Go to Hurricane Science Home Page

 

Meteorological Applications of
Remote Sensing Channels

Wavelengths that are used in meteorological applications of remote sensing span a wide range of the electromagnetic spectrum. The regions of the spectrum most commonly used are in the visible, infrared and microwave regions of the spectrum.

Fig. 1. The electromagnetic spectrum.
Image courtesy of Astronomy 162, Dept. of Physics & Astronomy, University of Tennessee.

Visible Channels (0.4 - 0.7 µm)

The visible region of the electromagnetic spectrum spans wavelengths from 0.4 µm to 0.7 µm. The only signficant source of visible radiation is the Sun. Remote sensors that utilize visible light are passive sensors, they measure the amount of visible light that is emitted by the sun and subsequently reflected by land, ice, sea and clouds. A disadvantage of visible images is that they can only be obtained during daylight hours.

Fig. 2. The visible region of the electromagnetic spectrum.

Image courtesy of Imagers: Interactive Multimedia Adventures for
Gradeschool Education using Remote Sensing, Goddard Space Flight
Center, NASA.

Visible imagery is useful for distinguishing between sea, land and clouds during daytime. Visible data are most strongly related to the albedo (the percentage reflectivity) of surfaces: highly reflective surfaces are relatively bright in visible images. Water appears uniformly dark, due to its low albedo below 10%. In general, land surfaces appear brighter than the sea, but darker than clouds. The albedo of land surfaces, however, varies widely, depending on the type of surface. Dense clouds, snow and ice have a high albedo and thus are bright features in visible images. Small clouds may not be visible in the image at all. They invoke, however, a brighter appearance of originally dark surfaces.

From visible images we can obtain several cloud characteristics such as size, shape, contrast, texture and brightness. Using physical models, the water content and optical depth of clouds can be obtained from visible images. Although the relationship between brightness and thickness is complex, it is usually accepted that an increase in brightness is associated with increases in total cloud thickness and in the probability of precipitation. Thus we can generalize that clouds with a high albedo have a large optical depth, a high cloud-water (or ice) content, and a small average cloud droplet size. Clouds with a low albedo have a shallow depth, a low cloud -water (or ice) content, and a large average cloud-droplet size.

Cloud height is another important factor for estimating the probability of precipitation. Visible images are not able to determine the height of clouds, therefore visible images must be used together with infrared images in order to estimate precipitation.

Images taken using a single visible wavelength can be used as high resolution black and white photographs of cloud cover, and thus are used extensively to monitor the position and movements of thunderstorms and hurricanes.

Infrared Channels

Fig. 3. Lower portion of the infrared spectrum, showing ranges of
wavelengths used to detect water vapor (H2O absorption) and thermal
infrared emissions from land, water and clouds (atmospheric window).

The infrared region of the spectrum begins with wavelengths just larger than red visible light, and extends up to wavelengths of 1 mm. The infrared region is further divided into three subregions:

near infrared: 0.75 Ð 3.0 µm

middle infrared: 3 Ð 30 µm

far infrared: 30 Ð 1000 µm (1 mm)

For meteorological purposes, the most useful infrared wavelengths are in the middle infrared region of the spectrum.

Atmospheric Window. In the 10-12 µm region, reflected radiation from the sun is negligible and it is thermal radiation emitted by the earth and clouds that produces the images. Radiation at these wavelengths is mostly emitted into space and is not absorbed significantly by atmospheric gases (including water vapor). Thus this portion of the middle infrared is referred to as the thermal atmospheric window. The satellite detector receives the thermal radiation emitted by the earth's surface and by the upper sides of clouds. The thermal images created by these wavelengths reflect the actual temperature differences between sea, land and clouds. This information can be used to distinguish cold cloud tops from warm cloud tops, thus inferring cloud top heights. GOES satellites employ an atmospheric window channel at 10.7 µm.

Fig. 4. Colorized GOES IR image of Hurricane Keith, September 2000.

Red and yellow colors mark the coldest (thus highest) cloud tops. In this image, the cumulonimbus towers around the eye of the hurricane show up as red. The narrow ring of green and blue inside the red region shows portions of the eye wall lower (thus warmer) than the tops of the cumulonimbus towers, and reveals the sloping nature of the eye wall.

In general, high clouds are composed of strongly undercooled water and ice crystals. Cirrus clouds contain exclusively ice crystals, and exhibit temperatures below -35¡C.. The majority of the high clouds are stratiform in texture (cirrostratus) and of large horizontal extent. Cirrus can be very thin and/or semi-transparent, thus cirrus clouds will not appear in the visible image. In this case the infrared image allows a distinction because the ice crystal clouds are very cold in contrast to the warmer background.

In the Meteosat 6 infrared image below, the high clouds (>7 km) are marked blue. In addition to the cirrus clouds, which are extremely well identifiable in the southern latitudes, the cumulonimbus clouds over equatorial Africa show up nicely, too.

Fig. 5. Met6 satellite image showing data from the Atmospheric Window Channel.

Comparisons of infrared temperature maps and radar echo maps show that low temperatures (high cloud tops) are closely correlated to rainfall. This correlation can be considered acceptable for up to a few hours after heavy rainfall, but is not so strong generally. This indicates that factors other than cloud top height need to be taken into account. For example, infrared images are not useful for obtaining the thickness of the cloud. Cloud thickness is also an important parameter for estimating the probability of precipitation. Infrared images must be used together with visible images.

GOES Shortwave Infrared (3.9 µm)

GOES satellites utilize a channel at the very bottom of the middle infrared range (3.9 µm). This channel is referred to as shortwave infrared. Water droplets emit less radiation at this wavelength than at longer infrared wavelengths. This property often permits the identification of fog, and the discrimination of water clouds and ice clouds. In addition, at night, this channel is capable of sensing low level clouds.

Water Vapor Channels

Infrared wavelengths from about 5.0 Ð 7.5 µm are strongly absorbed by water vapor at mid- and upper- levels of the atmosphere. Geostationary meteorological satellites measure upwelling radiation at these wavelengths, called the water vapor channel. GOES satellites employ a water vapor channel at 6.7µm.

Water-vapour images represent approximately the humidity of the middle troposphere. We can see the highest clouds, however, surface features cannot be detected. Instead, eddies and streaks of water vapour are clearly detected where low clouds are not shown. Water vapour images are generally animated to display motions in the atmosphere. Images showing the distribution of water vapor can infer mesoscale regions of moistening or drying, such as may be related to upward and downward movement of air in troughs and ridges. More water vapor is present in bright areas, and is often interpreted to show regions of rising air. Dark areas, with less water vapor are indicative of sinking motions.

Fig. 6. Image from the water vapour channel of the MET6 satellite.

 

Derivation of Upper Level Winds from the Water Vapor Channel

Sequences of water vapor images can be used to track short term movement of water vapor, and are used to estimate upper level wind fields. GOES Channel 3 imagery sequences can be used to derive upper-level wind vectors at three heights, 100-250 mb, 251-350 mb, and 351-500 mb, with the winds plotted on the image valid at the time of the winds. The layer in which cirrus clouds are present is assigned a height of 100-250 mb. Cirrus bands provide excellent targets to be tracked by the wind vector generating algorithm. The wind speeds and directions are derived from the images while the height assignments are done using model forecasts and quality assurance procedures to correct heights which may not fit the analysis.

Fig. 7. Application of Channel 3 water vapor derived winds, Hurricane Floyd on Sept. 15, 1999, 18:00 UTC.
Image provided by David Stettner, Space Science & Engineering Center, Univ. of Wisconsin, Madison.

Microwave Channels (1 mm Ð 30 cm)

Passive Microwave Sensors

For most meteorological purposes, the microwave region is defined as the interval of the electromagnetic spectrum with wavelengths of about 0.1 cm Ð 10 cm. In the microwave region of the spectrum, radiation emitted by the surface of land or water can penetrate water vapor and clouds much better than infrared radiation. In fact, surface temperatures can be retrieved using microwave techniques under complete cloud cover, although with slightly less accuracy, and provided it is not raining.

Precipitation droplets (rain), however, are the right size to interact strongly with microwave radiation and this enables them to be detected by microwave radiometers. Hence, microwave images are a more direct method for determining precipitation than visible or infrared images.

The TRMM Microwave Imager (TMI) is a passive multi-channel radiometer (similar to SSM/I) whose signals in combination can measure rainfall quite accurately over oceans and somewhat less accurately over the land. The TMI measures the microwave radiation emitted by earth's surface and by cloud and rain drops. Because large ice particles (often present in upper cloud regions) tend to scatter this emitted radiation, the TMI uses its various channels along with cloud models to discriminate between these processes and quantify the rain and ice responsible for the observed microwave signatures.

Detection of Wind Speed using Passive Microwave Imagery

With an air- or space-borne radiometer measuring the upwelling microwave radiation over a range of specific frequencies it is possible to retrieve the surface wind speed over the sea surface. Due to the high dielectric constant of seawater, the microwave emissivity of a calm sea surface (acting as a specular surface) is generally rather low - around 0.3-0.5. The graphic below shows an example of the result of a sea surface wind speed retrieval applied to data from the Special Sensor Microwave/Imager (SSM/I).

Fig. 8. Wind speed map for Sept. 24, 2001, as measured by the SSM/I radiometer.
Image courtesy of Marine Observing Systems Team, National Oceanic & Atmospheric Administration. Data processing and distribution performed by NOAA/NESDIS. The image mosaic contains the data from ascending paths of several consecutive satellite passes. Wind speed is given in knots.

Active Microwave Sensors

Precipitation Radar. The PR aboard the TRMM satellite is an active microwave sensor (a radar). It sends microwaves with frequencies of 13.796 and 13.802 GHz, with horizontal polarization, down toward the surface of the earth and measures the echo backscattered from rain. The strength of the echo is roughly proportional to the square of the volume of falling water. This allows the PR to produce very accurate estimates of rain profiles. PR can determine the vertical distribution of precipitation by measuring the "radar reflectivity" of the cloud systems and the weakening of a signal as it passes through the precipitation. As a result, it can measure the 3-D rainfall distribution over both land and ocean.

The disadvantage of microwave techniques for precipitation estimation is the poor spatial and temporal resolution, together with the difficulties in interpreting the images, especially over land.

Detection of Wind Speed using Active Microwave Sensors

A scatterometer is a microwave radar sensor used to measure the reflection or scattering effect produced while scanning the surface of the earth from an aircraft or a satellite. The SeaWinds scatterometer aboard the QuikSCAT satellite is a microwave radar designed specifically to measure ocean near-surface wind speed and direction. Its antenna sends 13.4 GHz microwave pulses to the ocean surface. When the pulses hit the surface of the ocean, the water causes a scattering affect referred to as backscatter. A rough ocean surface returns a stronger signal because the waves reflect more of the radar energy back toward the scatterometer antenna. A smooth ocean surface returns a weaker signal because less of the energy is reflected back to the antenna.

Fig. 9. QuikSCAT data showing the cyclonic surface winds of a
typhoon approaching Japan.

Image courtesy of the Marine Observing Systems Team, National Oceanic & Atmospheric
Administration, Ocean Surface Winds QuikSCAT Data Archive. QuikSCAT is a
JPL/NASA program. Data processing and distribution performed by NOAA/NESDIS.