be far away or unclear, hypothesized to include subtle advective nonlinearities in the wind field. The project will seek to automate and optimize wave identification algorithms (PIV, optical flow), and to extend the technique to infrared imagery (with lower resolution and dynamic range) for nighttime identification, to minimize spurious identifications, etc. To engage a community of expertise, the project will focus initially on well-studied recent situations such as field campaigns. As the algorithms are improved, they will be deployed to ingest many decades of satellite imagery archives. The result will be datasets of waviness, to be shared for community as well as project staff efforts to characterize and understand the sources, properties, and impacts of these mysterious but ubiquitous cloud-modulating wave phenomena.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.