THE OCEANOGRAPHIC PROBLEMS studied in the Earth Resources Experiment Package (EREP) investigations Program included locating the surface of the ocean relative to the center of the Earth as a function of latitude and longitude; measuring sea-surface temperature more accurately with data from a spacecraft; describing the variation of ocean color and the dynamics of floating ice in space and time; detecting ocean currents and ocean upwelling and the spatial and temporal changes in them; describing erosion, river runoff sediment transport, and the circulation within bays and estuaries bordering the ocean; assessing ocean dumping; and locating productive fishing areas. The meteorological problems were those of measuring cloud features; identifying the vertical and horizontal distribution of aerosols through the atmosphere both qualitatively and quantitatively; describing the radiation and energy budgets of the atmosphere; identifying characteristic airmass properties; and measuring the winds in the planetary boundary layer over the ocean, as determined by sea-surface roughness, to provide data for improved computer-based numerical weather predictions.
The wind, the oceans, the atmosphere overlying the oceans, and the land are resources for mankind. Because of energy-resource-use rates, NASA and Energy Research and Development Administration programs are designed to investigate modern methods of obtaining energy from the winds and the waves.
The oceans moderate and influence climate. Evaporation from them provides the water vapor that is transported by the atmosphere to provide rainfall for the land. Changes in oceanic properties can cause floods and drought. The ocean-atmosphere system provides carbon dioxide for plant life, while acting as a buffer to moderate variations in the amount of carbon dioxide. The oceans also serve as a mechanism for the disposal of waste products. Moreover, the oceans are the major means of transporting the goods of the world in ships. It is important, therefore, to understand sea ice, winds, and waves to improve shipping operations.
The study of the oceans and the atmosphere in an interdisciplinary manner involves many aspects of science and technology. Many features of the ocean have been studied using data obtained by spacecraft. The Gemini and Apollo missions yielded many useful photographs of coastal areas from which oceanographic features could be identified. The meteorological satellites currently being used for study purposes have defined such features of the oceans as the outlines of the Gulf Stream from day to day. The operational meteorological satellites used by the National Oceanic and Atmospheric Administration (NOAA) routinely provide such information as sea-surface temperatures and Gulf Stream boundaries.
All EREP instruments were used in the study of oceanographic problems. The Multispectral Photographic Facility (S190) and the Multispectral Scanner (S192) proved to be most useful in investigation of estuaries, bays, and coastlines; the other sensors proved to  be most useful for studying wide scales of variation in the ocean. On the basis of experience in studying the oceans from spacecraft, it was possible to define a new set of problems that could be addressed by the Skylab EREP instrumentation. The Microwave Radiometer/Scatterometer and Altimeter (S193) was primarily designed for use over the ocean to measure the geoid and to determine the wind fields near the surface of the ocean. However, other valuable information was obtained from this particular instrument as an additional benefit through the recognition of uses other than those originally intended.
The ability to use a resource intelligently often depends on a better understanding of the Earth in a geophysical context. The measurement of the position of the ocean surface (relative to the best-fit ellipsoid of revolution, as described later) is required before further steps can be taken to use oceanic resources. Experience gained in this area may eventually lead to better knowledge of ocean currents, to potential locations for undersea mines and oil wells, and to more accurate maps of the ocean floor.
One of the first applications of Earth satellites was the study of meteorological phenomena. Probably the most successful of the American satellites has been the Tiros meteorological series, begun in 1960. Tiros I transmitted more than 19 000 photographs of cloud formations and added a new dimension to the understanding of the physics of weather. Since then, an ever-expanding number of orbital and geosynchronous weather satellites, equipped with an impressive array of sensors, have been launched to aid in atmospheric research. The operational use of satellites in day-to-day weather forecasting has become commonplace.
With its nitrogen cycle, carbon dioxide-oxygen cycle, and hydrological cycle, the atmosphere is important to the production of foodstuffs on land. In addition, the atmosphere disposes of industrial wastes in the form of smoke and gases. The atmosphere carries the water evaporated from the sea surface over the land and provides rainfall. The properties of the atmosphere control the rate at which water returns to the atmosphere from the land by a process called evapo-transpiration. It is, therefore, not surprising that the EREP investigations Program for Skylab included the study of both the atmosphere and the oceans.
THE GEOID AS MEASURED BY THE S193 ALTIMETER
For small areas and distances, everyone has an intuitive understanding of the concepts of "level" and "vertical." If the factor of appreciable distance is introduced, however, these concepts need to be modified. For example, the support towers on each side of a long suspension bridge are several centimeters farther apart at the top than at the bottom because of the curvature of the Earth. Nevertheless, each tower is vertical. For distances of several hundreds of kilometers, the concepts of a level surface and a vertical line become more complicated.
The term "sea level" is quite familiar; yet, the sea is not level. Even if it were, the problem of determining the distance from the center of the Earth to any point on the sea surface would still exist. The term for the study of this problem is "geodesy," and the location of the level surface of the sea (even as extended over the continents) is called the geoid.
As a first approximation, the Earth is a sphere with a mean radius of 6371 km. In actuality, there is a difference between the equatorial and polar radii of the Earth of 21.4 km. If an ellipse with these radii is constructed and rotated about the polar axis, the result is an ellipsoid of revolution, or an oblate spheroid that more closely approximates the level surface of the sea.
Sir Isaac Newton was the first to derive the equations that show that the orbit of a spacecraft is an ellipse the orientation and eccentricity of which are determined by the method of placing the spacecraft in orbit. To derive this result, Newton had to assume that the Earth consisted of concentric ellipsoids of revolution, each homogeneous in mass within that layer. The problem is, of course, that for an orbit near the Earth, the Earth is very large compared to the orbiting spacecraft. His famous equation
states that the force attracting two masses is given by a universal constant G times the product of these two  masses m1 and m2 divided by the square of the distance r between them. The equation for the gravitational force between the Earth and a spacecraft had to be generalized from this equation by integrating over the volume of the Earth with a variable density assigned to the volume elements. It could be expected that the forces acting on a spacecraft such as Skylab would differ from place to place along the orbit depending on the nature of the Earth below. Such is indeed the case, and the actual orbit of a spacecraft, especially if it be rather low, departs substantially from the ellipse derived by Newton. The oblateness causes the orbit to change plane relative to the geometry of the stars, and the inhomogeneities in the distribution of mass in the upper layers of the solid Earth cause the path of the spacecraft to vary about the theoretical ellipse by substantial, measurable amounts.
The problem becomes more complicated over the ocean surface because the surface of the ocean is closer to these inhomogeneous concentrations of mass caused by such factors as the properties of the large plates that form the upper crust of the Earth. The oceans try to reach an equilibrium surface determined by an appropriate integration over the masses in the volume of the solid Earth. If there were no winds generating ocean currents, if there were no ocean tides, if there were no cooling at the poles and no heating at the Equator, and if the oceans did not vary in their saline content, then the surface of the ocean would be level and it would correspond to the concept of the geoid.
The ocean is very nearly level because all the effects cited cause it to depart from the geoid by at most 2 or 3 m (with a few notable exceptions, such as the tides in the Bay of Fundy), whereas the geoid moves toward and away from the center of the Earth by amounts that depart from the ellipsoid of revolution by as much as 100 m.
Sputnik, the first satellite, was launched by the U.S.S.R. in 1957 and tracked by British scientists who used the measurements of its orbit for comparison with previously made measurements based on land distances. The ellipticity of the Earth was confirmed and measured independently in this way. Many subsequent orbital spacecraft have been tracked very carefully, and the perturbations in their orbits have been used to calculate some of the characteristics of the geoid.
Before Skylab, the problems associated with using the orbits of other spacecraft to determine the geoid were becoming increasingly difficult to solve because greater accuracy was required. This difficulty arises because the effects of the smaller scale variations of the geoid fade rapidly with height and hence have little effect on the orbital spacecraft. A lower limit exists at which spacecraft can be orbited before the drag of the upper atmosphere causes them to slow down and fall back to Earth; hence, important details of the geoid cannot be sensed. Programs were developed to generate a geoid by combining spacecraft measurements and the limited number of measurements of gravity variability over the surface of the Earth. The results led to numerous geoids, one being the Marsh-Vincent geoid (or the Goddard Earth Model 6 (GEM-6)), shown in figure 5-1 (ref. 5-1).
The GEM-6 geoid contours the departures of the surface of the ocean, as continued by theory into the continents, in units of meters as if they were measured in terms of the distance from the ellipsoid of revolution. Several features are noteworthy. For example, to this level of definition, the surface of the ocean is 100 m closer to the center of the Earth at a point in the Indian Ocean than is the ellipsoid of revolution. Other areas are as much as 40 or 50 m farther from the center of the Earth than is the ellipsoid of revolution. This figure does not represent the correct geoid because the complete geoid has yet to be measured; nevertheless, many of the major features are correct.
Before the Skylab missions, analysis of orbital data made it possible to distinguish approximately 20 geoid oscillations around the Earth; the shortest oscillation wavelength that could be resolved was approximately 2000 km at the Equator. (The use of precisely obtained gravity measurements allowed shorter wavelengths to be determined, but gravity data are expensive and difficult to obtain.) Altimeter data from Skylab resolved oscillations that are 20 km long and thus produced an improvement in spatial resolution by a factor of 100 in the horizontal dimension.
An investigation of the problem of combining Skylab tracking data and calculations of the orbit with the altimeter measurements of the distance between Skylab and the ocean surface was conducted by Mourad et al. (ref. 5-2). The accurate calculation of the Skylab orbit....
...over a short arc depended on precise knowledge of the spacecraft location and velocity at the start of the arc. It was shown that, for reference ellipsoids that differ in mean radius by 1 part in 60 000 and in flattening by 5 parts in 30 000, the calculated altitudes for Skylab would differ by 10 m after traveling only 10° of longitude. The actual altitude of Skylab varied by as much as 1 km over a 100-km arc. It is necessary to determine these altitudes to an accuracy of a few meters before the full potential of spacecraft altimetry can be realized Large bias terms for different Z-axis-to-local-vertical passes indicated large differences from one orbital segment to another.
The high-frequency components of the geoid are nevertheless easily detected. Examples from the S193 altimeter are given by McGoogan et al. (ref. 5-3). The first pass to be discussed began as the Skylab spacecraft crossed the east coast of the United States and proceeded on a southbound orbit across the island of Puerto Rico. The height of the mean sea level relative to the reference ellipsoid as obtained from the difference of the altimeter range measurement and the computed satellite orbit is shown by the irregular line in figure 5-2(a). The altimeter trace fluctuates as much as I m vertically over the distance that was measured. These irregularities are caused by the difficulty of measuring time accurately for the noisy return of the radar pulse in which a change of I nanosecond represents a 15-cm change in range.
A smooth curve drawn through the altimeter geoid trace would very nearly represent the level surface of the ocean. Over the relatively short distance of approximately 220 km, the surface of the ocean is 12 m closer to the center of the Earth above the 8-km-deep Puerto Rico Trench (fig. 5-2(b)) because of the variation in the forces that produce the "level" sea surface. On the eastern side of the island is a slight dip but nothing as pronounced as the variation over the trench. A section of GEM-6 data along the same orbit is reproduced in figure 5-2(a). This trace differs from the altimeter measurement by 20 m east of Puerto Rico and by 30 m near the lowest part of the altimeter trace. The 20-m difference could have been caused by an error in locating Skylab in altitude. However, displacing the GEM-6 curve so that it coincides with the right-hand side of the altimeter geoid would still leave substantial differences for the geoid over the Puerto Rico Trench. The EREP passes over the Puerto Rico Trench were repeated many times for slightly different sections of the trench and similar results were obtained each time.
In figure 5-3(a), the altimeter geoid trace is shown from the pass over the Jaseur Seamount in the South Atlantic Ocean, east of Vitoria on the coast of South America. The water around the seamount is 4 km deep (fig. 5-3(b)). The seamount rises to within approximately 200 m of the sea surface. As shown in figure 5-3(a), the sea surface rises approximately 10 m with respect to the center of the Earth over this seamount. The reference geoid from figure 5-1 is essentially at the same distance from the center of the Earth along this entire track and, as the altimeter measurements indicate, there is a difference of 12 m from the left side to the right side of figure 5-3(a). Altimeter data from a pass over a seamount near the Cape Verde Islands (fig. 5-4) show the same phenomenon. The sea surface moves away from the center of the Earth by as much as 10 m. A smooth curve through the altimeter geoid would represent very closely the level surface of the sea except for minor effects of, at most, 1 m in these areas.
One of the striking features of the sea bottom is the continental shelf that is adjacent to some, but not necessarily all, continental coasts. Off the eastern coast of the United States in the general area of Florida, the edge of the shelf is called the Blake Escarpment. As shown in figure 5-5(b), the coastal water is initially shallow, then deepens to 1000 m, and finally, over a very short distance, deepens to approximately 4800 m.
It is important to note that the altimeter trace (fig. 5-5(a)) varies relatively slowly over the western portion out to the edge of the depth increase. Over the Blake Escarpment, the geoid trace moves toward the center of the Earth by 10 m over a very short distance and then rises slightly over the rise in the sea floor to the east. The GEM-6 geoid (fig. 5-1) does not show this very steep change over the Blake Escarpment.
Most of the time, the EREP was operated in the Earth-viewing mode for only 5 or 10 minutes, after which the Skylab spacecraft was returned to the solar-inertial mode. However, on January 31, 1974, the EREP was operated in the Z-axis-to-local-vertical mode for one complete Earth orbit; 26 segments of the geoid were determined for this orbit. (The locations of 23 of the 26 altimeter measurements are plotted on the GEM-6 model in figure 5-1.) The theoretically derived geoid from spacecraft and gravity data locates major features of the geoid and, along this subspacecraft track, the total range in the distance by which the geoid differs from the ellipsoid of revolution is approximately from -40 to 70 m. In general, the S193 altimeter measured a position for the geoid that reasonably approximated the theoretical curve, as shown in figure 5-6. The greatest discrepancies are for segments 202, 203, 222, and 223.
Each of these bursts of rapid fluctuations in the altimeter measurement, indicated as a pass on this diagram, could be expanded into a set of curves similar to those previously shown. Several segments during this around-the-world pass also showed some very interesting results.
Segments 209 and 212 and the corresponding submarine topography are illustrated in figures 5-7 and 5-8, respectively. Segment 209 shows no striking correlation with the submarine topography such as might be expected from the sharp increase in the depth of the ocean represented on the left side of figure 5-7(b). Of interest is that the theoretical geoid (GEM-6) agrees with the altimeter geoid represented on the right side of figure 5-7(a) and departs significantly from it (by approximately 10 m) on the left side.
The deepest point in the ocean is in the Marianas Trench. The S193 altimeter was operated over this.....
....trench but in a region only approximately 5500 m deep (fig. 5-8(b)). Nevertheless, the Marianas Trench produced a very sharp V-shaped feature in the location of the geoid (fig. 5-8(a)). Such effects on the geoid also occurred over the Mid-American Trench (figs. 5-9(a) and 5-9(b)).
The variation of the geoidal surface toward and away from the center of the Earth, as a response to variations in submarine topography, is clearly indicated in figures 5-2 to 5-5. The sea surface need not follow every rise and fall on the sea bottom as was shown in these illustrations. Some portions of the Earth's crust are in isostatic equilibrium; that is, they have risen or sunk as a function of mass to achieve an equilibrium position. Other portions of the crust near and at seamounts and oceanic trenches have not attained this equilibrium. If the seamounts and the trenches just illustrated were in isostatic equilibrium, the geoid would not be expected to respond to their presence. Over some parts of the ocean, this isostatic equilibrium very nearly exists and no remarkable difference in the altimeter geoid is observed over such regions (fig. 5-7).
In the preceding figures, the altimeter traces represent the level surface of the sea, except for relatively minor effects important to oceanography. These traces show that, generally, the level surface of the sea moves away from the center of the Earth over seamounts and toward the center of the Earth over trenches. The estimates of the geoid that were calculated until about 1968 did not include these important features and did not correlate very well with plate tectonics theory. When the radar altimeter concept was first proposed in 1968, Greenwood et al. (refs. 5-4 and 5-5) predicted that the altimeter would resolve the details of the geoid so well that the geoid would begin to show features that would be correlatable with plate tectonics. The illustrations that have been presented show that none of the finer scale fluctuations, such as the rapid variation over the Blake Escarpment, the rises over seamounts, and the dips over the submarine trenches, were in any previous geoidal model. They all correspond to various aspects of plate tectonics theory, and, thus, the theories of geodesy and tectonics are becoming more integrated.
Sensing systems on operational meteorological satellites routinely acquire image data in the infrared region of the electromagnetic spectrum. From these data, temperatures of cloud tops can be determined and, if an unobstructed view exists between the satellite and the sea surface, the temperature of the sea surface can be estimated. This measurement is made using one wavelength band in the infrared. However, it is necessary to apply corrections, which are predominantly based on climatological concepts, for the intervening atmosphere.
Infrared measurement of sea-surface temperature is difficult because the surface water is undergoing evaporation and there is a very sharp gradient in temperature at the surface of the sea, such that over a vertical distance of approximately 50 mm, the temperature decreases by as much as 1 or 2 K, being colder at the surface. This sharp gradient allows for the flow of heat from the deeper layers of the ocean through the air/sea interface into the atmosphere. What is actually measured is some temperature related to this skin temperature, so a small correction for this effect must be made.
The basic problem, however, is that, as the radiation from the sea surface travels toward the sensor on the satellite, it is both attenuated by the atmosphere and contaminated by atmospheric emission, with the result that signal strength at the satellite is usually lower than it would have been had it traveled through a vacuum. The overall effect is that the infrared temperatures sensed by the recording instrument can be as much as 8 K cooler than the temperature of the sea surface. The single wavelength band measurements can be corrected for this effect; but, because the correction depends on  the actual properties of the intervening atmosphere, occasional errors in measuring the properties can produce unacceptable sea-surface temperatures.
The infrared temperatures of a water surface that would be sensed by integrating over one wavelength band of the infrared Spectrometer (S191), as compared to integrating over another wavelength band, differ substantially because the extinction coefficients for the intervening atmosphere are functions of wavelength. The basic concept of the experiment conducted by Anding and Walker (ref. 5-6) was to sense two different temperatures in terms of the measurements at two different wavelengths and to calculate their relative extinction coefficients. Then, on the basis of valid linearizing assumptions, the two values of the temperature as plotted against the two relative extinction coefficients were connected by a straight line to extrapolate to the condition in which the relative extinction coefficient would be zero.
This concept was tested in two regions for which surface truth was available, one off Key West, Florida, with surface-truth data obtained on January 8,1974, and the other in Monroe Reservoir with surface truth obtained on June 10, 1973. In the Key West test, the actual sea-surface temperature was 296 K; the raw infrared measurements at the spacecraft indicated, for the most nearly transparent region of the atmosphere in the infrared, a temperature near 290 K; for the next most transparent region, approximately 289.5 K; and for the more opaque regions, approximately 286.5 K. The relative extinction coefficients were approximately 0.2, 0.33, and 1.4, respectively. When a straight line is fitted to these data, the intercept for a zero relative extinction coefficient is 293.6 K. The error in the method based on the Skylab measurements was thus 2.4 K; if the raw infrared temperatures for the best window had been used, a 6-K difference would have been measured in the infrared. As a check, a model of the theory produced a value of 294.2 K, which indicated some additional sources of error in the S191 data.
A similar experiment was conducted at Monroe Reservoir near Salem, Illinois, where the in situ temperature was measured as 298.0 K. The model yielded 298.6 K, and the value computed from the Skylab measurements was 297.7 K. In this instance, the measurement and the theory differed by only 0.3 K (fig. 5-10).
In addition to Skylab data, Nimbus data (specifically infrared interferometer spectrometer (IRIS)) were used....
....to test this hypothesis. It was found that similar results could be obtained with the IRIS data.
The results of this particular experiment show the value of bands in the infrared region for measuring sea-surface temperatures. The next generation of Tiros satellites will contain two bands in the infrared region for more accurate sea-surface temperature measurements.
Temperature, Salinity, and Sea-Surface Roughness
The presence of clouds is a problem in remote sensing of changes in ocean temperature and salinity because clouds preclude the measurement of temperature in the infrared. The main reason for studying the relatively low microwave frequencies, such as those used with the Skylab L-Band Radiometer (S194), is that clouds and rain should have a minor effect on the measurement; hence, in theory, sea-surface temperature should be measurable through clouds if an appropriately chosen frequency is used.
 The strength of the passive microwave signal received by the S194 was a function of numerous physical variables as studied by Hollinger and Lerner (ref. 5-7). These variables include the amount of sunglint reflected by the sea surface back to the antenna and the variations in temperature, salinity, and sea-surface roughness. The S194 data were studied in such a way as to isolate these variables to ascertain whether or not, upon removing the effects of all but one cause, the variability due to that cause could be detected.
When operated over the ocean, the S194 obtained data for conditions in which the oceanic salinity varied from approximately 32 to 36.5 parts per thousand. The sea-surface temperature as provided by the Fleet Numerical Weather Central varied from approximately 274 to 301 K, and the windspeed (the cause of sea-surface roughness) varied from 0.5 to 24 m/sec.
Hollinger and Lerner developed techniques to calculate the microwave temperature when the salinity, windspeed, and temperature are known. Then, the effects of two of these three variables were removed so that the remaining one could be studied. Reflected "sunlight" at this wavelength was eliminated simply by not using those data in which it was present. The dependence on salinity was quite weak, and the scatter in the microwave temperatures, after removal of windspeed and temperature effects, was 2 to 3 K about the theoretical curve. As the salinity varied from 32 to 37 parts per thousand, the theoretical temperature curve decreased from 98 to 95 K for a variation of 3 K. The temperature scatter was larger than the range of the salinities.
Near the mouths of rivers and estuaries, where the salinity can typically increase from values near zero in freshwater to 32 to 34 parts per thousand over 15 or 30 km as the ocean is approached, changes in salinity could be detected by an instrument having high spatial resolution. Because oceanographers require salinity measurements to two significant figures past the integer values (in parts per thousand), passive microwave systems are not appropriate remote sensors for salinity measurements in the open ocean.
The sensitivity of the S194 to variations in sea-surface temperature was also studied. The theoretical curve for the antenna temperature that would be measured by the S194 varies from approximately 96 K near a 273-K ocean temperature, through a modest maximum 0.5 K higher at a 288-K ocean temperature, and then dips to 94 K at an ocean temperature of 300 K. Because of the excessively large point scatter throughout this range, it was concluded that the S194, or any other instrument operating at a frequency of 1.4 GHz, was not suitable for measuring either salinities or temperatures of the sea surface on the open ocean.
The last ocean variable studied (ref. 5-7) in S194 measurements was the dependence of the passive microwave temperature on windspeed. The passive microwave temperature increased from approximately 96 K to approximately 103 K as the windspeeds varied from 2.6 to 25 m/sec for the passes that were studied. A regression line yielded an equation indicating that the passive microwave temperature increased at a rate of 0.31 K per 1-m/sec increase in windspeed. The regression equation that was obtained for antenna temperature is given by Ta = 0.31 W + 95.2, where W is the windspeed in meters per second. The mean-square difference between the windspeed predicted from the passive microwave temperature and the "observed" windspeed, after removing all these other effects by means of theoretical considerations, was approximately 4 m/sec. The temperature varied approximately 10 or 15 percent about an average value of 98 K.
All the results obtained in this investigation were anticipated, except for the relatively high scatter of the data. The theories concerning instrument measurement capabilities were verified remarkably well. The experiment was successful in verifying the theory of the microwave emission of the sea surface at a frequency of 1.4 GHz. With the theory confirmed at this particular frequency, measurements at a different frequency, at which a greater sensitivity to variations in sea-surface temperature exists, may ultimately permit the determination of sea-surface temperature through clouds. In fact, two five-frequency scanning passive microwave systems are being built for use on the Seasat-A and Nimbus satellites; theoretically, these instruments will be capable of determining sea-surface temperature, surface windspeed, and other sea-surface parameters through clouds.
The conventional methods of ascertaining water depth off a given coastline are sounding (dropping a line from a ship and determining when it hits the bottom) and echo ranging (sending a sound pulse to the bottom and measuring round-trip travel-time). Both methods require considerable time and extremely accurate navigation so that the position of the ship as a  function of latitude and longitude is known for each measurement. The most important areas for the determination of depth are those in which the water depth is less than approximately 30 m. Large amounts of money are spent annually by the maritime nations to update their water-depth charts, especially in areas of important harbors and active navigation. Over the course of 10, 20, or 30 years, the depths of the water in a given region can change. Sandbars can shift their location, new shoals can form, and others can be eroded away. The problem of correctly describing the depth of the water is thus an ongoing problem that requires continuous correction of navigational charts.
In ocean areas characterized by clear water, sunlight in the blue-green portion of the spectrum can penetrate to considerable depth. Although somewhat attenuated, the light can be reflected from the bottom and travel through the water to the spacecraft, if no clouds are present and if the bottom is not excessively deep. This technique is unusable where the water is turbid or discolored, but, over large areas of the ocean, bottom-reflected light can be viewed by the spacecraft and imaged by either a camera or a scanner system such as the S192. Trumbull (ref. 5-8) reported that a stereoscopic view of the bottom could be seen using S190 photographs. He stated that it is possible to separate bathymetric detail and turbid water effects by means of multispectral information in the photographs and in the multispectral scanner data. images from the wavelength range 0.6 to 0.7 µm show only turbidity features, but images from the wavelength range 0.5 to 0.6 µm show both bathymetric and turbidity features.
The two unwanted effects in the spacecraft images are the scattering of light by the intervening atmosphere and the specular reflection of diffuse skylight by the water surface. These effects produce a background signal that must be subtracted to obtain useful data. To determine the amount of unwanted signal to be subtracted from the signals where the bottom has been imaged, data must be collected over deep water where there is no bottom-reflected signal. When the unwanted signal is subtracted, a measured brightness remains that can be related to depth and other physical parameters. These parameters are the atmospheric transmittance, the solar irradiance on the water surface, the bottom reflectance, the index of refraction of water, the light attenuation coefficient of the water, the angle of observation (after refraction under water), and the solar zenith angle (after refraction under water). Some of these parameters can be calculated from theory; others have to be determined for the area in which the depth is to be computed. In particular, the attenuation coefficient and the bottom reflectance need to be determined for each area. Variations in bottom cover can cause differences in reflection.
Three methods for extracting depth information using the theory just described were developed. One was a single-band method. Another used two bands and exploited differences in bottom reflectance and underwater light attenuation differences. The third combined two bands of information in an optimum-decision technique.
The theories were applied to S192 data obtained over eastern Lake Michigan and the western coastal waters of Puerto Rico. The results of the analysis of the latter area are shown in figure 5-11. The vertical axis in the figure is logarithmic and shows the scale for the brightness of the point in the image minus the deepwater brightness level (V - Vs). Calibration points (squares for band 3 and circles for band 2) are points for which the depth is known. The agreement is quite good to a depth of approximately 16 m.
These theories were used to produce a bathymetric chart (fig. 5-12(a)) for the waters off the western coast of Puerto Rico with the use of S192 data from bands 2 and 3. For the same region, the conventional depth chart with soundings in fathoms is shown in figure 5-12(b). Over the Escollo Negro region, two line segments were selected for verification. Line 1450 of the digitized version (fig. 5-12(a)) was located and compared with the conventional bathymetric chart, as shown in figure 5-13. In such a verification effort, the observed depths are just as questionable as the Skylab......
.....depths because of the limited accuracy and the spottiness of data obtained by conventional methods. The agreement is quite good.
Applications to updating world navigational charts and mapping changes in near-shore bottom topography are forecast, based on the results that were obtained from analysis of the Skylab data.
According to Campbell et al. (ref. 5-10), pack ice begins to form in the Gulf of Saint Lawrence in December and reaches its maximum extent in March, after which it begins to melt and retreat. At that time of the year, the ice interferes with shipping; therefore, it is important to understand the growth and movement of this ice as an aid to safe navigation. The gulf extends from latitude 45° to 50° N and from approximately longitude 58° to 68° W.
The Skylab orbit enabled the Skylab 4 astronauts to obtain photographic and scanner data for the program conducted by Campbell et al. Because of the 50° inclination of the orbit, successive ascending nodes grouped near latitudes 50° N and 50° S. Consequently, Skylab passed over the gulf on January 6,11,14,18,19, 20, and 21, 1974, during daylight hours, and hand-held-camera photographs were obtained on all 7 days except for January 14, when S19OA and S19OB photographs and S192 imagery were obtained.
The skies were unusually clear, and the photographs enabled identification of many ice, cloud, and snow features in the gulf region. In winter, many of the extratropical cyclones that form over the United States deepen as they move northward and obtain full development as they travel toward the gulf; thus, this area is one of the cloudiest and windiest regions of North America. In the polar regions of the world, the continuous and effective mapping of floating ice by either photographic or infrared imaging systems on either aircraft or spacecraft is prevented by clouds.
In supporting the Skylab data acquisition, two aircraft, an oceanographic ship, a helicopter, three aircushion vehicles, and trucks were used to acquire....
....remotely sensed data during the Skylab overpasses. A side-looking imaging radar at a 3-cm wavelength was used on one aircraft, and a passive microwave imaging system was used on another. Other aircraft instruments included an infrared scanner, a multifrequency microwave radiometer, and two RC-8 cameras. The ships and the hovercraft were used to obtain photographs and to measure ice growth, ice thickness, and other conditions. The trucks obtained data on ice near the shores.
Examples of the spacecraft imagery obtained over the gulf from January 14 to 21, 1974, are shown in.....
....figures 5-14 to 5-17. Figure 5-14 shows details of the ice such as color, leads, and coverage around Prince Edward Island on January 14. The ice cover was not as extensive as it was on January 18. The photographs on January 20 document a remarkable growth of the ice over a 2-day period; the rapid eastward extension of the area covered by ice is also evident. The extent of open water areas around Anticosti island had changed appreciably.
Figure 5-18 is an ice reconnaissance map of the gulf for January 18. This map (one of four in ref. 5-10) is an operational real-time product that uses NOAA-2, Landsat, and aircraft reconnaissance data. Although the Skylab photographs were not used in preparation of the map, there is general agreement between the photographs and the map. Differences can be seen, however (e.g., the area of open water south of Anticosti island). The streaming and the eddying of the ice to the southeast of Anticosti island are important and their changes were traced from analysis of the other photographs.
The aircraft program yielded data from active and passive microwave systems that were analyzed to infer the age of the ice. An example of a side-looking imaging radar product is shown in figure 5-19. The amount of information in side-looking airborne radar imagery is impressive: pressure ridges, shear ridges, floes of all sizes, and plumes are clearly discernible. The radar data currently are useful for interpretation of the surface features of ice but cannot provide information on the age and thickness of the ice.
Data from the airborne passive microwave imaging scanner are shown in both horizontal and vertical polarization in figure 5-20(a). Supporting aircraft photographs of different types of ice are shown in figures 5-20(b) to 5-20(d). The microwave temperature (fig. 5-20(a)) is color coded from red as the "warmest" down through the color spectrum as an aid in visual interpretation. The brightness temperature (a term equivalent to the passive microwave temperature) is different for the two polarizations and for different....
.....kinds of ice. Analysis of the data showed that (1) the vertically polarized brightness temperature is greater than the horizontally polarized brightness temperature for all ice types, and (2) as the ice ages and thickens, the brightness temperature T increases in both the horizontal H and vertical V polarizations and the difference between the two decreases. These conclusions are summarized quantitatively in table 5-1. The values for vertical polarization increase by 75 K; the values for horizontal polarization increase by 160 K; and the difference Tv- TH decreases from 130 to 35 K. Passive microwave data are a good means of determining ice age and thickness.
A major conclusion of Campbell's study was that "the best all-weather floating ice remote sensing data can be obtained via the combined operation of passive and active microwave systems." This goal is being fulfilled, at least in part, by existing and planned active and passive microwave instruments on satellites.
Passive microwave imaging systems on Nimbus satellites routinely image ice in the Arctic and Antarctic at a fairly coarse resolution. A radar is being built for Seasat-A that will image ice with a resolution of 20 m and that will be capable of imaging a 100-km-wide swath through clouds day or night, whenever the satellite transits an interesting region. Extensive plans are being made for use of these data operationally by Canadian and U.S. scientists.
 Fisheries, Chlorophyll, Water Color, and Productivity
The world fishing effort exploits an ocean resource. The total weight of all the fish caught by the major fishing nations increased to a peak in 1972 and then decreased in both the following years. This trend suggests that certain areas are being overfished and that the management of this resource has not been directed toward the concept of maximum sustainable yield.
The presence of fish in a certain area of the ocean depends on the availability of food in the forms of zooplankton and phytoplankton. Phytoplankton, being small plants, require the equivalent of fertilizer and sunlight. The fertilizer, called nutrients, consists of the dissolved nitrates, phosphates, and carbon compounds in the water. Because sunlight only penetrates the first 100 m of the ocean, the food chain in the oceans and estuaries starts in the surface layers.
In many areas of the ocean, the nutrients are not abundant enough to support the food chain and produce large numbers of fish. Mid-ocean areas at latitudes 30° N and 30° S, where the surface waters sink, are particularly devoid of nutrients.
The ocean areas in which nutrients are plentiful are coastal areas with river runoff Arctic and Antarctic areas where the water overturns each winter to bring submarine nutrients to the surface, and upwelling areas off the western coasts of continents at certain latitudes. In these upwelling areas, the general circulation winds blow toward the Equator, and the drag of the wind on the sea surface forces the coastal water out to sea and thus causes nutrient-rich colder water from several hundred meters below the surface to rise.
Numerous investigations have indicated that water with certain colors, probably caused by phytoplankton, and certain temperatures, often rather narrowly defined, are preferred by various species of fish. If these regions of the ocean could be located, fishing vessels conceivably could be directed to them. The result would be a reduction in the cost of fish and an increase in the availability of these fish to the consumer.
Numerous properties of the ocean surface can be remotely sensed as an aid in locating potential fishing areas. These properties, called signatures, need not necessarily be the same for every area. Those that can indicate the presence of abundant phytoplankton are the temperature variations and the diverse colors of the water. The distinctive colors of the water may be caused by the presence of chlorophyll or by suspended sediments that may indicate nutrient-rich runoff
Several investigations within the overall concept of fisheries were concerned with detection of the nutrient-rich portions of the surface waters in oceans, bays, and estuaries by means of some direct signature that could be sensed by the Skylab instruments. Eight reports are of interest: Pirie and Steller (ref. 5-11), Nichols (ref. 5-12), Gordon and Nichols (ref. 5-13), Marshall and Bowker (ref. 5-14), Korb and Potter (ref. 5-15), Szekielda (ref. 5-16), Watanabe et al. (ref. 5-17), and Savastano (ref. 5-18).
Areas of upwelling occur along the continental coast and are the location of important fisheries. Coastal upwelling occurs when winds blowing parallel to the shoreline or slightly offshore cause the warm surface waters to move seaward. To replace that loss, waters from depth rise to the surface. The subsurface waters are cooler and contain more nutrients. Pirie and Steller (ref. 5-11) and Szekielda (ref. 5-16) used EREP data to study this phenomenon off the California and West Africa coastal areas, respectively.
An illustration of the sensing of upwelling using EREP data is shown in figure 5-21. The light areas in the water are caused by suspended sediments borne by the rivers that discharge into the ocean along this portion of the California coast. The suspended sediments flow southward along the coast and spread outward into the ocean at discrete regions. (The arrows indicate current direction.) This mixing of the coastal water containing the suspended sediment with the offshore water is not very efficient, with the result that sharp boundaries between the offshore water and the coastal turbid water can be seen. The five fingerlike projections in the upper left portion of the figure illustrate this phenomenon. Three regions (U) that are noticeably darker than the surroundings represent areas without the turbidity associated with the river water. Based on the location of these regions, oceanographers would conclude that they are composed of water that has upwelled from the deeper layers to the west.
Similar effects can be detected in the imagery obtained by operational spacecraft, as shown in figure 5-22. This image, obtained September 11, 1974, from NOAA-3 data (ref. 5-19), illustrates the same phenomenon on a wider scale off the California and Oregon coasts. In this thermal-infrared image, light shades indicate cold water; dark shades, warmer water.
The light gray along the coast is water with a temperature of approximately 287 K, and the dark gray offshore represents temperatures approximately 6 K warmer. The location and the strength of the upwelling areas along the coast are readily visible in this image; the general patterns are similar to those seen in figure 5-21. The upwelled water mixes with the river water. The result may be high nutrient concentration in the upwelled water, moderate concentration in the mixed river and upwelled water, and low nutrient concentration in the offshore water.
Examination of both these images shows that sharp boundary lines representing zones of discontinuity between two kinds of surface water are seen as far as 50 km from shore. Most conventional shipboard oceanographic measurement techniques cannot delineate these very complicated patterns. Thus, remote sensing provides oceanographers with the ability to resolve new scales and to investigate new problems in the circulation of the coastal waters. A review of the oceanographic literature more than 10 years old would reveal that most physical oceanographers were unaware of this scale of complexity in sea-surface properties.
That upwelling-or the absence of it-along the western coast of North America is important information desired by the fishery industry is shown in the re....
 ...-cent work of Bakun (ref. 5-20) and Bakun and Nelson (ref. 5-21). The first reference provides a tabulation of the daily and weekly upwelling intensities along the western coast of North America from 1967 to 1973. The second discusses the method by which these quantities are computed and describes the climatology of upwelling processes off the coast of Baja California.
These two studies, although not a part of the EREP Program, show how upwelling patterns determined from imagery could eventually be correlated with winds measured by a radar scatterometer. In both cases, the details of the wind fields near the coast are determined from the conventional meteorological information and are then used to compute the stress of the wind on the sea surface, the currents produced by this stress, and the area of upwelling. The volume of water that may be involved in the upwelling seen in figure 5-21 was calculated to be as much as 343 m3/sec per 100 m of coastline. The amount of upwelling peaks in June and July and is still nearly always present both at latitude 39° N, longitude 125° W, and at latitude 36° N, longitude 122° W, in September. For both areas, weekly averages usually exceed 50 m3/sec per 100 m of coastline with peak amounts as high as 250 m3/sec per 100 m of coastline.
An oceanic region similar in circulation and climate to the region off the western coast of North America is off the coast of Spanish Sahara and Mauritania in an area from latitude 18° N to 22° N, longitude 16° W to 18° W. The well-known upwelling phenomenon in this area is caused by meteorological conditions similar to those off California (ref. 5-16). An S190A color photograph off Cape Blanc, where the land is desert, is shown in figure 5-23. The different colors of the water are clearly evident.
Numerous images from both Landsat-l and Skylab have been obtained for this region and were used to map the ocean color boundaries (fig. 5-24). The plots revealed that the boundaries fluctuated from 10 to 20 to perhaps 40 km from one image to the next and that they were quasi-persistent features of the sea surface in that area. (A feature is defined as quasi-persistent if it lasts several weeks before marked change is observed.)
Analysis of the EREP data shows that upwelling features are relatively small scale compared to major ocean features studied by means of shipboard oceanographic measurements. The same features were observed in the imagery off southern California, and it is evident that remote-sensing techniques make it possible to delineate these areas more accurately. Because the features are quasi-persistent, ample time exists for a fishing fleet to reach the more productive areas.
Many fish caught in the continental shelf waters off the U.S. east coast are adults that matured from fry in the nearby estuarine areas. The fish are hatched in the estuaries and spend their early years in the marshes and shallow areas until they are large enough to venture to sea. Understanding the processes and the phenomenon of these estuaries is thus important in fishery research. A productive food chain in and near the mouths of these estuaries requires dissolved nutrients and sunlight, the same essential conditions that are present off California and north Africa. The dissolved nutrients are derived through runoff from the adjacent landmass and by means of the treated and untreated sewage effluents that are discharged by the coastal cities and towns. A substantial amount of the fertilizer that is used by farmers on their fields is dissolved by rainwater and carried by small streams and brooks into these estuaries as a continuing process.
In contrast to upwelled water, the nutrient-rich estuarine water can have many different origins. Perhaps a primary indicator of its origin is the water color, which depends on the amount of sediments that have been washed down with the water from the land. Three investigators (refs. 5-12 to 5-14) attempted to classify as to chlorophyll and productivity indices the surface features of the waters off the U.S. east coast near Assateague island, in the lower Chesapeake Bay, and in the Rappahannock River. The results were only partly successful in these particular cases and showed no discernible water-color parameter in the imagery that could be strongly correlated with chlorophyll in the water and only a fair correlation of chlorophyll with turbidity.
The overall result of the study was that it was possible to classify the estuarine waters into numerous major categories according to color and that these categories....
....and their positions in the river could be explained on the basis of the tidal circulation in the estuaries. The investigators indicated that repetitive coverage every 2 or 3 hours is needed to study tidal effects. Such coverage is impossible from a low-orbit spacecraft such as Skylab. It may be possible to study phenomena such as these from geostationary satellites on an hour-to-hour basis. However, because of the great distance between the sensor and the target surface, the low resolution of the imagery is a major problem. An alternative is aircraft reconnaissance, and many of the shortcomings of aircraft reconnaissance were discussed in the reports.
The study by Korb and Potter (ref. 5-15) highlights some of the problems of studying ocean features with remote sensing. The surface-truth part of the program was successful; but, of the two EREP data sets obtained for the study areas, one was badly contaminated by Sun glitter and the other had 95 percent cloud cover. They found that the values of chlorophyll in the Matagorda Bay, Texas, area ranged from 1.0 to 14.5 mg/m3 and that the values of turbidity ranged from 0.9 to 25.0 Jackson turbidity units. The correlation coefficient calculated for the chlorophyll and turbidity data was 0.68. This value indicates that the levels of chlorophyll and turbidity were correlated at a confidence level of greater than 99 percent for the bay and near-coast data. Because turbidity can easily be sensed remotely, this result, if it is generally characteristic of coastal regions, would have important implications for studies concerned with the remote sensing of chlorophyll in these regions.
Watanabe et al. (ref. 5-17) used image masking and stereography to study the S190A photographs of Japanese waters in an area south of Hokkaido near latitude 40° N, longitude 145° E. In image masking, the difference between visible (0.5 to 0.6 µm) and near-infrared (0.7 to 0.8 µm) wavelengths is used to identify oceanic patterns or green sea areas and to distinguish them from atmospheric formations. When overlapping photographs were viewed stereoscopically, the clouds could be differentiated from the sea surface and the ocean patterns identified. These latter areas were correlated with abundant phytoplankton that demarked the boundary between the Kuroshio and Oyashio Currents.
In the cited studies, parallel theoretical and observational efforts were made to acquire a better understanding of the various regions and to obtain adequate surface truth. The theoretical effort is assimilated into the body of knowledge for both the direct observations and their future applications to other remote-sensing pro" grams. Just as instruments and systems evolve with time and become progressively better, so do the theories and the techniques used in interpreting the data. It is fully expected that the results of these theoretical investigations will aid in the design of new remote-sensing systems.
The problem of catching fish economically and efficiently was studied using EREP data (except for S193) for the northeastern Gulf of Mexico in cooperation with sport and professional fishermen (ref. 5-18). These fishermen recorded the type, the location, and the time of all catches, and these data were then correlated with  information obtained by Skylab and by high-altitude aircraft that overflew the area simultaneously. The primary purpose of the experiment was to ascertain whether remotely sensed data could be correlated with surface measurements and with the types and numbers of fish that were caught.
Conventional meteorological and oceanographic data obtained during this investigation were ocean depth, wave conditions, distance from the shore, chlorophyll content of the water, sea-surface temperature, salinity, water transparency, water color, atmospheric surface pressure, and air temperature. The water temperatures were sensed remotely by two different aircraft over a portion of the total area that was investigated. Chlorophyll-a was sensed remotely and measured at the surface. The remotely sensed chlorophyll data were obtained by a special spectral radiometer flown on a light aircraft at an altitude of 3000 m. The instrument measured the radiance in the spectral region from 390 to 1100 nm and was calibrated at 57 wavelengths in that range. An important feature in the measured chlorophyll content of the waters near the surface of the
Gulf of Mexico is that it varies by fairly large amounts over relatively small areas. The content cam be as much as four times greater in one place than it is in another just a few kilometers away, and small pockets or regions of zero chlorophyll-a have been observed. The correlation of the values of chlorophyll-a, as measured from an aircraft and as measured from water samples obtained in situ for one flight line, is shown in figure 5-25.
The fish involved were the blue marlin, the white marlin, the sailfish (under the classification of billfish), and the dolphin and wahoo (under the classification of other game fish). During a 2-day period, 67 fish were "raised but not hooked," which probably means that they were sighted and followed the bait for a while but did not take the bait. Of the 171 fish that were hooked (approximately the same number on each day), 58 got away; therefore, 113 fish were caught.
Models based on aircraft data and conventional surface-truth data were developed for predicting the abundance of white marlin in the area. The best model using these techniques had a correlation coefficient of 0.489 and was significant at the 60-percent level. The effect of....
 ....remotely sensed EREP data on the prediction models was next determined. Although cloudiness in the area precluded the use of certain sensor data, the S192 data were processed and the radiance values from bands 1 to 7 and from band 13 were used for the cloudless areas in the original test regions. The information content of bands 4 and 5 was eliminated.
The final prediction model for white marlin for one particular day based on a Skylab pass over the area yielded a correlation coefficient of 0.892 (compared to 0.489 without space data) and was significant at the 90 percent level (as compared to the 60-percent level). These results, though based only on data for I day and on a subset of the total area that was included in the prediction model, were further evaluated using a model developed from the same surface-truth parameters that were used in the previous model, except that only those test areas for which S192 data were available were used. It was concluded, therefore, that the increase in precision of the model could be attributed to the data from the S192 sensor. Other aspects of the problem, such as trying to remove the effects of sunglint by differencing the measurements made in the different bands, were investigated; the effort was unsuccessful.
The application of these techniques to larger areas of the ocean on a routine basis to improve the efficiency of the various U.S. fishing fleets will not be achieved in the near future. However, this experiment demonstrated the feasibility of a technique that may eventually become extremely useful to the fishing industry.
Coastal Water Interpretation
As demonstrated by their reports, the Principal investigators who used the EREP data are pioneers in studying potential techniques for improved understanding of the coastal waters around continents and islands. A typical marine scientist conducts research from small, well-equipped coastal vessels from which he acquires point measurements such as catching fish, netting various kinds of plankton, determining chemical and nutrient contents of the water, and measuring currents. The problem with this approach is that it is not possible to cover a large area in great detail, and difficulty is encountered in relating one set of measurements taken at one time and one place to another set of measurements obtained at a different time and a different place.
A significant aspect of the Skylab missions is that the coastal imagery obtained can be used in many different ways to aid in the interpretation of the conventional data obtained by marine scientists. The methods developed in the EREP Program potentially may aid marine scientists in their studies.
Eight Principal investigators applied photo-interpretative techniques such as stereoscopic viewing, densitometric analysis, and color enhancement to EREP imagery to study many coastal and oceanic features. The areas studied were Chesapeake Bay by Nichols (ref. 5-12) and Gordon and Nichols (ref. 5-13); Delaware Bay by Klemas et al. (ref. 5-22); the Seto Naikai, or inland Sea, and other waters around Japan by Maruyasu et al. (ref. 5-23); New York State water resources (Lake Ontario and Conesus Lake) by Piech et al. (ref. 5-24) (sec. 6 of this report); Sam Francisco Bay and California coastal waters by Pirie and Steller (ref. 5-11); Puerto Rico waters by Trumbull (ref. 5-8); Block Island Sound by Yost (ref. 5-25); and the Gulf Stream by Maul et al. (ref. 5-26).
The southern Chesapeake Bay, including the Rappahannock Estuary (refs. 5-12 and 5-13), is an area rich in oysters and fish and with relatively mild conditions of tide and river inflow. The tides in such an estuary cause marked changes in the amount of suspended materials that are indicated in circulation patterns. For example, when the water from the lower part of Chesapeake Bay enters the Rappahannock, it can concentrate on one side of the estuary and leave the other side untouched. The tidal currents produce small-scale mixing patterns controlled by changes in the shape and the width of the estuary. In this investigation, S190A and S190B data were analyzed with densitometers and four different estuarine water types were mapped that were related to the water transparency, turbidity, and suspended-sediment load. It was possible to locate small-scale mixing patterns caused by local tidal currents.
In the Delaware Bay study (ref. 5-22), surface truth obtained from boats included measurements of Secchi depth, suspended-sediment concentration, transmissivity, temperature, salinity, and water color. In this investigation, emphasis was placed on coastal land use and vegetation mapping as prepared from EREP photographs and S192 scanner imagery. These products were used to map, at a scale of 1:125 000, 10 land use and vegetation categories that included delineation of the wetlands. A particularly valuable....
.....aspect of the imagery was discrimination of small dispersed areas of marshland that are particularly important in the estuarine food chain. With regard to the marine studies, it was possible to monitor suspended sediment concentrations, to map surface-current circulations, to locate boundaries of internal systems, to track surface slicks, to follow ocean-waste dispersion, and to monitor ship traffic.
In studies of data obtained during a pass over Japan, Maruyasu (ref. 5-23) used the different colors in the water on a single S190B photograph to trace the tidal currents in the inland Sea in the vicinity of Kojina Bay. A coinvestigator (Ochiai) analyzed S190B photographs of the Sea of Bingo from the same Skylab pass and higher resolution imagery from an aerial multispectral scanner to map the boundaries of industrial effluents around the coastal industrial zone and to detect oil pollution. An area of red tide was detected in the aerial photograph and in an enlarged portion of an S190B photograph. As described by Ochiai, "The monitoring of the red tide is considered [to be] the most important task for fishery [scientists]." The yellow-colored vortex in figure 5-26 is one red-tide pattern that was sighted in the Sea of Bingo during the observation flight on January 11, 1973; it was also identified in an enlarged S190B photograph (fig. 5-27).
California is an excellent coastal area for demonstrating the potential of use of low-Earth-orbit sensor systems for studying coastal and estuarine processes because of the varied types of features that are encountered. The northern coast is rocky with silt-laden streams and rivers. The southern coast has long, sandy beaches with eroding coastal bluff formations. The streams and rivers of southern California are usually dry during the summer months. In San Pablo Bay, which is in the northern part of San Francisco Bay, sediment transport was traced to areas of known deposition with Skylab imagery and was correlated closely with plots of sediment distribution obtained during the same period by ship surveys (ref. 5-11).
Color-composite enhancements of S192 imagery (bands 4, 6, and 7) provided detailed current and sediment transport patterns. The brightness in the image is proportional to the amount of suspended sediment. The sediment can be seen to flow southward past Alcatraz Island into South Bay. Some of this sediment also flows under the Golden Gate Bridge and can be seen as a boundary far out to sea. An S190A black-and-white....
....photograph of San Francisco Bay illustrates the variation in the sediment load (fig. 5-28).
The patterns of dredged sediment discharges were plotted over a 3-month period. It was found that lithogenous particles, kept in suspension by the freshwater from the combined Sacramento and San Joaquin Rivers, were transported downstream to the estuarine area at varying rates depending on the river-discharge level. To measure the transport in San Pablo Bay, dredged sediments were marked with iridium before discharge near the Carquinez Strait. For May, June, and July 1974, the movements of these tracer sediments were plotted after collection and processing by 82 stations within the bay. This information matched the movements predicted from interpretation of EREP imagery and photographs.
The photographs of San Francisco Bay were taken during a period of exceptionally high freshwater and suspended-sediment discharge. A three-pronged surface sediment pattern is visible where the Sacramento-San Joaquin River enters San Pablo Bay through the Carquinez Strait. The three prongs extended to areas where maximum deposition historically occurs-the central channel, the southeast shore near Pinole Point, and the northwest flats near the Petaluma River mouth. The S19OB color photographs were excellent for spectral and spatial resolution. Spectral analysis of the photographs indicated that the sediment reflection peak was near 0.55 µm. Northwesterly wind was moving surface waters into the southeast bay near Pinole Point.
The S190 photograph and others were processed using narrow-band filters and a densitometer, so as to produce contours of suspended-sediment load as shown in figures 5-29(a) and 5-29(b). A suspensate concentration of approximately 2 mg/liter is quite sufficient to tag a surface-current system and, by using progressively longer wavelength filters, the surface structure of currents with more than 250 mg/liter can be imaged.
Measurements of the suspended-sediment load passing through the Carquinez Strait on January 26, 1974, were made. In the center of the channel, the concentration was approximately 250 mg/liter. A total of approximately 6.3 million metric tons of material passed into the San Francisco Bay during the 1973-74 season. Analysis of the S190A photographs indicates a reflectance shift toward the green from the blue as sediment load increases. This shift explains the excellent detail in the 0.5- to 0.6-µm and in the 0.6- to 0.7-µm bands.
Dredging may be required in the Berkeley Flats area of San Francisco Bay. Use of satellite and aircraft information in this respect will be beneficial, because the sites of shoaling and deposition are detectable. Cost savings using EREP-type data would vary with the placement and the extent of required dredging, but it is possible that savings of several million dollars could result. The techniques outlined in reference 5-11 should be applicable for coastal and estuarine processes studied in other areas of the world.
Excellent EREP photographs were obtained for the study of the coastal processes and waters surrounding Puerto Rico (ref. 5-8). The study of these photographs yielded information concerning many important aspects of the region. The important feature shown in figure 5-30 is the large anomalous blue area, which is also evident in four of the six frames of a simultaneously obtained set of S190A photographs. This intensely dark blue, almost black, area occupies much of the Bahia de Mayaguez (approximately 70....
....km2) and is most strongly developed offshore from the city of Mayaguez, in and downwind of an area of known discharge of oily wastewater from tuna packing plants and other industries. The boundary of this deep-blue area on the south coincides with the location of a reef edge; depths in the lighter colored area to the south are generally in the range of 5 to 8 m, whereas depths in the deepest blue area are approximately 183 m. The intensity of blue in the dark-blue area is clearly not a simple function of depth, because nearby deeper areas are not nearly as blue. It seemed evident to Trumbull that this deep-blue area was caused by some effluent on the water surface that changed the spectral reflectance. The presence of the anomaly was unknown before the EREP study.
Among the conclusions of the study were the following. The synoptic nature of the EREP information permitted the detection and study of phenomena impossible by any other existing technique. The S190B photographs contained incomparable bathymetric detail in which depths to a maximum of 26 m can be seen in areas of clear water. The turbidity of coastal waters near Puerto Rico is commonly high, reducing water penetration severely or eliminating it entirely. However, depth contours could be produced for certain areas.
Trumbull concluded that Skylab-quality data have particularly high potential for studies of bathymetry, patterns of coastal currents, coastal erosion, sediment transportation and accumulation, effects of coastal works of man, and oil-slick detection in the less well developed coastal areas of the world. Coral reefs and areas of coastal erosion were detected from orbital photographs. Potentially economic quantities of offshore deposits of sand, gravel, and mixed sand and gravel were readily detectable on orbital photographs of S190 quality, where bottom reflectance can be seen. Field examinations are required for accurate differentiation assessment of the resources.
Currents can be studied in turbid coastal water. (A disadvantage is that, because of current variability in time and space, frequent coverage is necessary.) Effluent discharges and oil slicks of relatively small dimensions are readily detectable. Large, diffuse discharges, potentially dangerous ecologically, can be detected and studied by means of orbital photographs. A limiting requirement for oil-slick detection seems to be that the slick must not be in the Sun-azimuth direction from the center of the photograph.
Photographs of S190B quality can portray the patterns and boundaries of bottom-dwelling plant and animal communities in clear water. It is, however, necessary to make a field check because these patterns are somewhat difficult and sometimes impossible to distinguish from bathymetric patterns. Although stereoscopic effects can be seen in the Skylab photographs, they are not strong enough to show the apparent relief of bathymetric features. Techniques such as those previously discussed are therefore preferred, as compared to attempting to contour depths from stereophotographs. An important point, however, is that the depth of the water can actually be seen from spacecraft altitudes in stereopsis. Both aircraft and spacecraft photographs can be used to identify problem areas and to guide fieldwork from research vessels in a more intelligent way.
In an investigation of Block island Sound and adjacent New York coastal waters (ref. 5-25), photographic techniques were used that greatly enhanced subtle low-brightness water detail. Photographic contrast-stretching techniques applied to S190A photographs enabled differentiation between two water masses having an extinction coefficient difference of only 0.07. By digitizing S190A multispectral data in registration, a nonhomogeneous vertical stratification of Block island Sound waters with differences in suspended solids of 1 mg/liter was detected. Significant differences between conventional tidal-current charts and the actual patterns of waterflow in Long island Sound were established. One such difference is the existence of two large counterclockwise gyres heretofore undetected. The average extinction coefficient for white light was measured by ship at Block island Sound to be 0.335, with a value of 0.400 for the blue band and of 0.554 for  the red band. These optical water characteristics of nonhomogeneous surface and subsurface water can be charted. Estimates of suspended particles larger than 5 µm can be made.
Both this section and that on fisheries indicate that marine scientists could profitably use imagery systems such as those on the Landsat satellites but with the spectral bands specially selected to provide high spectral resolution and information in the violet, blue-green, and yellow portions of the visible spectrum, plus one red band. For these bands, the effect of the atmosphere is important because the light from the image is scattered more and, hence, is attenuated. However, the many theoretical analyses in these reports and in the study of aerosols show promising techniques to compensate for the intervening atmosphere. The signal-to-noise ratio of the particular bands in the spacecraft sensor may still be a limiting factor. The problem remains of relating a series of images, often obtained many weeks apart, to point measurements made by conventional techniques to relate the time variations at selected points to the image over an extensive area. The images can serve to assess the problem and identify points at which measurements would be needed.
Some of the energy that drives the global atmospheric circulation is provided by evaporation of water from the ocean surface. The Sun is the primary source of evaporative energy. The water vapor is transferred by vertical air currents through the boundary layer, condenses, and produces clouds and latent heat. Thus, the latent heat of condensation is a driving force of atmospheric circulation, particularly at the lower latitudes.
Further advances in man's understanding and ability to predict require a greater knowledge about the radiation transfer through the atmosphere, the sea-surface temperature patterns, the vertical and horizontal wind fields at different altitudes, and the physical thermodynamic characterization of clouds.
In addition to the need for improving knowledge of the physics of solar and terrestrial radiation transfer through the atmosphere, it is of major importance that better techniques be developed to correct for the effects of atmospheric attenuation. The Skylab meteorological program placed major emphasis on the study of radiation transfer.
The atmosphere is composed of dry gases (nitrogen, oxygen (O2), argon, carbon dioxide (CO2), and trace gases), water vapor (H2O), and aerosols. Because incoming solar radiation and outgoing terrestrial radiation interact with these atmospheric constituents, the transmission of the beam is selectively modified as a function of wavelength because of the spectral nature of the interacting mechanisms. Radiation at some wavelengths is absorbed by gas molecules (removal of photons from the beam), whereas radiation at other wavelengths is refracted or scattered by molecules and aerosols. Total atmospheric extinction, therefore, is the result of the combined attenuation due to scattering plus absorption.
Atmospheric transmittance may be defined as , where is the atmospheric optical depth. An optical depth of 1.0 implies am extinction that would occur in an equivalent vertical path through the mass of the clear atmosphere. Figure 5-31 shows the relationship between atmospheric transmission and wavelength for the spectral interval covered by the S192, excluding the thermal infrared. In the visible region of the spectrum is a slight amount of absorption caused by ozone (O3) at the shortest wavelengths. However, the most important attenuation mechanisms are those due to scattering by the gaseous (including water vapor) molecules and aerosols.
Chang and Isaacs (ref. 5-27) measured the attenuated direct solar beam on the Great Salt Lake Desert, Utah, during the Skylab pass on June 5,1973. Figure 5-32 illustrates the spectral modification of incoming solar radiation. The reduced solar intensity that reaches the ground is available for evaporation and warming of the surface, and some may be back-reflected to space by the surface. This back-reflection produces photographs and signals for the S192, and the atmospheric effect has occurred on both the incoming solar beam and the exiting reflected beam.
The absorption and scattering functions for the dry gases in the atmosphere are generally well known (Rayleigh scattering). Extinction by water vapor, liquid water, and ice crystals (cirrus) is not as well understood, even though this factor appears to be of major importance in the heating and cooling of the atmosphere as....
....well as in limiting the accuracy of remotely sensing certain Earth surface features. The extinction of energy is expressed in terms of optical depth.
Aerosol layers may be composed of dry haze, watercoated solid particles, or ice crystals. Sources of aerosols in the atmosphere are numerous and include volcanic eruptions, windborne soil particles, industrial and aircraft pollution, insects, protozoa and other microorganisms, and dust of extraterrestrial origin. The particles usually vary considerably in size, shape, chemical composition, and optical characteristics. In addition, aerosol layers are almost never homogeneous in either the vertical or the horizontal planes. In the atmosphere, am aerosol layer changes the radiative balance and produces both heating and cooling effects. The aerosol backscatter of incoming solar radiation increases the total Earth-atmosphere albedo, whereas the absorption of solar and terrestrial radiation increases the atmospheric temperature (and thus reduces the net radiative cooling). The aerosol number density, particle-size distribution, and location within the atmosphere may be of significant influence on the....
....strength of the atmospheric circulation. Eight Skylab investigations were devoted either primarily or in part to improving understanding of visible and infrared radiation transfer through aerosol layers (refs. 5-6 and 5-27 to 5-33).
Photographic examples of the results of atmospheric attenuation, refraction, and molecular scatter are seen in figures 5-33 and 5-34. In figure 5-33, the full Moon is seen setting beyond the Earth limb. The distortion of the Moon in the bottom frame results from refraction of the backscattered-reflected solar energy from the Moon through the Earth's atmosphere.
Excellent examples of the manner in which the white light of the Sun is refracted into its component colors are shown in figure 5-34, two series of photographs of the Earth limb as the Sun rises above (fig. 5-34(a)) and sets below (fig. 5-34(b)) the horizon. The red color predominates in the more dense portion of the atmosphere because all other colors of shorter wavelengths have been attenuated; that is, scattered or absorbed. As the molecular density decreases with height above the Earth's surface, the remaining color components appear, until all colors are seen as white light. The blue above the white light is due to scattering by the last remnants of the atmospheric molecules.
Within the dispersion of colors, the horizontal layers seen in frames 4 to 7 (fig. 5-34(a)) are indicative of large changes in the index of refraction. The values of these indices of refraction and their heights are significant and need to be evaluated in the radiation-transfer equations. Future limb-analysis investigations are expected to result in more detailed optical parameters through analysis of forward-scatter radiation, using techniques similar to those used by Tingey and Potter (ref. 5-32) with backscattered radiation.
The residence time of aerosols in the stratosphere has been conservatively estimated to be at least 18 months. Because of the concern about possible chemical and thermodynamic changes resulting from the release of aerosols into the stratosphere by man, many efforts have been made over the past decade to determine aerosol distribution. These efforts have included literally hundreds of aircraft, rocket, and balloon flights as well as ground-based searchlight and laser measurements.
Models based on these measurements yield typical aerosol optical depths of 0.5 for aerosols at altitudes above 20 km with probable variations as great as 5 to 10 times this amount. A stratospheric aerosol optical depth....
....of 2.0 has been shown to cause an error of several percent in classification of ground targets (ref. 5-32). Thus, both scientific and practical reasons exist for the study of the stratospheric aerosol layers.
The Skylab spacecraft usually orbited in a solar-inertial mode. During these periods, it was possible to obtain quantitative measurements of the solar backscatter of the Earth limb (fig. 5-35) to use in studying the distribution of aerosols in the higher atmosphere. Tingey and Potter (ref. 5-32) developed techniques to use limb-brightness measurements (by S19OA, S191, and S192) to determine attenuation coefficients of haze layers in the stratosphere. By ratioing the coefficients of attenuation due to aerosols and to dry gases (Rayleigh), they were able to show the location and relative magnitude of haze layers. The results for one S192 pass are shown in figure 5-36; attenuation coefficient peaks associated with haze layers measured by S190A are shown in figure 5-37.
From this work, it was ascertained that several aerosol layers could be identified and that the attenuation coefficients could be evaluated quantitatively. It should be noted, however, that the results of this study were limited by the pointing accuracy of the spacecraft sensor and the absolute calibration of the radiometers. This study unquestionably verified that the limb-brightness technique is useful for evaluating the particulate content of the stratosphere. Aerosol layers were noted at 20-, 40-, 50-, 60-, and 67-km altitudes. Layers at approximately 40, 50, and 55 km appear to be more responsive to longer wavelengths (0.71 µm), whereas layers at 59 and 67 km are more easily detected in the bandpass centered at a wavelength of 0.53 µm. This approach can be used with data from future orbiting platforms to monitor the changes and variations in the stratosphere between 10 and 70 km.
Tropospheric Solar Radiation Attenuation
From the earliest days of aerial photography, it has been known that visible radiation attenuation in the Earth's atmosphere is primarily caused by aerosol scattering. Pioneer work by Lord Rayleigh (ref. 5-34), Mie (ref. 5-35), and, later, Van de Hulst (ref. 5-36) developed general concepts for describing the scattering of light from spheres. (Although it is recognized that most dry aerosols and ice crystals are not spherical, the equations for treating scattering from nonspherical particles are not yet available.) The development of the.....
.....mathematical basis for modeling the transfer of radiation through scattering atmospheres is reviewed in the various EREP reports.
A primary function of solar radiation-transfer models is to correct for atmospheric albedo to obtain surface albedo. For the most part, the existing models are only theoretical constructions and thus have very limited applicability to remote sensing. To develop operational models, one must include a synthesis of radiation-transfer theory, surface-reflectivity characteristics, and an appropriate atmospheric description. The fact that few operational visible-spectrum transfer models have been developed is understandable for the following reasons. First, specialized types of accurate data input that are of uncertain availability are required. Second, the high costs associated with model computation time and ground-truth field measurements are often disproportionate to the resources available. Third,  many transfer models for atmospheric correction are not sufficiently accurate to warrant the atmospheric attenuation calculation. In the last case particularly, the discrepancy between theory and measurement results from a lack of understanding of the physical complexities of multiple scattering, the effects of variations in chemical and physical properties of the haze layers, and the lack of knowledge of the distribution of the haze layers in space and time.
Skylab provided a unique opportunity to make accurate comparisons of the results produced by the scattering models to high-quality measurements. The Skylab Concentrated Atmospheric Radiation Project (SCARP) (Kuhn et al., ref. 5-30) included a major field program designed especially to obtain field and aircraft data that could be used to test and compare visible radiation transfer models through wet, dry, clean, and dirty atmospheres. A distinguishing feature of the study was the use of an aircraft to obtain atmospheric aerosol data. In addition, spectral measurements were made of the surface albedo, and direct measurements of aerosol optical depth were also obtained. The unique set of data collected for this study was considered representative of a large range of varying Earth atmospheres and surface targets with Skylab, aircraft, and ground-based observations collected for a maritime, humid atmosphere near Houston, Texas; a continental, hot, dry atmosphere near Phoenix, Arizona; and a combination dry-atmosphere and low-Sum-angle condition at White Sands, New Mexico.
From this study, it was concluded that single-scattering models are most sensitive to aerosol-refractive-index input and that aerosol optical depth is a critical input for the more refined multiple-scattering models. Of particular interest and practical importance was the evidence that more refined visible-radiation-transfer models are not improved by use of aerosol measurements from aircraft. Only measurements of aerosol optical depth made at the surface are needed to optimize model accuracy. Furthermore, by means of the SCARP measurements, techniques were developed to invert mathematically the surface optical-depth measurements to obtain the aerosol-size-distribution function. These two findings are considered of major importance for future remote-sensing activities.
Numerous Skylab investigations in fields varying from mineral exploration to oceanography included atmospheric corrections to provide more accurate information on true surface albedo (e.g., refs. 5-26 and 5-31). These investigations, although interesting from the standpoint of radiation transfer through the atmosphere for different air masses and slant ranges, were primarily directed toward improving the measurement accuracy of remotely sensed surface phenomena.
Tropospheric infrared Radiation Attenuation
In the previous subsection, radiation originating from the Sun, which has a temperature near 6000 K, was discussed. Radiation originating from sources having temperatures of usually less than 320 K is discussed herein.
On a worldwide basis, an average of 27 percent of the direct solar radiation and 20 percent of the energy reflected downward by or conducted from the atmosphere (or a total of 47 percent of the solar energy that reaches the top of the atmosphere) is absorbed by the Earth. The energy that is not used for evaporation and photosynthesis heats the surface and subsequently is reradiated at infrared wavelengths from the surface. In some wavelengths, it is absorbed by certain constituents of the atmosphere, primarily carbon dioxide, ozone, and water vapor, and thus heats the atmosphere. The combined effects of geographical variations in surface temperature and atmospheric absorption drive the "atmospheric engine." Measurements of the radiation budget components from satellites provide data on the energy budget of the Earth. These data comprise a primary information source for numerical weather forecast models. Hence, improvement in understanding the physics of infrared terrestrial radiation transfer through the atmosphere would directly contribute to improved numerical models for global weather forecasting.
A small part of the outgoing infrared radiation originates at the surface and in the clouds and escapes through the atmosphere through radiation "windows," which are the spectral regions between the absorption bands of the gases. Information on various aspects of....
...the atmosphere and its behavior can be obtained by measuring the upflux of terrestrial radiation at these selected wavelengths. Of particular importance in the measurement of cloud-top and Earth surface temperatures is the atmospheric window in the 8- to 12-µm region of the infrared spectrum. Because atmospheric gases cause little attenuation in this spectral band (fig. 5-38) and because the terrestrial radiation peaks in this region, most of the remote-sensing measurements of surface temperatures for studies in geology, agriculture, oceanography, etc. have been made in this spectral interval. Unfortunately, even in the early days of satellite meteorology, it was noted that at no place in the infrared spectrum was the atmosphere completely transparent. Even in the 10- to 11-µm region, a small amount of outgoing terrestrial radiation is absorbed by the water vapor in the atmosphere. In addition, a significant but usually unknown amount of attenuation occurs as the radiation passes through haze layers and invisible cirrus cloud layers (ref. 5-37). As a consequence, many corrections must be made before radiation measurements from Earth-orbital satellites can be translated into meteorologically meaningful parameters.
During the past decade, numerous theoretical and empirical models have been developed to permit rigorous analytical derivation of Earth surface and meteorological data from infrared-sensor measurements. Most models of infrared radiation transfer developed for remote-sensing applications differ significantly from one another in the manner in which the atmospheric transmissivity for the various gases and particulates is included in the radiation-transfer equations.
Determining the accuracy and utility of these models and investigating their application to the estimation of cloud-top, land, and sea-surface temperatures were major efforts of the EREP investigations Program (refs. 5-8, 5-29, 5-30, 5-33, and 5-38).
One of the primary objectives of SCARP was to obtain field measurements of the critical parameters used in evaluating these models. Surface, aircraft, and balloon-borne sensors obtained pertinent data under the Skylab spacecraft at targets in southeastern Texas and in the Gulf of Mexico (maritime, moist air masses) and at White Sands, New Mexico, and Phoenix, Arizona (continental, dry air masses). These data, together with the Skylab S191 measurements, afforded better insight into the mechanisms of infrared radiation transfer through the atmosphere, as well as a statistical comparison of the accuracy of the transfer models in current use (figs. 5-39 and 5-40). The SCARP investigation showed that, when sufficient information is available, the present numerical modeling is adequate for predicting the transfer of infrared radiation through the atmosphere with an accuracy of approximately 1 K.
For most studies, however, ancillary data from surface, aircraft, or balloon measurements are not available. To eliminate this requirement, Anding and Walker (ref. 5-6) developed a technique based on the use of the differential optical properties of the atmosphere in the infrared-window region to infer the atmospheric attenuation. They then used the attenuation values to correct for the effect of the absorption of atmospheric gases on radiometric data. As shown in figure 5-41, they demonstrated that this method of calculating the spectral radiance arriving at the top of a maritime atmosphere compared quite favorably with the S191 measurements. The agreement is excellent in the bandpasses from approximately 11 to 13 µm. For this evaluation, a 100-percent-maritime (wet) aerosol with 23-km sea-level visibility was assumed. To test the application of this method of analysis, the in-band brightness temperatures were computed and plotted against relative extinction coefficients (fig. 5-10).
In addition to the size, shape, and nature of the aerosol particle, the effect of the degree of wetness must be considered. In the Anding-Walker study and in the SCARP maritime study, the aerosol particles were assumed to be water coated. The White Sands and Phoenix aerosols were assumed to be composed of dry quartz. Examples of transmissivity of aerosol layers based on these assumptions are shown in figure 5-42.
The optical depth of the aerosol layers was measured by aircraft instruments for SCARP. When actual measurements were not available, assumptions of optical depth were based on horizontal visibility by Anding and Walker (ref. 5-6), Maul et al. (ref. 5-26), Turner (ref. 5-33), and others. The aerosol problem is further complicated because these layers are tenuous; that is, they change almost continuously in composition, size distribution, and location within the atmosphere. Although the Skylab EREP investigations provided helpful new data, the problem of aerosol attenuation of both visible and infrared radiation has not been solved.
Atmospheric Water, Clouds, and Precipitation
Water, because of its peculiar radiation-absorption characteristics, its changes of state within the normal ranges of atmospheric temperatures, its capability to absorb much of the heat radiated from the Earth's surface, and its large gradients of concentration, both horizontally and vertically in the atmosphere, is an important factor in the energy budget of the atmosphere and a dominant factor in the production of weather events. Earth-orbital satellites have provided major new contributions to knowledge of atmospheric water and have made possible the global assessment of the quantities, the forms, and the distribution patterns of atmospheric moisture as functions of time and space.
Clouds are of major meteorological significance for several reasons.
Much information about the state of the atmosphere can be derived simply by noting the changing state of the clouds. Nephanalysis, the science of weather analysis from cloud-pattern studies, provides clear indications of atmospheric pressure, temperature, and air circulation, as well as the distribution of moisture itself Many meteorological characteristics of an airmass may be inferred from satellite-based observations of the clouds associated with it. Sheetlike clouds are usually indicative of slow, widespread lifting of moist air. This type of cloud is often associated with orographic lifting on the windward side of mountain ranges, in the frontal zones between contrasting airmasses, and in areas where the water vapor in warm air passes over cool surfaces, condenses, and produces low stratus or fog. Tower clouds, on the other hand, are formed in areas of marked atmospheric instability with strong vertical air currents.
It is often possible, then, to assess the extent of cloud cover, the type of cloudiness, and the degree of vertical motion from nephanalysis of satellite measurements of clouds. The thickness of unstable layers can be estimated from cloud-top temperatures. The height of cloud bases can be calculated from the geometry of the cloud and from cloud shadow locations on the satellite photograph. Temperature inversions may be inferred by noting where the flattening of cloud tops occurs; such flattening is indicative of the subsidence of warmer, drier air aloft. Phase-change level may be determined by noting the change in cloud outlines  where the sharply defined water clouds give way to the diffuse ice crystal layers.
In addition, it is often possible to identify areas of strong convection from shadows of tall cumulonimbus towers (particularly under low-Sun-angle conditions). Jet-stream cirrus often casts shadows on lower cloud decks, and thin, high cirrus may be identified by its influence on the image of lower clouds. Satellite photographs of cloud patterns show that the cloud elements may be either randomly distributed or organized into some regular formation. Such formations are normally associated with one or more atmospheric and/or topographic factors.
Because other satellite programs are dedicated to obtaining detailed information on the state of the atmosphere, the number of meteorological experiments on Skylab was limited. The capability of man to take high oblique photographs and stereophotographs is unique, and the information gained by the Skylab crewmen's photographic documentation demonstrates forcefully the value of manned space programs. The Skylab 4 crew acquired stereophotographs of numerous classical and unique cloud patterns, including thunderstorms, tropical and extratropical cyclones, mountain wave clouds, convection in cold air passing over warm water, jet-stream cirrus, island vortex and convection effects, sea breezes, cloud streets, and subsynoptic-scale atmospheric circulations.
The cloud-street orientation seen in figure 5-14(a) is an example of the manner in which cloud rows are used as an indicator of low-level windflow in areas for which radiosonde and surface meteorological observations are not available. Pitts et al. (ref. 5-28) conducted a field program to obtain hourly radiosonde soundings concurrent with photographic data for a cloud-street pattern over Fort Sill, Oklahoma, in June 1973. They learned that the cloud-street orientation conformed to the wind vector at the base of a temperature inversion (cloud-top height) in this case.
The remote measurement of air movement at different altitudes has been attempted by means of various techniques ranging from measurements of cloud-street orientation to the satellite monitoring of the movements of constant-pressure balloons.
In another experiment, Villevieille and Weiller (ref. 5-39) related vertical-wind profiles with cloud-street parameters using satellite photographs. This study presents the theoretical development and algorithms of techniques for calculating important atmospheric stability and wind-shear parameters. Further discussion of the methodology of vertical-wind-profile calculations is included in section 6.
In earlier studies, Kuettner (ref. 5-40) and LeMone (ref. 5-41) reported that the range of the ratio of the spacing between the horizontal streets to the height of the temperature inversion was 2 to 4. The EREP data indicated that this spacing ratio was 1.7 for the Fort Sill, Oklahoma, study, which was slightly less than the ratio range found in the earlier studies.
Investigations to describe cloud physical structures using cloud radiance measurements met with varied results. Curran et al. (ref. 5-29) attempted to compare the cloud-top altitudes measured by using 11-µm thermal-infrared (S192 channel 13) radiation with those....
....using stereoscopic techniques on S190 photographs. A qualitative comparison of cloud-top temperatures with the water phase of the cloud top is shown in figure 5-43 and indicates that future studies of this type can provide useful meteorological information.
Skylab EREP investigations confirmed that infrared sensors on satellites tend to either underestimate or overestimate cloud-top altitudes based on their blackbody temperatures. High clouds (15 000 m) were underestimated by approximately 1000m and low clouds (3500 to 7000 m) were overestimated. These errors were to be expected unless the measurements were corrected for gaseous and aerosol attenuation, because even small errors in temperature result in quite large errors in altitude estimates.
Curran et al. (ref. 5-29) found that the ratio of cloud reflectance at wavelength 1.61 µm to that at wavelength 0.754 µm as a function of the cloud optical thickness at 0.754 µm can be used to distinguish between clouds composed of ice crystals and those composed of liquid droplets (fig. 5-44). When applied to the multispectral scanner, appropriate channel data enabled determination of the thermodynamic phase of the cloud tops. Alishouse et al. (ref. 5-38) and Pitts et al. (ref. 5-28) also ratioed reflectances in the narrow bands in the visible and near infrared to distinguish between ice crystal and water droplet clouds.
Skylab EREP experiments provided some evidence that discrimination is possible among cloud ice crystals, cloud water droplets, and surface snow. However, the need exists to extend discrimination to include supercooled water droplets, mixes of ice crystals and water droplets, ice crystal stabilization (i.e., agitated in cumulus top as opposed to tropopause stratification), and mixes of surface snow crystal structures.
Because of its spectral reflective and thermal characteristics, snow cover greatly affects both the energy budget at the surface and the regional water balance. New advances in snow-cover mapping using Skylab sensors are discussed in sections 4 and 6.
Although the determination of soil moisture content is normally considered in the realm of agriculture, the meteorologist is interested in spatial and temporal variations of soil moisture, particularly in the first few....
 ....centimeters below the surface, because the moisture content of the surface soil strongly influences soil thermal properties and evapotranspiration rates. Because water has a greater specific heat than does mineral soil, for a given heat input, moist surface soils with be cooler than dry soils during the day.
Marwitz (ref. 5-42), Davies-Jones (ref. 5-43), and Sasaki (ref. 5-44) have shown that the inflow air source for severe thunderstorms is from the near-surface layer of the atmosphere. Heat and moisture from the soil transferred through this inflow air provide additional energy to the storm system. Beebe (ref. 5-45) reported that the tornado frequency maximums in the Texas Panhandle were centered in a region of extensive irrigation. He concluded that the increased water vapor supplied to these tornado cloud systems was a result of evapotranspiration from the irrigated fields.
In theory at least, the use of microwave frequencies is a direct approach to the measurement of soil moisture. Water has a very high dielectric constant; soils have very low constants. Moist soils therefore have a dielectric constant that is proportional to the relative amounts of water, soil, and air present.
The influences of soil type, surface roughness, and vegetative cover on microwave emission are all wavelength dependent with the strongest effects at the shorter wavelengths (ref. 5-28). A significant advantage of the longer wavelengths is that measurements are not restricted to cloudless skies. At L-band (approximately 21 cm) wavelengths, the atmospheric transmission is close to unity with little influence from clouds or gaseous absorbers.
Several studies were conducted to evaluate the microwave L-Band Radiometer (S194) for soil moisture determination. These results are described in section 6. Pitts et al. (ref. 5-28) compared L-band measurements with an index of antecedent precipitation for two Skylab passes across Oklahoma, New Mexico, and Texas. The antecedent precipitation index (API) is a simple method of characterizing the precipitation history in which
where Pi is the daily precipitation for each day from n days previous to the current day i and K characterizes the loss of moisture from the soil due to evapotranspiration and deep percolation and is a function of soil type, slope, season, and vegetation.
The antecedent precipitation index is compared with S194 brightness temperature for a Skylab pass across southwestern Oklahoma and northeastern Texas in figure 5-45. From the studies, it is concluded that the L-band of the microwave is well suited for remote sensing of synoptic soil moisture over large areas under a wide variety of weather, vegetation, and terrain conditions.
Measurement of Sea-Surface Winds
The major purpose of the radiometer-scatterometer (ref. 5-46) experiment was to obtain simultaneous measurements of radar backscatter and passive microwave temperatures to demonstrate that the passive microwave temperatures could be used to correct for atmospheric attenuation and that the backscatter measurements, after correction, could be used to determine windspeed and wind direction.
The winds over the ocean surface are very difficult to measure. The windspeeds increase with height, and the rate of increase depends on the difference in temperature between the water and the air. Moreover, the.....
....winds are turbulent and fluctuate about an average value in both speed and direction. For improved numerical weather prediction methods, the properly averaged winds need to be measured on an ocean-wide scale and on a uniform grid of points. Data obtained from the S193 provided a scientific breakthrough in the field of the meteorology of ocean wind fields.
The winds generate waves of all lengths simultaneously on the ocean surface. These wavelengths vary from 0.6 cm to more than 600 m, with the highest waves traveling in the wind direction. Some waves travel in directions that are ±90° relative to the wind direction. As the windspeed increases, waves of all lengths grow in height.
That the height of high-frequency capillary waves increases with an increase in windspeed has been shown by measurements made by Mitsuyasu and Honda (ref. 5-47) in a wind-water tunnel. This experiment provides high-frequency-spectrum data to support the theory that capillary wave structure is a dominant factor in radar backscatter, and the data show a power-law windspeed dependence. Figure 5-46 shows that the spectrum of the waves grows with windspeed in the 5 to 30-Hz frequency range when observed as a function of time at a point. The winds in the tunnel for the five curves shown had nominal velocity values of 5, 7.5, 10,12.5, and 15 m/sec; the curve for the 15-m/sec windspeed corresponded to 33-m/sec winds at an elevation of 10 m above the sea surface.
The growth in height of intermediate-length waves and the increasing roughness of the sea with increasing windspeed are strikingly illustrated by a series of photographs taken from the weather ship Papa while stationed in the North Pacific (ref. 5-48). The sea surface becomes increasingly rough as the speed of the wind, as measured just above the surface, increases. The S193 measured this increase in roughness and, at incident angles of 50°, 43°, and 32°, the radar backscatter that was measured for a given relative wind direction increased with windspeed. The measured radar backscatter was correlated both theoretically with sea-surface roughness and winds (ref. 5-49) and directly with the winds by means of multiple-regression techniques (ref. 5-49). Sea roughness is dependent on both wind direction and windspeed. The Skylab results demonstrate that, if the wind direction is known from an independent source, the windspeed can be determined from the backscatter measurement. The Advanced Applications Flight Experiments (AAFE) Langley Radscat Program showed that, as the windspeed increases by a factor of 5, the backscatter increases by more than a factor of 10 (fig. 5-47).
Fung and Chan (in ref. 5-49) succeeded in using the spectral form for the capillary waves shown in figure 5-46 and the available information on the slopes of the longer waves (perhaps 10 m long and longer) to derive a composite theory for backscatter. The large-scale wavy surface was tilted back and forth (in theory), and the backscatter from the small waves, which were approximately 2 cm long, was calculated for the different slopes and summed over the slopes. A typical result is shown in figure 5-48, where the parameter ao describes the angular spread of the capillary wave spectrum.
Three methods were used to determine the wind-field parameters which were correlated with the S193 data. For Hurricane Ava and tropical storm Christine, one of the methods, used to compute the wind-field characteristics for a 160 000 km2 area, was developed for a program independent of Skylab. The model required input information regarding the speed of movement of the storm over the ocean surface, the central pressure, and the radius of maximum winds. Each cell scanned during the Skylab pass over the cyclone was assigned an appropriate wind direction and speed based on this theory.
A synoptic analysis for tropical storm Christine based on conventional shipboard data is shown in figure 5-49, in which the rectangular area indicates the S193 scanned width. Conventional data provide very little information on the characteristics of the wind field in a tropical storm because mariners avoid these areas.
Although the winds in Ava and Christine were determined from a theoretical boundary-layer model for an intense moving vortex, the theory was cross checked against data on winds and other parameters obtained by aircraft flights into both storms. For Hurricane Ava, an NOAA C-130 aircraft penetrated to the eye and measured winds near the sea surface at an elevation of approximately 150 m and at other elevations within the storm at the eyewall. These data, together with central pressure determined from dropsondes into the eye, and glitter patterns in the periphery of the storm were used to check the reasonableness of the final model wind field. An example of one check is shown in figure 5-50, in which the winds measured by the aircraft are correlated with the theoretical winds for the appropriate sector of the tropical hurricane (ref. 5-50).
Figure 5-51 (tropical storm Christine) provides a graphic representation of the meteorologically determined vector wind and the measured backscatter value before correction for attenuation and without any concern with the variability of the backscatter value with wind direction. The arrows that appear to lie in the horizontal plane (with some perspective) are the values of the windspeeds and wind directions, as shown by the appropriate scales, at each of the cells. The spacecraft passed over the bottom most row of cells, which correspond to the nadir measurements. For incidence angles of 31°, 42°, and 50°, the values of the vertically polarized and vertically received backscatter are graphed as vertical bars at each cell. The exact range of variation as a function of windspeed depends on the incidence angle, so these variations in backscatter should be studied along each line at a given incidence angle. For the 31° plot, the backscatter for the low winds in the lower left is approximately-18 dB, increases to greater than-5 dB for one of the cells near the eye, and then decreases with decreasing windspeed to the upper right point. Similar remarks are appropriate for each of the lines scanned for the other two incidence angles. This figure is considered highly significant in that it constitutes the most convincing demonstration of the evident correlation between the winds near the surface of the ocean and the measured value of the radar backscatter.
A second method used to determine the winds in the S193-scanned area was to plot data from all available shipping and coastal reports and perform a streamline isotach analysis of the wind field (ref. 5-49). Streamlines are curved lines everywhere parallel to the reported wind direction, and isotachs are contours of.....
...equal windspeed. For the subtropical areas of the world, this analysis was performed for numerous Skylab 2 and 3 passes. The windspeed and the wind direction at each of the cells scanned were then read from the streamline isotach analysis.
The third method of data analysis was a computer technique (ref. 5-49) developed for determining wind in middle-latitude extratropical cyclones and in other changing pressure systems. These systems produce the most important day-to-day wind fields over the ocean. In this analytic technique, the quality of the various reports in the area is weighed by source, depending on whether the reports are made by weather ships, ships with anemometers of known height, ships with anemometers of unknown height, or transient ships that estimate the winds from wave appearance. In some....
....areas, no reports were available and the winds had to be calculated from the pressure gradients. For many of the cells scanned by the S193 during Skylab 2 and 3, the winds were determined in this way; this technique was used for all the cells scanned during Skylab 4. The procedure corrects for the wind variation with height above the sea surface and refers all winds to an elevation of 19.5 m. It also corrects for atmospheric instability, which produces an additional variation. In effect, the wind with which the backscatter measurement was compared was the wind that would have produced the same wind stress on the sea surface for a neutrally stratified atmosphere.
The scale of the wind fields in extratropical cyclones is illustrated by an analysis performed by Ross (ref. 5-50) as shown in figure 5-52. This cyclone was one of the most intense of the past 20 years. The isobaric pattern and the windspeeds reported by ships near the time of a Skylab pass are shown. The scales involved can be inferred by comparison with the size of the Gulf of Saint Lawrence, which can be seen in the upper left.
The final result of these analyses was provided in an appendix to reference 5-49, in which the measured backscatter values, the passive microwave temperatures, the windspeed and wind direction, the latitude and longitude of the cell, and the time of the observation are tabulated. Table 5-11 is taken from reference 5-49. The S193 azimuth is the direction toward which the radar beam is pointing. The aspect angle is the direction of the wind vector relative to the pointing direction of the radar beam. For zero degrees, the beam is pointing upwind; plus is clockwise and minus is counterclockwise.
Correction of the backscatter measurements for the effects of attenuation using the passive microwave measurements was accomplished by using the brightness temperature at an incidence angle of 50°. Differences between the temperatures that were measured and the temperatures that would have been measured (as determined by the sea-surface temperature in the absence of attenuating effects) were calculated. These differences can be shown to be caused by the intervening cloud droplets and rain between the spacecraft and the sea surface. This excess microwave temperature was correlated with attenuation based on independent measurements of the atmospheric temperatures and corresponding attenuations made with a number of ground-based upward-looking passive microwave receivers. This correlation determined the two-way attenuation in decibels for the scatterometer.
For Skylab 2 and 3, a histogram of the attenuation as calculated at each 50° incidence angle is shown in figure 5-53. Because the backscatter values range over 20 dB or more as the windspeeds vary from 3 to 25 m/sec, a 0.2 or 0.3-dB correction is quite small and barely affects the calculation of the windspeed. Attenuation could not be computed for the Skylab 4 data because the antenna was damaged during the repair of the scanning subsystem. However, based on Skylab 2 and 3 results, the inability to correct the attenuation could not have appreciably affected the calculation of the winds.
For certain cells scanned by the S193, the excess microwave temperatures were extremely high compared to the usual values of the excess microwave temperature just described. These "hot" spots are indicative of large cloud droplets and falling rain. In this situation, the measured backscatter cannot be used to calculate windspeeds because most of it is coming from the clouds and rain. Figure 5-54 shows the marked increases in passive microwave temperature on one of the lines of cells of constant incidence angle over tropical storm Christine. Even for heavy cloud cover in a tropical storm, the attenuation values are not prohibitively large except for the scans in which very sharp peaks.....
....have occurred. The five small R's in figure 5-51 show instances in which the winds could not be calculated from the backscatter in tropical storm Christine because of this effect.
Many different methods were used to study the relationship between the measured backscatter values and the windspeed, all based on the assumption that the meteorologically determined wind direction was correct. All the techniques were variations of multiple-regression schemes. A functional form for the dependence of windspeed on backscatter was used. The various unknown constants in that functional form were determined by minimizing the sums of the squares of the differences between the windspeeds that would be predicted from the radar measurements by means of this analytical form and the meteorological winds. The...
...results of the analysis of the paired sets of meteorological winds and radar winds were then graphed.
The graph for tropical storm Christine and Hurricane Ava, for all three nadir angles and for cross-polarized radar backscatter, is shown in figure 5-55. The distance of the plotted points from the true value can be caused either by errors in the radar part of the theory or by errors in the determination of the meteorological wind. It was therefore very important to obtain an independent estimate of the errors in the wind, and this estimate was accomplished only for the objective synoptic-scale analyses. It was found that the meteorologically determined surface-truth wind has substantial error.
The winds over the ocean are determined from analysis of ship reports of windspeed, wind direction, surface atmospheric pressure, air temperature, and sea temperature. Some reports are made from weather ships that remain in one place to make scheduled observations like those of a weather station on land. In general, these reports, made by trained meteorological....
....personnel, are the most accurate. Transient ships provide reports of variable quality depending on whether or not they have anemometers and whether or not the height of the anemometer has been reported. If the ship has no anemometer, the windspeed and wind direction are estimated. Moreover, the reports are from ships that are scattered unevenly over the oceans, being concentrated along the shipping lanes and widely spaced otherwise.
The analysis of the ship reports involves the boundary-layer theory of Cardone (ref. 5-51), in which relationships among the pressure gradients, the air/sea-temperature differences, and the winds are used to form a  continuous field of the vector winds at a fixed height above the sea surface. The accuracy of such an analysis in determining the actual winds depends on the quality and spacing of the available ship reports.
To determine the accuracy of the existing conventional methods for determining the winds near the surface of the ocean, the data were analyzed separately for the entire period by means of a "withheld weather ship" technique. First, the wind fields were analyzed as just discussed with the weather ship data incorporated. Then the weather ship data were removed and the analysis was repeated. The differences between the two resultant winds at points near the weather ship were quite large and depended on the quality of the data that replaced the weather ship data for the area in question. Errors in the specification of the winds were larger than those for a weather ship report when the winds were reported by an ordinary transient ship and still larger if the wind had to be determined from the isobaric pattern and the available boundary-layer theory.
It was possible to partition the total mean-square difference between the radar winds and the meteorological winds into a contribution from the errors in the meteorologically specified winds and a contribution from the radar specified winds. The regression equations yielded a windspeed predicted from the radar backscatter measurement, given the wind direction, so that for the three highest nadir angles, pairs of values of the radar windspeed Ur and the meteorological windspeed Um were the result. The total variance of these quantities for N samples given by
where all samples from 1 to more than 800 were addressed, is a measure of the variation between two different ways of determining the wind. Both Ur and Um contain errors and differ from the true but unknown value of the windspeed UT, with the error difference given by
Because the total variance is known, it is possible to compute the error variance of the meteorological winds and the error variance of the radar winds that constitute the total variance.
The term for the meteorological variances was determined by the withheld weather ship technique. The term for the radar variance was determined by comparing different polarizations. The results are provided in table 5-III, as stratified according to the quality of the meteorological surface-truth and windspeed ranges. More than 800 separate cells were scanned by the S193 for these studies during 14 Z-axis-to-local-vertical passes. This amount of data far exceeds that of all previous aircraft programs. The total variance for a large sample should equal the sum of the two parts. It does not because of sampling variability; however, the sums nearly balance. For some categories, the two terms on the right add up to less than the term on the left. The difference is shown under the unexplained variance. The italicized values represent those categories where the two terms on the right exceed the term on the left. Although the errors in the meteorological wind were determined from a different data base, the results nearly balance category by category. Most of the difference between the radar wind and the meteorological wind is due to "errors" in the meteorological wind.
Cardone et al. (ref. 5-49) concluded that the windspeeds computed from the backscatter measurements at each cell, after correction for attenuation and under the assumption that the wind direction was correct, were at least as accurate as those that would have been recorded by a weather ship located at (or near) each cell. It was also stated that this conclusion was a conservative interpretation of the results of the study. The standard deviation of the errors of the radar-measured wind may well be less than half that of the errors in the winds presently reported by weather ships.
The radar backscatter theories developed in this program and the AAFE data were used as a guide in formulating the regression equations that were used. These regression equations with unknown constants, determined from Skylab backscatter measurements and the winds, need not necessarily have agreed with either the theory or the AAFE data. However, they did. The results of one regression method in which radar backscatter is plotted against azimuth angle for three different windspeeds are shown in figure 5-56. The agreement with figure 5-47 as to shape and relative separation is very good. The disagreement as to absolute level can be attributed to S193 recalibration problems during Skylab 4.
All Skylab EREP instruments were found to be valuable in studying the oceans. Sea-surface temperatures were measured from orbit by the Infrared Spectrometer. Atmospheric effects on the measurements were corrected to an accuracy of ± 1 K by analysis of data in two selected wavelengths sensed by the instrument. The Multispectral Scanner acquired data that were used to portray thermal patterns of ocean currents and upwellings and to determine depth of clear water to 18 m. The Microwave Radiometer/Scatterometer obtained surface roughness data that verified theories and techniques that will be used in the future to observe sea ice, sea state, and winds over the oceans on a global scale not possible by any other method. Investigators used Microwave Altimeter data
....to demonstrate that the contour of the ocean surface can be measured to an accuracy of ± 1 m or better. Information was obtained in passive microwave data that could be used to determine surface roughness and precipitation in the atmosphere.
The L-Band Radiometer results indicate that sea-surface roughness measurements as related to windspeed may be obtained at L-band frequencies when other frequencies are not usable because of precipitation interference. The L-band data were not suitable for measuring sea-surface salinities or temperatures on the open ocean.
The photographs from the Multispectral Photographic and Earth Terrain Cameras were used to dis-....
...-cern and record visible phenomena such as water color, turbidity, depth, current and wave patterns, shoal extent and location, chlorophyll content, and ice types and patterns.
The availability of data acquired simultaneously from two or more instruments provided bases for conclusions that would not be supportable if only one instrument had been used. Study of living marine resources is an example. Certain temperatures are favorable to growth of plant and animal life but locating areas of appropriate temperatures yields inconclusive results in any search for living organisms. If in addition to data on suitable temperature, water-color information is available from a scanner or a camera, the search area can be reduced. If supplementary information on currents, measured by scanners, cameras, or altimeters, is provided, the results can be further improved.
Subtle gradations of contour, texture, color, and temperature, which are impossible to obtain when measured at or near the surface, can be observed from space and interrelationships can be evaluated. Several investigators used Skylab EREP data containing such information to study ocean currents, which involve most, if not all, of these factors. The superiority of the data acquired from orbit was demonstrated for many applications.
The Skylab EREP investigations resulted in significant advances in the understanding of the physics of the atmosphere and the interaction between the atmosphere and the land and ocean surfaces. Without doubt, the most significant development in satellite meteorology resulting from Skylab was the use of the microwave spectrum to measure the surface wind over the ocean. The possibilities for obtaining markedly improved wind information across areas of the open ocean will be of major usefulness not only to weather forecasting but also to shipping.
Analytic techniques were developed for and new information was gained on the location and concentration of particle layers in the stratosphere. Skylab provided, for the first time, measurements for studying the spectral transfer of visible and thermal radiation through aerosol layers in the troposphere and permitted realistic evaluation of methods for correcting the effects of atmospheric attenuation for remote sensing of the Earth's surface.
The Skylab EREP photographs provided information for studying the phenomenon of cloud streets- the orientation and spacing of cloud bands as a function of the horizontal wind field. New knowledge was gained and mathematical algorithms were developed to describe the relationship of physical cloud parameters to the vertical wind field.
The Skylab experiments developed techniques for discriminating between cloud ice crystals, cloud water droplets, and surface snow. These experiments will lead to the experimental designs for programs during the Space Shuttle flights to solve the problem of discriminating supercooled water droplets, mixtures of ice crystals and water droplets, and the like.
Several EREP investigations were conducted to ascertain the usefulness of microwave radiometers for soil moisture determination. Although the state of the art in longer wavelength microwave radiometry permitted only synoptic-scale measurements with an instantaneous field of view of 100 km and more in diameter, it was learned that the L-band is well suited for monitoring surface soil moisture under a wide variety of weather, vegetation, and terrain conditions.
Oceans and Atmosphere
The EREP experiments demonstrated that the oceans and the atmosphere over the oceans must be studied jointly and simultaneously. The measurement of the winds over the ocean depends on the properties  of the waves generated by the winds and on ocean temperature measurements. Measurements by means of either infrared or passive microwave instruments have to be corrected for the effects of the atmosphere.
On the basis of Skylab-derived knowledge, improved versions of the S191, S192, S193, and S194 instruments will continue to be built for use on unmanned spacecraft. Multispectral scanners are already onboard the Landsat; an improved altimeter is onboard GEOS-3, and an even more accurate model will be used on Seasat-A. Passive microwave sensors will be included on Nimbus and Seasat-A; dual infrared bands, on Tiros. Radar backscatter will be measured on Seasat-A. The combined study of the oceans and atmosphere will be possible with these new instruments to an extent never before possible.
The Skylab experiment also demonstrated the value of having the spacecraft manned by trained crewmen. The crew's ability to acquire and track a target with the S191 instrument proved useful in acquiring spectral information at a variety of viewing angles. Hand-held camera photography supplemented the EREP data for the heating ice experiment. The experience gained in the combined uses of trained crewmen with complex, new, unproven instruments can be the basis for operational plans for using the Space Shuttle in further scientific study of the Earth.
5-1. Vonbun, F. O.; McGoogan, J.; Marsh, J.; and Lerch, F. J.: Sea Surface Determination From Space: The GSFC Geoid. NASA TM X-70959, 1975.
5-2. Mourad, A. G.; Gopalapillai, S.; Kuhner, M.; and Fubara, D. M.: The Application of Skylab Altimetry to Marine Geoid Determination. NASA CR-144372, 1975.
5-3. McGoogan, J. T.; Leitao, C. D.; and Wells, W. T.: Summary of Skylab S-193 Altimeter Altitude Results. NASA TMX 69355, 1975.
5-4. Greenwood, J. Arthur; Nathan, Alan; et al.: Radar Altimetry From a Spacecraft and its Potential Applications to Geodesy. Remote Sensing Environ., vol. 1, no. 1, Mar. 1969, pp. 59-70.
5-5. Greenwood, J. Arthur; Nathan, Alan; et al.: Oceanographic Applications of Radar Altimetry From a Spacecraft. Remote Sensing Environ., vol. 1, no. 1, Mar. 1969, pp. 71-80.
5-6. Anding, David C.; and Walker, John P.: Use of Skylab EREP Data in a Sea Surface Temperature Experiment NASA CR-144479, 1975.
5-7. Hollinger, James P.; and Lerner, Robert M.: Analysis of Microwave Radiometric Measurements From Skylab. NASA CR-147442, 1975.
5-8. Trumbull, James V. A: The Utility of Skylab Photointerpreted Earth Resources Data in Studies of Marine Geology and Coastal Processes in Puerto Rico and the Virgin Islands. NASA CR-147437, 1975.
5-9. Polcyn, Fabian C.; and Lyzenga, David R.: Skylab Remote Bathymetry Experiment. NASA CR-144482, 1976.
5-10. Campbell, W. J.; Ramseier, R. O.; Weaver, R. J.; and Weeks, W. F.: Skylab Floating ice Experiment. NASA CR-147446,1975.
5-11. Pirie, Douglas M.; and Steller, David D.: California Coastal Processes Study. NASA CR-144489, 1975.
5-12. Nichols, Maynard M.: Southern Chesapeake Bay Water Color and Circulation Analysis. Water Color and Circulation Southern Chesapeake Bay, Part 1. NASA CR-141404, 1975.
5-13. Gordon, Hayden H.; and Nichols, Maynard M.: Skylab MSS vs Photography for Estuarine Water Color Classification. Water Color and Circulation Southern Chesapeake Bay, Part 11. NASA CR-141404, 1975.
5-14. Marshall, Harold G.; and Bowker, David E.: The Use of Skylab in the Study of Productivity Along the Eastern Shelf Waters of the United States. NASA CR-144908, 1976.
5-15. Korb, C. Laurence; and Potter, John F.: Ocean Properties. NASA TM X-72961, 1976.
5-16. Szekielda, Karl-Heinz: Dynamics of Plankton Populations in Upwelling Areas. NASA CR-144480, 1975.
5-17. Watanabe, Kantaro; Kuroda, Ryuya; Hata, Katsumi; and Akagawa, Masaomi: Study of Sea ice in the Sea of Okhotsk and its Influence on the Oyashio Current. NASA CR-144472, 1975.
5-18. Savastano, K. J.: Application of Remote Sensing for Fishery Resources Assessment and Monitoring. NASA CR-147507, 1975.
5-19. Breaker, Lawrence C.: A Qualitative Look at Sea Surface Temperature off the West Coast of the United States as Seen by UHRR imagery From the NOAA-3 Satellite. NOAA, Redwood City, Calif., 1974.
5-20. Bakun, A.: Daily and Weekly Upwelling indices, West Coast of North America, 1967-73. NOAA Tech. Rep. NMFS SSRF-693. NOAA National Marine Fisheries Service, 1975.
 5-21. Bakun, A.; and Nebson, C. S.: Climatology of Upwelling Related Processes off Baja California. Paper presented at the Symposium on Fisheries Science at the Autonomous Univ. of Baja Calif., Ensenada, Baja Calif, Feb. 16-22,1975.
5-22. Klemas, Vytautas; Bartlett, David S.; et al.: Skylab/EREP Application to Ecological, Geological, and Oceanographic investigations of Delaware Bay NASA CR-144910, 1976.
5-23. Maruyasu, Takakazu; Ochiai, Hiroaki; and Nakano, Takamasa: investigation of Environmental Change Patterns in Japan. NASA CR-144569, 1975.
5-24. Piech, Kenneth R.; Schott, John R.; and Stewart, Kenton M.: 5190 Interpretation Techniques Development and Applications to New York State Water Resources. NASA CR-144499, 1975.
5-25. Yost, Edward F.: In Situ Spectroradiometric Calibration of EREP imagery and Estuarine and Coastal Oceanography of Block island Sound and Adjacent New York Coastal Waters. NASA CR-147469, 1975.
5-26. Maul, George A.; Gordon, Howard R.; et al.: An Experiment to Evaluate Skylab Earth Resources Sensors for Detection of the Gulf Stream. NASA CR-147454, 1976.
5-27. Chang, David T.; and Izaacs, Ronald G.: Experimental Evaluation of Atmospheric Effects on Radiometric Measurements Using the EREP of Skylab. NASA CR-144500, 1975.
5-28. Pitts, David E.; Sasaki, Yoshikazu; and Lee, Jean T., compilers: Severe Storm Environments: A Skylab EREP Final Report. NASA TM X-58184, 1976.
5-29. Curran, Robert J.; Salomonson, Vincent V.; and Shenk, William: The Application of Satellite Data in the Determination of Ocean Temperatures and Cloud Characteristics and Statistics. NASA CR-147539, 1976.
5-30. Kuhn, P. M.; Marlatt, W. E.; and Whitehead, V. S.: The Skylab Concentrated Atmospheric Radiation Project. NASA CR-144481, 1975.
5-31. Thomson, F. J.: Machine Processing of S-192 and Supporting Aircraft Data: Studies of Atmospheric Effects, Agricultural Classifications, and Land Resource Mapping. NASA CR-144503, 1975.
5-32. Tingey, David L.; and Potter, John: Quantitative Determination of Stratospheric Aerosol Characteristics. NASA CR-147444, 1975.
5-33. Turner, Robert E.: investigation of Earth's Albedo Using Skylab Data. NASA CR-147445, 1976.
5-34. Rayleigh, J. W. Strutt: On the Light From the Sky; its Polarization and Colour. Phil. Mag., vol. 41, 1871, Pp. 107-120 and 274-279.
5-35. Mie, G.: Contributions to the Optics of Denze Media of Special Colloidal Metal Solutions. Ann. Physik, vol. 25,1908, pp. 377-445.
5-36. Van de Hulst, H. C.: Scattering in a Planetary Atmosphere. Astrophys. J., vol. 107, no. 2,1948, pp. 220-246.
5-37. Marlatt, W. E.: The Effect of Weather Modifications on Physical Processes. The Microclimate in Ground Level Climatology, R. H. Shaw, ed., AAAS (Washington, D.C.), 1967, pp. 295-308.
5-38. Alishouse, John; Jacobowitz, Herbert; and Wark, David: A Cloud Physics investigation Utilizing Skylab Data. NASA CR-147474, 1975.
5-39. Villevieille, A.; and Weiller, A. B.: The Possibility of Evaluating Vertical Wind Profiles From Satellite Data. NASA CR-147475, 1975.
5-40. Kuettner, J. P.: Cloud Bands in the Earth's Atmosphere: Observations and Theory. Tellus, vol. 23, no. 4-5, 1971, pp. 404426
5-41. LeMone, Margret Anne: The Structure and Dynamics of Horizontal Roll Vortices in the Planetary Boundary Layer. J. Atmos. Sci., vol. 30, no. 6, Sept. 1973, pp. 1077-1091.
5-42. Marwitz, J. D.: The Structure and Motion of Severe Hailstorms. Part 1: Supercell Storms. J. Appl. Meteorol., vol. 11, 1972, pp. 166-179.
5-43. Davies-Jones R. P.: Discussion of Measurements inside High-Speed Thunderstorm Updrafts. J. Appl. Meteorol., vol. 13, 1974, pp. 710-717
5-44. Sasaki, Y. K.: Mechanism of Squall-Line Formation as Suggested From Variational Analysis of Hourly Surface Observations. Preprints of the Earth Conference on Severe Local Storms, American Meteorol. Soc. (Boston, Mass.), Oct. 15-17, 1973.
5-45. Beebe, R. C.: Large Scale irrigation and Severe Storm Enhancement. Symposium on Atmospheric Diffusion and Air Pollution, American Meteorol. Soc. (Boston, Mass.), 1974, pp. 392-395.
5-46. Moore, Richard K.; and Ulaby, Fawwaz T.: The Radar Radiometer. Proc. IEEE, vol. 57, no. 4, Apr. 1969, PP. 587-590.
5-47. Mitsuyasu, A.; and Honda, T.: The High Frequency Spectrum of Wind Generated Waves. J. Oceanograph. Soc. Japan, vol. 30, no. 4,1974, pp. 29-42.
5-48. Pierson, Willard J.: The Theory and Applications of Ocean Wave Measuring Systems At and Below the Sea Surface, on the Land, From Aircraft, and From Spacecraft NASA CR-2646, 1976.
 549. Cardone, Vincent J.; Yoong, James D.; et al.: The Measurement of the Winds Near the Ocean Surface With a Radiometer Scatterometer on Skylab. A Joint Meteorological Oceanographic and Sensor Evaluation Program for Experiment S193 on Skylab. NASA CR-147487, 1976.
5-50. Ross, Duncan: A Comparison of Synoptic and Skylab S193/194 Determinations of Ocean Surface Windspeeds. NASA CR-147540, 1975.
5-51. Cardone, V. J.: Specification of the Wind Field Distribution in the Marine Boundary Layer for Wave Forecasting. Rep. TR 69-1, Geophys. Science Lab., New York Univ., Dec. 1969.