High/Low Cloud and Snow Discriminator - Focus Tutorial

Introduction

True Color Image Containing Snow, Low Cloud and High Cloud	High/Low Cloud Product Provides a Better Depiction of These Elements
It is often very difficult to discriminate between clouds and snow-covered land using standard visible and infrared satellite imagery. This is because both clouds and snow are highly reflective at short (solar) wavelengths and appear cold at long (infrared) wavelengths. Enlisting information from the mid-range wavelengths (1 to 4 micron window channels) allows us to discriminate between snow and clouds. Once clouds and snow are decoupled, further discrimination between low (water phase) and high (ice) clouds can be achieved through additional reflectance and emission screening techniques. The result is a false-color product that represents land as green, snow as white, low clouds as yellow, and high clouds as dark orange to magenta.

Background

Combining Multiple Channels to Decouple Clouds and Snow
The slide above gives a basic overview of the channels used for the high/low cloud over snow product. While MODIS is emphasized here, similar enhancement can be produced with the AVHRR instrument. The graphic illustrates the result of an intelligent combination of the channel information listed in the bullets above, with relative contributions to the Red, Green, and Blue color guns from land, snow, low cloud, and high cloud represented as vertical bars. A full vertical bar corresponds to a large contribution, while a small bar corresponds to a minimal contribution. For example, snow displays a large contribution in Red, Green and Blue, and therefore is enhanced as white (see tutorial on MODIS Dust Enhancement for background information on interpreting false color enhancements). Low clouds, on the other hand, produce minimal contributions in the Blue color gun and hence appear as yellow.

Limits

Large Ice Crystals Can Appear as
The current enhancement is not without its pitfalls and limitations. There are many instances where the optical properties that constitute the scene translate to less-than-optimal coloration in the enhancement. Some mixed-phase clouds or supercooled mid-level clouds may appear as shades approaching the appearance of snow. In other cases, large ice aggregates lofted by deep convection (e.g., example above) can mimic the radiative signatures used to detect snow and therefore be falsely enhanced as such. A good rule of thumb is to place the pixels in question in the context of the larger image. In most situations the analyst can readily identify those clouds exhibiting false snow signatures.

Examples

Detection of Overlapping Cloud Layers
The transparent nature of some varieties of cirrus allow for viewing of a cloud layer (or surface) beneath it. In the example above, a shield of pre-frontal cirrus fans across a snowy landscape. Patches of low cloud or fog are clearly evident atop the snow backgrounds as illustrated, even though they reside beneath the upper deck of cirrus. This can prove useful, for example, for advising sorties flying between cloud decks on target acquisition in a broken low cloud field.

Varying Appearance of Cirrus
When cirrus is thin enough to allow for emissions from the background below the cloud (lower cloud layer or the surface) to contribute to the satellite observation, its appearance in the false color enhancement is altered according to the properties of this background. In the above example, thicker areas of cirrus (magenta) give way to thinner cirrus on the edges where a higher green contribution from the land background alters the red/green/blue composite color. It is not uncommon to observed a thin shield of cirrus transition from purple to orange as it crosses from a water (dark) to land (bright) background. An eye for these nuances is best developed by using the product on a regular basis and gaining familiarity with performance under a variety of conditions.

Sun Glint Over Water Can Lead to False Low Cloud Signature
The region of specular (mirror) reflection of the solar disk upon the relatively smooth surfaces of water bodies can give rise to false yellow tonality in this product. Most of the time it is relatively easy to pick these regions out, but sometimes local winds, currents, or suspensions in the water can give rise to brightness variability in the glint zone. It is recommended that the analyst always cross reference with the true color products to ascertain areas of possible glint contamination. This is not an issue over the land bodies.