Earth Observations from Space: The First 50 Years of Scientific Achievements (2008)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

3 Weather Because many aspects of daily life are affected by the S. Grant signed a joint resolution of Congress authorizing weather, understanding and predicting the weather has been the Secretary of War to establish a “Division of Telegrams a human quest for millennia. Space-based observations have and Reports for the Benefit of Commerce” as part of the played pivotal roles in the history of weather forecasting. U.S. Army Signal Service Corp (NWS 2006). This divi- Their contributions to forecasting at all spatial scales can be sion became the Weather Bureau in 1890 and the National grouped into three areas, which are described in this chapter: Weather Service in 1967. weather imagery, atmospheric properties, and numerical The development of radiosondes (weather balloons that weather forecasting. transmit information to a fixed location) in the 1930s yielded Fundamentally, weather forecasting is a four-dimensional measurements that resulted in major advances in weather problem, involving three spatial dimensions and time. The forecasting by adding the crucial vertical dimension to air that is now affecting point B was yesterday at point A meteorological observations. About half of the stations in the and tomorrow will be at point C. Similarly, the storm system Integrated Global Radiosonde Archive depicted in Figure 3.1 centered at point A yesterday is centered at point B today and make observations twice per day. Though numerous, these has a different shape, size, and intensity than it did yesterday. stations cannot give the desired global picture of the current To forecast tomorrow’s weather at a point, one must know state of the weather. The ocean is especially data sparse. today’s weather over a broad region surrounding that point. After World War II, rockets were sufficiently advanced The farther one wants to forecast into the future, the larger to be able to lift cameras high above the clouds to take pho- the area must be where one knows the weather today. tographs of weather systems and to indicate the potential In 1846 the state of the art in weather forecasting was for weather observations by Earth-orbiting satellites. This succinctly stated by François Arago: “Whatever may be the led Harry Wexler to publish a paper in 1954—3 years progress of sciences, NEVER will observers who are trust- before the launch of the first satellite and 6 years before worthy, and careful of their reputation, venture to foretell the Television Infrared Observation Satellite (TIROS) (see the state of the weather.” In the United States, however, Chapter 2)—titled “Observing the Weather from a Satellite isolated observers were communicating among themselves Vehicle” (Wexler 1954). Thus, weather observations from to understand the horizontal extent of the weather, but they space were not serendipitous but were eagerly anticipated. communicated by mail, which meant that the weather could Since the beginning of the space age, perhaps 200 not be forecasted, only understood in retrospect. With the weather satellites have been launched, as nations around the development of the telegraph, Joseph Henry and James world recognized their value and as technology advanced to Espy experimented with transmitting weather observations make more capable instruments possible. in real time to a central site, where weather maps could be drawn and forecasts made. By 1870, President Ulysses WEATHER IMAGERY  1786-1853, director of the Paris Observatory and permanent secretary The first weather satellites attempted simply to “take of the French Academy of Sciences. pictures” from space. By the mid-1960s, engineers had  For a history of these developments, see Fleming (1990).  1797-1878, first secretary of the Smithsonian Institution and one of the founding members of the National Academy of Sciences.  1911-1962, chief of the Scientific Services Division of the Weather  1785-1860, America’s first national meteorologist. Bureau. 18 

WEATHER 19 FIGURE 3.1  Locations of Integrated Global Radiosonde Archive stations. SOURCE: Data from the National Climatic Data Center, Durre et al. (2006). Reprinted with permission from the American Meteorological Society, copyright 2006. developed the capability to fly satellites in sun-synchronous orbits, in which an instrument on a single satellite could view the entire Earth twice per day, once in daylight and once at night (Figure 3.2). Then meteorologists could tile the pictures together to form the long-sought global picture of Earth’s weather (Figure 3.3). Also during the mid-1960s, the first geostationary satellites were launched. These satellites orbit Earth in the equatorial plane at the same angular velocity that Earth rotates on its axis; thus, they stay “stationary” over the same point on the equator. Although they do not view the entire Earth but only one hemisphere (Figure 3.4), they can make images frequently, not just twice per day. These images can be assembled into movies that allow forecasters to watch the weather in motion. This is an invaluable tool for weather analysis and forecasting (Box 3.1). Geostationary satellites rapidly became the choice of weather services worldwide, such that today they form a ring around the equator, provid- ing coverage of the entire tropics and midlatitudes. Many accomplishments in weather forecasting have been achieved using the imagery from weather satellites. Only a few can be mentioned in this document. Perhaps the most dramatic accomplishments relate to observing and pre- FIGURE 3.2  One day’s orbits of a sun-synchronous satellite. A dicting hurricanes and tropical storms. In 1900 a “surprise” single instrument views the entire Earth. SOURCE: Kidder and hurricane roared out of the Gulf of Mexico over Galveston Vonder Haar (1995). Copyright Elsevier, 1995. Island killing at least 8,000 people; it was the largest natural disaster in the United States (Blake et al. 2006). Since then an important scientific accomplishment occurred sometime in the 1960s: with the continuous monitoring of weather by sat- ellites, no tropical cyclone anywhere on Earth escapes detec- tion (Figure 3.5). Indeed, Robert C. Sheets, former director  A circular, near-polar orbit in which the orbital plane keeps a constant relationship to the Sun-Earth line, such that the satellite passes over a point of the National Hurricane Center (NHC), has written: on the Earth near the same time every day. 

20 EARTH OBSERVATIONS FROM SPACE: THE FIRST 50 YEARS OF SCIENTIFIC ACHIEVEMENTS FIGURE 3.3  First complete view of the world’s weather, photographed by TIROS 9, February 13, 1965. Image assembled from 450 indi- vidual photographs. SOURCE: Publication of the National Oceanic and Atmospheric Administration (NOAA), NOAA Central Library. FIGURE 3.4  Example of the hemispheric coverage of a geostationary satellite. Taken by NASA’s Applications Technology Sat- ellite 3 (ATS 3) at 1402 UTC on July 21, 1970. Note that Tropical Storm Becky can be seen in the Gulf of Mexico near Florida. SOURCE: NOAA Photo Library. 

WEATHER 21 descends and spreads out at the ground, producing an outflow boundary. When this boundary interacts with an adjacent BOX 3.1 storm, the intensity and destructive potential of both storms is Weather “Movies” amplified (Purdom 1976, 1986). In addition, satellite images revealed that a cold, V-shaped structure in the anvil of a thun- One of the best ways to understand the impact derstorm is a signature of severe weather (Fujita 1978). It was that observations from space have had on weather also discovered that atmospheric turbulence is signaled by forecasting is to watch the weather in motion in a cloud patterns in the lee of mountain ranges (Ellrod 1989). sequence of satellite images. Unfortunately, this Data from weather satellites allowed scientists to identify and is impossible in a printed document. Readers are forecast many other atmospheric phenomena too numerous urged to visit one of the many Internet sites that to mention here (Kidder and Vonder Haar 1995). offer satellite movies and try a bit of “nowcasting”: Where is that storm going, and when will it arrive at the reader’s location? Here are a few sites: ATMOSPHERIC PROPERTIES In addition to taking pictures from space, satellites have • NOAA’s Geostationary Operational Environ- allowed scientists to make radiometric measurements of the mental Satellites (GOES) site: http://www.goes.noaa. electromagnetic spectrum, from the ultraviolet to the micro- gov (click on the MPEG or Java Applet icon) wave regions. From these measurements, scientists are able • Japan Meteorological Agency’s Multi-functional to retrieve properties of the atmosphere that are important Transport Satellite (MTSAT) site: http://www.jma. to forecasters, especially the vertical temperature structure, go.jp/en/gms/ winds, and moisture content (see Chapters 4 and 5), which • EUMETSAT’s Meteosat site: http://www. are essential for numerical weather prediction. eumetsat.int (under “Image Gallery” choose “Real- Temperature profiles can be retrieved by several means. time Images”) Given measurements at several wavelengths near an absorp- tion band of a well-mixed gas, such as the 15-µm band of carbon dioxide or the 5-mm band of oxygen,, the radiative transfer equation can be used to retrieve a temperature versus height profile that is consistent with the measured radiances (Chahine 1968, Smith 1970; see also Box 5.2). This has The greatest single advancement in observing tools for tropical meteorology was unquestionably the advent of been done with a large number of satellite instruments start- the geosynchronous satellite. If there was a choice of only ing with the Satellite Infrared Spectrometer (SIRS) and the one observing tool for use in meeting the responsibilities Infrared Interferometer Spectrometer (IRIS), both launched of the NHC, the author would clearly choose the geosyn- on the Nimbus 3 satellite on April 14, 1969. Recently, the chronous satellite. (Sheets 1990) Atmospheric Infrared Sounder (AIRS) on the Aqua satellite (launched May 4, 2002) has provided high vertical resolu- From these observations we learned that tropical tion soundings using an infrared instrument that measures cyclones go through a life cycle that can be recognized and atmospheric radiation at 2,378 wavelengths (Chahine et al. categorized in satellite imagery. The Dvorak scheme (Dvorak 2006). 1975) for estimating the intensity of tropical cyclones, which Radio occultation provides another way to measure ranks storms between 1 and 8 based on wind speed and other atmospheric temperature profiles. Radio signals from Global features, is used worldwide. Tropical cyclone studies have Positioning System (GPS) satellites are refracted (bent) as benefited tremendously from satellite data (e.g., Special they pass through the atmosphere. This bending angle can be Sensor Microwave/Images, Advanced Microwave Sound- measured from a second satellite. A sequence of refraction ing Unit, and QuickScat), which have been used to develop angles are measured as the GPS satellite rises or sets through algorithms for monitoring and predicting hurricane intensity, the atmosphere. These measurements can be converted into tracks, and wind structures (Kidder et al. 1980, Demuth et a vertical profile of index of refraction of the atmosphere al. 2000). and thus into a vertical temperature sounding with high Since the advent of satellite imagery, scientists have vertical resolution (Ware et al. 1996). The first such instru- learned to remotely identify many previously known weather ment, GPS/MET, was launched on the MicroLab 1 satellite features: fronts, high- and low-pressure systems, fog, low on April 3, 1995; a constellation of six satellites (Formosa clouds, cirrus, and thunderstorms (Bader et al. 1995). Satel- S ­ atellite [FORMOSAT-3]/Constellation Observing System lite observations also led to the discovery of thunderstorm for Meteorology, Ionosphere, and Climate [COSMIC]) was clusters, called mesoscale convective complexes, which are launched on April 14, 2007. unrelated to classical storm systems (Maddox 1980). Sci- A second property of the atmosphere that is necessary entists also learned that rain-cooled air from thunderstorms for weather forecasting is its water vapor content. It can be 

22 EARTH OBSERVATIONS FROM SPACE: THE FIRST 50 YEARS OF SCIENTIFIC ACHIEVEMENTS FIGURE 3.5  At 1745 UTC on August 28, 2005, Hurricane Katrina was observed by the Geostationary Operational Environmental Satellite (GOES 12) near the time of its maximum wind speed, 150 knots (173 miles per hour). SOURCE: National Hurricane Center. Reprinted with permission from the National Hurricane Center, copyright 2005. retrieved from satellite measurements by some of the same the images, the wind speed and direction can be calculated methods that are used to retrieve atmospheric temperature (Hubert and Whitney 1971). Figure 3.7 shows an example (e.g., Weng et al. 2003; see also Box 5.2). Figure 3.6 is an of the winds obtained by tracking clouds. example of the vertically integrated water vapor content of Finally, there are several other atmospheric parameters the atmosphere over the ocean measured through clouds in retrieved from satellite data that are useful to forecasters and the microwave portion of the spectrum. These images are are beginning to be used in numerical weather prediction used by forecasters to monitor tongues of moisture from the models. For example, wind speeds around tropical cyclones tropical oceans that can cause heavy rain and flooding when are important for mariners and emergency managers. They they encounter land. are estimated by the Dvorak technique (mentioned above) Winds, or the atmospheric flow field, must be known and also by using microwave soundings. Microwaves pen- to forecast the weather. Scatterometers, such as QuikScat etrate the clouds and allow measurement of the magnitude (launched June 20, 1999), which measure wind speed and of the warm core of tropical storms and its radial gradient, direction near the ocean surface and have “revolutionized the from which wind speeds can be derived (Kidder et al. 1980, analysis and short-term forecasting of winds over the oceans Demuth et al. 2004). at NOAA’s Ocean Prediction Center” (Von Ahn et al. 2006), Precipitation is a fundamental part of the hydrologic are discussed in Chapter 8. Another way to measure winds cycle and is further discussed in Chapter 6; it is often one of is to track clouds in sequences of satellite imagery, usually the first concerns when people think about the weather. Many geostationary imagery. Small clouds travel with the wind. By satellite techniques have been developed to estimate rainfall observing the location of the same cloud in two successive (see, e.g., Barrett and Martin 1981). Today, daily rainfall satellite images and knowing the time difference between estimates are available worldwide on the Internet. The 

WEATHER 23 most advanced precipitation estimation is from the Tropical • The discovery of the mathematical principles that Rainfall Measuring Mission (TRMM) satellite, launched on govern atmospheric flow and the change of phase of water, November 27, 1997 (Simpson et al. 1988, 1996). A joint • The invention of computers and the numerical tech- U.S.-Japan mission, TRMM carries passive sensors of the niques with which to solve these equations, and visible to the microwave portion of the spectrum and is the • The development of observing systems to supply first precipitation radar in space. Many other satellite-derived the needed initial state of the atmosphere. parameters are important to forecasters, including cloud height, cloud top temperature, and cloud phase; fog, smoke, Without doubt, one of the chief reasons for the success and aerosol identification; and skin surface temperature (see, of weather forecasting is that Earth-orbiting satellites provide e.g., Kidder and Vonder Haar 1995). an accurate global initial atmospheric state, which the numer- ical weather prediction models project into the future. Today, satellite data constitute the vast majority of the data available NUMERICAL WEATHER PREDICTION for the initialization of numerical weather prediction models There are several reasons why Arago was wrong in and has the greatest impact of any measuring technology 1846 and why today weather can be forecasted as much as in improving forecast skill. Table 3.1 lists the satellite data 10 days ahead: currently used to initialize models run by NOAA’s National Centers for Environmental Prediction (NCEP). FIGURE 3.6  Vertically integrated water content of the atmosphere (in kilograms per square meter) derived from microwave measurements on six sun-synchronous satellites, three NOAA satellites, and three Defense Meteorological Satellite Program (DMSP) satellites. SOURCE: Data from NOAA; drawing by S. Kidder. 

24 EARTH OBSERVATIONS FROM SPACE: THE FIRST 50 YEARS OF SCIENTIFIC ACHIEVEMENTS FIGURE 3.7  Winds obtained by tracking clouds in successive infrared images. The height of the cloud is determined by the cloud’s temperature. Note that where there are no track- able clouds, no winds can be retrieved. SOURCE: NOAA. TABLE 3.1  Satellite Data Used to Initialize Numerical Figure 3.8 shows a time series of one measure of the Weather Prediction Models in 2006 skill of a representative numerical weather prediction model for 3-, 5-, 7-, and 10-day forecasts. The top line for each set Satellite Data of curves is for the northern hemisphere, where nonsatel- HIRS sounder radiances lite observations are plentiful; the bottom line is for the AMSU-A sounder radiances southern hemisphere, where nonsatellite observations are AMSU-B sounder radiances GOES sounder radiances woefully few. Due largely to our increasing ability to use GOES, Meteosat, GMS winds satellite observations effectively—that is, to assimilate the GOES precipitation rate observations into numerical weather prediction models (e.g., SSM/I precipitation rates Kalnay 2003)—the difference between northern hemisphere TRMM precipitation rates forecasts and southern hemisphere forecasts has steadily SSM/I ocean surface wind speeds decreased, and the overall forecast skill has increased to the ERS-2 ocean surface wind vectors QuikScat ocean surface wind vectors point that global 7-day forecasts are now as good as northern AVHRR SST hemisphere 5-day forecasts were 25 years ago. AVHRR vegetation fraction In addition, tests at NCEP show that data from just one AVHRR surface type satellite instrument, the Advanced Microwave Sounding Unit Multisatellite snow cover (AMSU), extend forecast usefulness by 1 day in the southern Multisatellite sea ice SBUV/2 ozone profile and total ozone hemisphere and by about a half day in the data-rich northern AIRS hemisphere (Lord 2006). AIRS is also improving forecasting MODIS winds skill (Chahine et al. 2006), and there is evidence that satellite Altimeter sea-level observations data, particularly QuikScat winds, are improving hurricane Note: Monthly statistics on the data used in NCEP’s models are available track forecasts (Zapotocny et al. 2007). Without question, at http://www.nco.ncep.noaa.gov/sib/counts/. improvement in numerical weather prediction—on which HIRS = High-Resolution Infrared Radiation Sounder; AMSU = Advanced all forecasts more than a few hours ahead are based—is a Microwave Sounding Unit; GOES = Geostationary Operational Environ- major scientific accomplishment of Earth observations from mental Satellites; SSM = Special Sensor Microwave; ERS = European space. Remote Sensing Satellite; AVHRR = Advanced Very High Resolution Radiometer; SST = sea surface temperature; SBUV = Solar Backscattered Ultraviolet. SOURCE: Lord (2006). 

WEATHER 25 FIGURE 3.8  Anomaly correlation of 500 hPa height forecasts by the European Centre for Medium Range Forecasting. SOURCE: Updated from Simmons and Hollingsworth (2002). Reprinted with permission from the Royal Meteorological Society, copyright 2002.