A. Stratus/fog detection
1. Nighttime (IR Channels 2 and 4)
Prior to the GOES I-M satellites, only a single IR channel was available for the detection of fog or low clouds at night - the longwave IR window at 11 microns. The addition of a shortwave channel at 3.9 microns (Channel 2 on GOES I-M) has greatly improved our ability to monitor low ceilings and visibilities. Its lower emissivity in the presence of low level water clouds results in a lower brightness temperature than the longwave IR, while cloud-free areas show little temperature difference.
The maximum information is obtained from a difference between the two IR channels. The difference image shown over New Mexico in the left panel shows the valley fog as white, cloud-free areas gray, and cirrus black. That image is a great improvement over the enhanced Channel 4 IR image in the right panel.
At times, cloud areas which appear to be fog may have elevated bases , such as stratocumulus cloud.
Recent research has shown that the magnitude of the temperature difference between Channels 2 and 4 is related to cloud thickness. This allows us to estimate fog/stratus depth by applying an appropriate enhancement. In the New Mexico example, the thicker fog and stratus near Albuquerque did not burn off until late afternoon. This technique is only an approximation, accurate to 50-100m, and may give ambiguous results when higher cloud layers are present.
2. Daytime (Channels 1, 2 and 4)
During daylight hours, fog and low clouds have distinctive characteristics in visible (Channel 1) imagery, such as smooth tops and distinct edges. In areas of snow or ice coverage, however, detection can become difficult with visible data. Channel 2 (3.9 microns) can be helpful in those situations because of the higher reflectance of water droplet clouds compared to ice and snow. It will also provide better discrimination than the longwave window IR.
B. Aircraft Icing
GOES satellite imagery can assist in monitoring areas where aircraft icing is likely. Even though satellites observe only the tops of clouds, recent research has shown that large amounts of supercooled liquid water can accumulate near cloud tops. Many of the meteorological conditions conducive to icing can be inferred or measured via satellite. For example:
1. Channel 1 (Visible) - Shows area coverage. Brightness relates to cloud depth, water content, and high RH. The visible image for 13 April 1995 shows extensive stratiform clouds over the Ohio Valley and Northeastern U.S. Texture shows cloud phase and the presence of convection. In this example, smooth cloud tops indicates an absence of convection.
2. Channel 2 (3.9 micron IR) - Brightness temperature during daylight hours includes both an emitted thermal and reflected solar component. The reflectance is higher for liquid phase clouds, so those types of clouds are significantly warmer than Channel 4 (10.7 microns). A sample collected for actual icing reports shows that this difference can be as much as 30 degrees C when the solar zenith angle is large.
Information on cloud drop size is also present, because smaller drops are more reflective , thus appear warmer than clouds with large drops. The latter often contain drizzle sized drops that are known to cause severe icing. The CH2 image on 13 April 95 shows cloud top temperatures higher than 0C everywhere, but considerable variation horizontally. The heavier icing tends to occur in lighter gray (cooler) areas in this channel.
3. Channel 4 (10.7 micron IR) - This data shows cloud top temperatures, as well as horizontal coverage. Most icing occurs with temperatures between 0 and -20C. The top of the icing layer can be determined with temperature profiles from radiosondes or numerical models. In the CH4 image on 13 April 95, cloud top temperatures were in a favorable range (0 to -15C), except where cirrus clouds were present. This translated to heights of 7 to 12 kft above Mean Sea Level (MSL). We don't see as much variation in temperature as in CH2, suggesting that cloud tops are fairly uniform, and that CH2 is providing information on cloud drop sizes.
4. Channel 5 (12 microns) - When this channel is combined with CH4, we can objectively identify thin cirrus clouds that show temperatures in the correct range for icing in CH4. This will be useful in producing a derived product from GOES imagery, especially during the daytime.
C. Jet Streams/ Clear-Air Turbulence (CAT)
1. Jet Streams
Animated satellite imagery, especially the water vapor data (CH3) is very useful in the identification of jet streams and other significant upper tropospheric features important for air route planning. Derived cloud motion wind vectors can verify the location of jet stream axes. Many of these occur along the poleward edge of persistent cirrus shields or moisture seen in the IR or WV imagery.
2. Clear Air Turbulence
Although the jet stream is predominately smooth, clear air turbulence (CAT) can occur in some situations. There are often satellite signatures that suggest the presence of CAT, although the altitude cannot be determined accurately.
a. Transverse Cirrus Bands
One of the indications is the presence of well-defined cirrus bands oriented nearly perpendicular to the upper winds. The most intense turbulence is related to wide, thick bands easily seen in IR imagery or visible. The bands occur in the presence of strong vertical and horizontal wind shears below the jet. Winds often show pronounced backing with height.
During the warm seasons, the jet stream is often enhanced by mesoscale convective systems (MCS's). The result is stronger-than-forecast winds aloft, especially in the left front quadrant of an MCS (looking downstream). In the example for 16 July 96, 200-mb winds of 100-110 kt were observed by the upper air stations from Green Bay WI to Detroit MI, at least 30 kt more than the model forecasts. Moderate CAT was observed through the day along the jet, and in the right front quadrant.
b. Water Vapor Imagery
Strong, widespread CAT can usually be found in positively-tilted (NE-SW) upper troughs where the polar and low latitude jet streams converge. In these situations, we frequently see progressive darkening in the WV imagery, such as that which occurred on 11 January 1996. The darkening is caused by strong subsidence within a tropopause fold, induced by upper level convergence. A cross section through the leading edge of the dry zone shows the structure across the trough axis and jet stream on the morning of 11 January. The isentropes slope downward as they approach the north Texas area, where the dry slot is present. The synoptic scale flow patterns that most often result in WV darkening have been identified.
D. Mountain Wave Turbulence
1. Low Level Wave Clouds
Low level (trapped) mountain waves occur when strong low level winds blow across mountain ranges in the presence of a strong inversion near the mountain tops. Weak directional shear is another requirement. When moisture is present, these waves become visible in satellite imagery. The improved resolution of GOES I-M satellites permits much better detection of these waves at night. The spacing we see between cloud bands in the images is directly related to mean layer wind speed. We can thus expect longer wavelengths to be associated with stronger low level turbulence. This has been verified by comparisons of cloud band spacing with actual aircraft turbulence intensities.
2. High Altitude Waves
Along the east slopes of some large mountain ranges, such as the Rockies in the western United States, strong mountain waves at the lower levels can propagate their energy vertically and produce steep mountain waves that can cause severe or extreme turbulence at high altitudes. The conditions that produce these mountain waves are similar to those that cause strong Chinook wind storms along the lee of the mountains. The principle satellite signature is a high, cold cirrus plume with a dark (warm) subsidence zone that appears in water vapor images near or just to the lee of the mountain ridges. For non-turbulent waves, the upstream edge of the cirrus plume is aligned near the ridgeline.
On 24 April 1996, moderate to severe turbulence probably related to a mountain wave was reported near Sheridan, Wyoming at 43-45,000ft above Mean Sea Level. The IR, water vapor and visible imagery at the same time showed a pronounced sinking zone (Foehn gap) east of the Big Horn Mountains. The wind profiler at Medicine Bow, Wyoming, located farther south, showed that winds near mountain top level were 30-40 m/sec. High altitude severe turbulence was reported near Denver later in the afternoon.
Strong straight-line surface winds produced by convective storms sometimes are accompanied by signatures in satellite and radar data. Strong gradients can often be found in the IR cloud top temperatures and WSR-88D reflectivity at the leading edge of storms containing strong surface winds. These regions can often be observed with bow or comma-shaped patterns in the thunderstorm systems.
F. Volcanic Ash Detection
The "split window" channel (CH5 IR - 12.0 microns) on the GOES-8/9 satellites can be used in conjunction with the other longwave IR data to effectively detect high altitude volcanic ash clouds better than single channel IR or visible data. In early March, 1996, Mt Popocateptl near Mexico City emitted large volumes of volcanic ash up to the stratosphere. The split window (CH5-4) image on the afternoon of 12 March showed the ash plume more clearly than other image channels.
II. MARINE APPLICATIONS
Sea fog forms when moist air flows over open waters whose temperature is close to the dew point of the overlying air. Detection of sea fog is made difficult by its shallow nature. The nighttime, bispectral technique can be successfully used, but some fog areas are not as easily seen since the emissivity differences between the two IR channels are not as great for thinner cloud layers.
Extensive sea fog occurred over the northeastern Gulf of Mexico and over the Atlantic shelf waters on the morning of 26 February 1996. The dual IR technique depicted the location of the low clouds offshore much better than single IR images.
Sea surface temperature (SST) distribution showed that the fog occurred where SST is in the range 15-18 C.
B. Oceanic Convection
1. Cellular Patterns/Cloud Streets
The most common type of oceanic convection is the cellular pattern of cumulus/cumulonimbus that forms open rings, arcs, or circular patches. These cloud patterns are similar to classic Benard cells observed in the laboratory.
a. "Open" Cells
Open cells form when there is strong instability below the marine inversion, as in cold advection situations to the rear of an oceanic cold front. In weaker flow situations, they form circular or elliptical rings, while in strong flow, open arc lines are most likely. Open cells usually indicate surface winds of greater than 25 knots. In areas of open cell convection, the diameter of the cells is approximately 15 times the cloud top heights.
b. "Closed" Cells
Closed cells form in conditions of subsidence and anticyclonic low level flow. They usually indicate surface winds less than 25 knots. Sometimes brighter, thicker strands are observed within the closed cells. These are often associated with light precipitation and reduced visibilities.
c. Cloud "Streets"
Cloud streets are narrow plumes of cumuliform clouds that develop in strong, cold flow across warmer waters. Cloud top heights are low at first, then increase farther offshore as the marine inversion is lifted by sensible heat transfer into the boundary layer. They are analogous to lake effect snow bands, and can produce locally moderate rain or snow with reduced visibility. Since they form in conditions of little change in wind direction with height, the orientation of the streets is a good indicator of surface wind direction. The spacing between cloud streets is approximately 5 times the cloud top height.
2. Initiation of deep convection
Deep convection over oceanic regions is forced by many of the same conditions as convection over land: low level moisture, instability, dynamic forcing such as an upper trough, cold front, or surface convergence zone, etc. To provide unstable conditions, the marine inversion must be weak or absent. Thus, we rarely see thunderstorms in regions of cool sea surface temperatures or within surface anticyclones unless there is extremely cold air aloft.
Deep convection often develops along boundaries of warm ocean currents such as the Gulf Stream, and the loop current in the Gulf of Mexico. The sharp gradients in SST are reflected in the temperatures through the lowest several hundred meters, as shown by research flights. Thus, low level thermal boundaries are usually present to serve as a triggering mechanism for convection. The strength of the convection then depends on the synoptic conditions.
Diurnal variation of convection over large water bodies is different than over land. Numerous studies using satellite data have shown peaks in rainfall and convection to be late at night to just before sunrise over oceans.
An example of a major thunderstorm outbreak occurred on 1 April 96 over the Gulf Stream waters. At 0815Z, strong new cells were developing along the "West Wall," while convection moving off the North Carolina coast was weakening. By 1215Z, the thunderstorm activity was at a peak, as the upper level system with its pronounced dry slot approached. Model data showed the Lifted Index to be -2 to -3 C in this region.
Even in stable condtions, lines of towering cumulus and showers can occur with sufficient low level convergence.
C. Strong Wind Zones Near Intense Cyclones and Cold Fronts
1. Main Satellite Indicators
Cloud streets (closely spaced) to the rear of surface cyclone or cold front Presence of a "ring cloud" (a nearly circular, eye-like ring of convection in the low center) suggests a warm core structure with winds decreasing with height, such as shown by aircraft data from the ERICA study Darkening within slot of deepening extra-tropical cyclone (shows transport of strong winds downward to near surface) A hooking, sharply tapered, comma cloud head (observed coincident with several ship/buoy/offshore drilling rig upsets)
On 25 March 1996, a strong ocean cyclone in the Pacific showed many of these characteristics. The strongest winds (>50kt) occurred just west and southwest of the center in the region of open cell cu/cb and cloud streets. A ring cloud of low (warm top) convection developed between 0330 and 0630 UTC as the upper center closed off.
SatMet Website INDEX