Atmospheric Characterization of the Space Environment: Unique Observations from Haleakala

Many space-based applications are impacted by atmospherics. For example, atmospheric optical turbulence distorts the wave front of incoming light. Water and ice based clouds are often major contributors in the ability of ground based systems to observe their targets. These atmospheric disturbances can often times impact critical, time-sensitive missions from space imaging to optical communications, therefore the ability to detect these disturbances are vital.

Recently, a unique atmospheric monitoring station has been developed and deployed to the summit of Haleakala, HI for the purposes of characterizing the atmosphere above it. Many past observations of atmospheric seeing conditions have been conducted at the summit. Indeed, the site of the Inouye solar telescope was partially chosen due to its favorable atmospheric seeing conditions. However, atmospheric transmission characteristics are much less understood. A whole sky, Infrared Cloud Imager (ICI) has been developed and provides on board, calibrated radiances of the sky from horizon to horizon at intervals up to 20 seconds. The dynamic range of this instrument is from approximately 0 Wm-2to nearly 30 Wm-2. These calibrated radiances are post processed and reduced with the aid of a co-located laser ceilometer to provide atmospheric transmission loss at each field of view. A laser ceilometer provides back scatter profiles of the atmosphere from the ground up to 13 km above ground level at 6 second intervals allowing for near continuous monitoring of the atmosphere. This ceilometer is providing unique insight into the very local atmospherics above it and its potential impacts on SSA and other applications. Nearly two years of continuous data have been collected and processed thus far and reveal unique characteristics of the atmosphere above Haleakala Summit. For example, water based clouds, although measured only 30% of the time, have transmission losses of less than 0.5 (3 dB for communication applications) approximately 50% of the time. These observations imply at a minimum, favorable outcomes for many space based applications.

With the assistance of recent computing advances at the Maui High Performance Computing Center (MHPCC), deep learning algorithms are applied to this data in order to learn how to predict when challenging atmospherics are likely to occur. Specifically, a multi-layer perceptron (MLP) model is trained from inputs from the ICI and ceilometer. Results from the training and evaluation of the MLP along with these first of a kind observations will be shared at the conference.