Kansas Water Science Center

Solar-Irradiance Variations and Regional Precipitations
in the Western United States

Charles A. Perry
Water Resources Division, U.S. Geological Survey, Lawrence, Kansas USA

Table of Contents

Figures

    Figure 1. Depth of extinction of the solar radiation spectrum in water
    Figure 2. Maps showing SST anomalies in the Pacific Ocean for (a) 1986 and (b) 1987
    Figure 3. Schematics of the effect of (a) warm water and (b) cool water on the atmosphere
    Figure 4. Empirically modelled values of (a) monthly irradiance, (b) monthly differences in irradiance and (c) annual averages of monthly differences of solar irradiance
    Figure 5. Correlation coefficients for regression of annual regional precipitation with annual averages of monthly irradiance for:
      Figure 5AB. (A) 0-year and (B) 1-year lag times, 1950-88
      Figure 5CD. (C) 2-year and (D) 3-year lag times, 1950-88
      Figure 5EF. (E) 4-year and (F) 5-year lag times, 1950-88
      Figure 5GH. (G) 6-year and (H) 7-year lag times, 1950-88
    Figure 6. Correlation of precipitation and irradiance for (a) coastal Oregion region 1 and for (b) s outh-east Washington region 10.

Abstract

Changes in total solar irradiance can be linked to changes in regional precipitation. A possible mechanism responsible for this linkage begins with the absorption of varying amounts of solar energy by the tropical oceans which creates ocean temperature anomalies. These anomalies are then transported by major ocean currents to locations where the stored energy is released into the atmosphere, altering atmospheric pressure and moisture patterns that can ultimately affect regional precipitation.

Correlation coefficients between annual differences in empirically modeled solar-irradiance variations and annual state-divisional precipitation in the United States for the period 1950-88 were computed with lag times of 0 to 7 years. The most significant correlations occur in the Pacific Northwest with a lag time of 4 years, which is approximately equal to the travel time of water within the Pacific Gyre from the western tropical Pacific Ocean to the Gulf of Alaska. Precipitation in the Desert Southwest correlates significantly with solar irradiance lagged 3 and 5 years, which suggests a link with ocean-water temperature anomalies transported by the Equatorial Countercurrent as well as the North Pacific Gyre. With the correlations obtained, droughts coincide with periods of negative irradiance differences (dry high-pressure development), and wet periods coincidewith periods of positive differences (moist low-pressure development).

Introduction

Variations in the total solar irradiance may be responsible for short-term climatic variations. Until documentation of the long-term variation in total solar irradiance its relationship to the solar magnetic cycle, therehave been no well-documented external-forcing process to explain climatic cycles that are in the range of 1 to 10 years. Earth-satellite measurements since 1978 have revealed that total solar irradiance has an average variation of at least 0.1% over the 11-year period of a sunspot cycle (Willson and Hudson, 1988). However, monthly and single year variations can approach the magnitude of the long-term trend. The 10.7-cm solar flux is a proxy for the bright faculae regions. These regions contribute to total irradiance by increasing total radiative output. The 10.7-cm flux has been measured on Earth since 1947, and it too shows a pattern that corresponds with sunspot patterns (Lean and Foukal, 1988). Sunspots contribute to the total irradiance, but by decreasing the radiative output. During periods of high sunspot number, however, there are a proportionally greater number of bright faculae regions which have a net effect of increasing total irradiance. Estimates of total solar-irradiance variations between 1874 and 1988 are available, based on an empirical model using past solar activity (Foukal and Lean, 1990). Solar irradiance may play an important role in the global climatic system, but the variations are small, and their effect must be amplified to cause significant climatic variations.

One possible medium of amplification could be through the World's oceans. Variations in the temperature of the ocean water, specifically the sea-surface temperature (SST) have been linked to atmospheric-pressure anomalies (Wallace et al., 1990). There are indications that anomalously cool SST in the eastern Pacific were responsible for the severe 1988 North American drought (Palmer and Brankovic, 1989). Ocean currents serve as the major conveyors of energy from the tropics toward the poles. It is possible, therefore, that these currents could transport anomalously warm or cool pools of water to latitudes where they could alter atmospheric pressure and moisture patterns that may affect regional precipitation. As these pools of water move around the ocean gyres a succession of climatic patterns may ensue at any one location and a regional pattern may be seen on a global scale. Langbein and Slack (1982) identified national and regional patterns in runoff and frequency of dry years in the United States and noted that wet and dry patterns migrated from west to east across the continent in a time frame of about 5 years.

Short-term regional climatic variations in the United States may be affected by ocean temperature patterns which could be, in part, a result of solar-irradiance variations. Temperature patterns in the Pacific Ocean should have the greatest effect on weather in the Western United States. The response time of the climatic variable, precipitation, is shown to be commensurate with the time of travel of water within the Pacific Ocean Gyre.

Solar/Climate Mechanism

The mechanism proposed for the coupling of solar-irradiance variations with regional climate consists of three basic components. These are: (1) absorption of solar energy by the transparent tropical oceans in a deep surface layer, (2) transport of that energy by major ocean currents, and (3) transfer of that energy into atmospheric moisture and low pressure systems that would be advantageous for precipitation formation (Perry, 1992). Each individual component has inherent complexities that are difficult to separate. However, the sun's energy is the driving force for weather and climate, and any variations in that energy have the potential to affect precipitation formation and distribution.

Absorption

Although SST's show significant coupling with atmospheric parameters (Wallace et al., 1990), it is not the ocean's surface that stores the majority of the solar energy. The visible spectrum contains about one-half of the total energy available from the Sun at the Earth's surface (Loiv, 1980), and those wavelengths can penetrate well below the ocean's surface. Lewis et al. (1990) showed that solar radiation in visible frequencies, usually assumed to be absorbed at the sea surface, penetrates to a significant depth below the upper mixed layer of the ocean that interacts directly with the atmosphere. Figure 1 shows the depth of extinction of the solar spectrum in water.

The transparency of the tropical oceans is dependent upon the amount of biogenic material, phytoplankton pigments, and the degradation products that are present. In the Pacific Ocean, transparency increases from east to west, with greatest penetration of solar energy occurring in the western tropical Pacific. The net radiative transport of heat downward through the base of the mixed layer (which varies from 10 m in the eastern and western Pacific to about 60 m in the central Pacific) is approximately equivalent to the estimated climatological net surface-heat flux into the ocean over much of the western Pacific (Lewis et al., 1990). This heat eventually returns to the ocean's surface months or years later to interact with the atmosphere as the general circulation of the Pacific Gyre transports the water northward and eastward toward North America.

Other factors that affect absorbance of solar energy by the ocean include latitude, season, sea-surface roughness, atmospheric particulates, and cloud conditions. All factors contribute in various degrees to the development of anomalous ocean temperatures. However, solar irradiance fluctuations may a significant effect upon the initial formation of these ocean temperature anomalies.

Transport

The North Pacific Ocean Gyre is the largest in the World, with an outer circumference of approximately 25,000 km. It has a clockwise circulation, with the fastest surface currents in the northwest section, just east of Japan, and the slowest currents west of Mexico. Typically ocean current velocities range from 1 to 16 km/day (Niiler, 1986). Using an average speed of 5-8 km/day, one rotation of the gyre could take approximately 9-12 years. Time of travel from the western tropical Pacific to near North America would take somewhat less than one-half this time (4 to 5 years) because the currents are generally faster in the northwestern one-half of the gyre.

Two minor circulation patterns in the Pacific warrant discussion. The Equatorial Counter Current flows eastward between the westward flowing sections of the North Pacific and South Pacific Gyres. Upon this counter current rides the infamous and elusive oceanic warming of El Nino and the oceanic cooling of La Nina. The other minor circulation pattern is the counterclockwise flow of the Gulf of Alaska Gyre. It could be considered a large eddy of the North Pacific Gyre. This circulation is driven by the eastward flowing North Pacific Drift Current, which splits west of British Columbia. The northward-flowing part warms the Alaskan coast and the southward-flowing part becomes the California Current.

The Pacific Gyre and its minor circulations are the conveyors of absorbed solar energy from the central and western tropical Pacific to locations north and east. If incoming solar energy varies on time scales of months to years, then different parts of the gyre will be warmed to different temperatures as it rotates. During a period of diminished irradiance, a part or pool of the tropical ocean would receive less energy and be anomalously cool, whereas increased irradiance would result in an anomalously warm pool. These pools of slightly warmer or cooler water would be drawn around the Pacific Gyre like riders on a carousel.

Plots of annual SST anomalies in the Pacific Ocean during 1962 to 1990 show large areas of both positive and negative temperature anomalies that persist for several years and move around the Pacific Gyre. For example, figure 2 shows the center of a very large cool SST anomaly in the mid-Pacific in 1986 moving 20 degrees longitude (approximately 2000 km) eastward by 1987 at a rate of approximately 5.5 km/day. By 1988, this SST anomaly was well established off the west coast of the United States and is noted as a possible cause of the severe 1988 nationwide drought (Palmer and Brankovic, 1989).

Transfer of Energy to the Atmosphere

In the tropics, the net flux of energy is downward into the ocean, whereas in the higher latitudes the net flux is upward. Energy is transferred from the ocean to the atmosphere by two processes. One is by conduction of infrared radiation to the air and the resulting convection of the warmed air just above the surface to greater heights, and the other is by the release of water vapor. The amount of energy released by evaporation greatly exceeds that released by conduction and convection because of the approximately 580 cal needed to convert 1 g of liquid water to vapor.

Evaporation from the surface of the ocean is a mechanism that could amplify the effect of solar-irradiance variations. The vapor pressure of water increases by approximately 7% for each 1 °C of increase in water temperature between 5 and 15 °C. Therefore, a 1 °C positive anomaly in SST could result in about a 7% increase in the amount of energy available to the atmosphere. Variations in the vapor pressure at the sea surface can significantly alter atmospheric moisture fields, from which further amplification of solar variations can occur through dynamic atmospheric processes.

Schematic illustrations of the effects of warm and cool ocean water anomalies on the atmosphere are shown in figure 3. Warm water produces more water vapor for the atmosphere than cool water. Water vapor is less dense than dry air, resulting in lower atmospheric pressure over warm water than over cool water. Low pressure allows convergence of air near the surface, followed by lifting and adiabatic cooling of the air until condensation and formation of clouds, and ultimately the release of the latent heat into the upper levels of the atmosphere. The release of the latent heat increases the buoyancy of the air and further lifting occurs. The greater the supply of energy available at the ocean surface, the greater the probability of cloud formation, precipitation, and the injection of energy into the upper atmosphere. Energy patterns in the upper atmosphere are responsible for the persistent upper air flow patterns that interact on a global scale. The strength, curvature, and location of the upper level winds, which are critical factors in precipitation formation, are largely dependent upon temperature patterns generated by the oceans and continents.

Upper air flow patterns can generate new surface or low level pressure systems. For example, an upper level low may be formed initially by a surface low that extends upward into the atmosphere. In the middle latitudes, the predominant westerlies would advect that upper level low downstream (to the east) away from its initial surface-energy supply. As the upper level low moves eastward, it could induce low level cyclogenesis. Energy released from the induced low could be added to the upper level low, maintaining or even strengthening it. In the absence of available energy in the atmosphere near the surface, the upper-level low would not be reinforced, expending its energy and dissipating.

Irradiance/Precipitation Hypothesis

The hypothesis that the combination of absorption of solar energy by water in the western Pacific Ocean, the transport of that energy by the currents of the Pacific Ocean Gyre, and the modification of the atmosphere by warmer or cooler pools of water affecting precipitation distributions in North America is tested in this paper. For increased precipitation in the western parts of the United States, the following scenario is suggested:

The irradiance of the Sun generally increases each month for a year. Water slowly moving westward along the southwestern part of the Pacific Gyre absorbs this increasing energy and stores it at depth, increasing the temperature slightly in a large volume of water. Nearing the Asian landmass this slightly warmer than normal volume of water turns northward. Then, irradiance of the Sun begins a year of decrease, and the water now in the southwestern part of the Pacific Gyre receives less energy. The result is that the leading pool of water has more stored energy than the trailing pool. As both pools travel around the gyre, the stored energy at depth begins to be expelled at the surface by evaporation. The leading pool with its greater energy content could spawn rain showers or thunderstorms that would inject moisture into the mid-levels of the atmosphere. The trailing pool, relinquishing less energy to the atmosphere, would be shadowed by a fairer weather pattern.

Once the leading pool enters the North Pacific Drift, at least 1 to 2 years have passed since it turned northward. Here, the energy within the warm pool supports extratropical storms that are driven eastward by the prevailing westerlies. Still far from North America, these storms would expend their precipitation over the open ocean but, nonetheless would affect global atmospheric flow patterns. One or 2 additional years pass, and the leading pool of water, cooler now but still warmer than the pool following it, continues to supply energy to the atmosphere. North America is much closer now, and the storms make landfall bringing precipitation to the northwestern part of the United States. If the pool of warmer than normal ocean water is large enough, a part of it may be drawn into the eddy of the Gulf of Alaska Gyre, supplying energy to the Aleutian low-pressure system for an extended length of time. If the pool is smaller, it may exhaust its surplus stored energy and lose its ability to significantly alter the atmosphere. Also, some of the warm pool may be drawn into the California Current and continue to affect atmospheric conditions farther south and east.

The trailing pool of cooler than normal water would have a negative effect on the formation of storms. As this pool moves across the Northern Pacific, fewer storms would be associated with it, and less precipitation would be the result. A large pool of cooler than normal water could be incorporated into the Gulf of Alaska Gyre, and surface low-pressure formation would be suppressed for an extended period of time, bringing multi-year droughts to North America.

The entire Pacific Gyre could be considered as a slowly rotating oblate disk, with its elements moving at various velocities, gaining or losing energy as a function of solar irradiance, latitude, season, sky conditions, water transparency, and surface conditions. The elements of anomalously warm or cool pools of ocean water within this disk could affect surface and upper atmospheric temperature, moisture, and wind patterns including the position and strength of the jet stream (Maxwell, 1992) several years after their initial formation.

Data

Data needed to test the hypothesis that solar variations affect regional climate include the independent variable, solar irradiance fluctuations, and the dependent climatic variable precipitation. Annual regional precipitation was chosen as the unit of measure for several reasons. Annual values eliminate the seasonal precipitation variations and provide a better time scale for global influences. Utilization of regional data reduces the variability of point precipitation data. The use of regional data also provides the opportunity to detect regional variations of climate that may have been forced by a single global factor, ie. solar irradiance variations.

Solar Irradiance

Solar irradiance has been measured in space by sensors of the Earth Radiation Budget Experiment (ERB) on the Nimbus-7 satellite since November of 1978 (Hickey et al., 1980). Presently (1993), these measurements account for more than 14 years of nearly continuous data. The ERB data correlate well with data collected by the Active Cavity Radiometer (ACRIM) that flew on the Solar Maximum Mission. However, 14 years is a short time for comparison with climatic data. Fortunately, Lean and Foukal (1988) developed an empirical model for total solar irradiance that is based on changes in excess radiation from bright magnetic faculae and on changes in reduced radiation from dark sunspots. Using this model, estimates of bright magnetic faculae were made back to 1954 using daily 10.7-cm flux (Lean and Foukal, 1988). Estimates of irradiance were later extended back to 1874 by Foukal and Lean (1990) using monthly means of the sunspot number in place of the 10.7-cm flux. This later model relied on the correlation between the monthly sunspot number and monthly 10.7-cm flux between 1947 and 1988. Because the 10.7-cm flux was not measured before 1947, the 42 year period 1947-88 of monthly values of irradiance was assumed to have the greatest reliability for the period from 1874 to 1988. Irradiance values generated by this later model are shown in figure 4a. This paper is based on the hypothesis that increases or decreases in solar irradiance have an effect on climate. Therefore, monthly differences (Figure 4b) and the annual average of monthly differences of solar irradiance were determined (Figure 4c).

Regional Precipitation

Monthly precipitation data for the United States were arranged into 344 regions according to the State divisions of climatological data (National Oceanic and Atmospheric Administration, 1989). These regions included all the states except Alaska and Hawaii. Each annual regional precipitation average is an arithmetic mean of the precipitation for the stations within that region from January through December. The number of stations within any one division varies from less than 5 to more than 50.

Analysis

Correlation coefficients were calculated between the annual average of the monthly differences of the modeled solar-irradiance values hereafter referred to as the annual average irradiance difference and the annual average precipitation values for each of the 344 regions for the period 1950-88. The differences between monthly solar irradiance values were summed for the interval between January and December and divided by 12 to obtain the annual average irradiance differences. Eight series of annual average irradiance differences, with time lags from 0 to 7 years, were correlated with precipitation data. For example, in a 2-year lag correlation, precipitation was correlated to the annual average irradiance difference that occurred 2 years previous. Correlation coefficients for each lag time were mapped for the 48 contiguous States. Correlation coefficients for each of the 344 regions for all lag times are shown in Figures ( 5ab, 5cd, 5ef, 5gh). Correlation coefficients are plotted at the centroid of each region; correlation coefficients between 0.20 and -0.20 were not contoured.

Correlation coefficients (R) for the eight lag times and the 344 regions ranged between -0.51 and +0.65. To be significant at the 1 per cent level, R must be less than -0.37 or greater than 0.37. The highest correlation coefficient (R = 0.65) was obtained in the Pacific Northwest when the annual average irradiance differences were lagged 4 years. At this time lag, droughts coincided with periods of lower irradiance averages (decreased energy availability), and greater than average precipitation coincided with periods of higher irradiance averages (increased energy availability). Examples of the correlation between annual average irradiance differences lagged 4 years and precipitation averages for two regions are shown in Figure 6. A region of abundant precipitation (the 39-year average is approximately 200 cm) is shown in Figure 6a, using data from Oregon's coastal region 1. A region of meager precipitation (the 39-year average is approximately 44 cm) is shown in Figure 6b using data from southeastern Washington region 10, the Palouse Blue Mountain area. Correlation coefficients of R > 0.60 were obtained in five other regions in Oregon, eastern Washington, and western Idaho. Coefficients of R > 0.50 were obtained for surrounding regions, including a large region in northern California ( Figure 5ef). Positive correlations coefficients of R > 0.50 also were obtained for the southeastern part of the United States and along the eastern seaboard for the 4-year lag time. This could be a reflection of a long-wave, trough-ridge-trough pattern as forced by Pacific Ocean temperatures. Weak negative correlation coefficients occur in Texas, New Mexico, eastern Montana, northeastern Wyoming and western North and South Dakota, midway between the areas of positive correlations, supporting the trough-ridge-trough condition.

The strongest positive correlations were obtained with the 4-year lag time but significant correlations exist at other lag times suggesting the influence of other oceanic areas. Significant positive correlation coefficients exist for regions in California for a lag time of 3 years. The 3-year lag positive correlation could be caused by temperature anomalies in water arriving west of Baja California after traveling from the western and central tropical Pacific Ocean along the Equatorial Counter Current. Significant positive correlations coefficients are obtained from Texas to Nebraska with a 2-year lag, indicating the possibility of ocean-temperature anomalies in the Caribbean Sea and the Gulf of Mexico affecting precipitation in the plains States.

Lag times of 0, 1, 2, 3, 5, 6, and 7 years showed virtually no positive correlation coefficients in the Pacific Northwest. However, for a 0-year lag time, the Pacific Northwest and the central United States had significant negative correlations R less than -0.37 indicating an inverse relation between annual average irradiance differences and regional precipitation. Coefficents less than -0.37 were obtained for regions in Washington, Idaho, Montana, Wyoming, Utah, Iowa, Missouri, Illinois, and Vermont. Nearly all of the other regions in the United States had weak negative correlations for the 0-year lag time. The negative correlations could be an indication of an inverse relation between irradiance and relative humidity. For example, increased irradiance over a continental landmass would immediately raise air temperatures. With no large source of water at the surface to evaporate, relative humidities would decrease, hampering precipitation formation and encouraging drier conditions. At a lag of 1 year, negative correlation coefficients less than R = -0.37 move eastward, appearing in Nebraska, South Dakota, Iowa, and New Hampshire.

A possible cause of the persistent drought in the west coast States from 1984 through 1991 may be the a period of persistent negative solar-irradiance differences that occurred from 1980 through 1987 (Figure 6). This period appears to be the longest within the last 42 years of modeled irradiance and may have helped create the very large areas of negative SST anomalies observed in the North Pacific from 1984-90. Significant increases in solar irradiance that occurred in 1988 may have induced the greater than average precipitation in these same west coast States four years later during the winter of 1992-93.

The patterns seen in the correlation coefficient maps for the various lag times could be a result of persistent upper level pressure systems and low level moisture patterns caused by persistent ocean-temperature anomalies moving around the Pacific Gyre. These ocean temperature anomalies may be created by solar-irradiance variations. The Pacific Ocean temperature anomalies probably have an effect on precipitation patterns across the entire North American continent but are most evident in the Pacific Northwest. Ocean temperature anomalies in the Gulf of Mexico, Caribbean Sea, and the Atlantic Ocean probably have an increasing influence in the eastern parts of the continent.

Solar-irradiance variations correlate with precipitation in the Pacific Northwest and other areas in the United States, but are there other locations or phenomena that show a link to irradiance variations? One such phenomenon may be the El Nino. The 1982-83 El Nino, which had by far the most intense warming of the SST since 1950 at Puerto Chicama, Peru (Enfield 1988), occurred 3 years after the largest total increase in solar irradiance during the same time period. The Equatorial Counter Current could have transported a very warm pool of water from the west-central Pacific to the eastern Pacific during that 3-year interval. The El Nino of 1986-87 may have been weakened substantially by the negative solar irradiance differences of 1983-84 (Figure 4c). This relation between solar-irradiance and El Nino strength may help explain the persistence of the 1991 through 1993 El Nino which followed solar irradiance increases during 1988 to 1990.

Summary and Conclusions

Annual precipitation data for the 344 state-divisional regions in the United States were correlated with solar-irradiance data lagged 0 through 7 years. Annual averages of monthly differences of empirically modeled solar-irradiance values show significant correlations with annual regional precipitation at certain lag times. Periods of increased annual average irradiance differences correspond to periods of increased precipitation, whereas periods of decreased averages correspond to decreased precipitation in the Pacific Northwest, when irradiance is lagged 4 years. Travel time for water moving from the warm western tropical Pacific Ocean to the Gulf of Alaska within the Pacific Gyre also is approximately 4 years, further evidence for an oceanic transfer mechanism.

These significant correlations may be due to a solar-climate mechanism which involves a combination of the processes of absorption, transport and transfer of varying amounts of energy from the Sun to the oceans, and back to the atmosphere. Large quantities of energy in the visible spectrum can be injected into ocean below the mixing layer. This energy then is transported by major ocean currents to locations where the energy flux becomes upward into the atmosphere. The effects of solar-irradiance variations may be amplified by the 7% increase in the vapor pressure of water for every 1 °C increase in ocean temperature. The energy is transferred to the atmosphere and becomes available for the formation of storms that produce precipitation.

The persistent drought in the western States from 1985-91 may be related to a period of decreasing solar irradiance that occurred between 1981-87. Relatively large increases in solar irradiance difference that occurred from 1988 to 1990 may have helped break that drought four years later during the winter of 1992-93 when greater than average precipitation fell.

References

Beer, T., 1983, Environmental Oceanography, Pergamon Press, Oxford, 206 pp.

Favorite, F. and McLain, D.R., 1973, Coherence in transpacific movements of positive and negative anomalies of sea surface temperature 1953-60: Nature, v. 244, p. 139-143.

Foukal, P. and Lean J., 1990, An empirical model of total solar irradiance variation between 1874 and 1988: Science, v. 247, p. 556-559.

Hickey, J.R., Pellegrino, P., Mashhoff, R.H., House, F., and Vonder Haar, T.H., 1980, Initial solar irradiance determination from Nimbus 7 cavity radiometer measurements: Science, v. 208, p. 281.

Kuhn, J.R., Libbrecht, K.G., and Dicke, R.H., 1988, The surface temperature of the Sun and changes in the solar constant: Science, v. 242, p. 908-911.

Langbein, W.B. and Slack, J.R., 1980, A marked pattern in U.S. and regional runoff: Water Resources Review, January 1980, p. 18.

Langbein, W.B. and Slack, J.R., 1982, Yearly variations in runoff and frequency of dry years for the conterminous United States, 1911-79: U.S. Geological Survey Open-File Report 82-751, 88 p.

Lean, J. and Foukal, P., 1988, A model of solar luminosity modulation by magnetic activity between 1954 and 1984: Science, v. 240, p. 906-908.

Lewis, M.R., Carr, M., Feldman, G.C., Esaias, W., and McClain, C., 1990, Influence of penetrating solar radiation on the heat budget of the equatorial Pacific Ocean: Nature, v. 347, p. 543-545.

Liov, K.N., 1980. An Introduction to Atmospheric Radiation, Academic Press, New York, 392 pp.

Maxwell, B.D., 1992, Effects of North Pacific Sea Surface Temperature Anomalies on the Atmosphere over the North Pacific and North America: Master's Thesis, University of Kansas, Lawrence, Kansas.

National Oceanic and Atmospheric Administration, 1989. Climatological data for the United States: National Climatic Center, Asheville, North Carolina.

Niiler, P.O., 1986. The observational basis for large-scale circulation, in Abarbanel, H.D.I. and Young, W.R., eds., General Circulation of the Ocean: Springer-Verlag, New York, 291 pp.

Palmer, T.N. and Brankovic, C., 1989. The 1988 U.S. drought linked to anomalous sea surface temperature: Nature, v. 338, p. 54-57.

Perry, C.A., 1992, A Correlation between precipitation in the Western United States and solar-irradiance variations, in Proceedings of the American Water Resources Association Conference Managing Water Resources During Climate Change, Reno, Nevada, November 1-6, p. 721-729.

Wallace, J.M., Smith, C., and Jiang, Q., 1990, Spatial patterns of atmosphere-ocean interaction in the northern winter: Journal of Climate, v. 3, p. 990-998.

Wlllson, R.C., and Hudson, H.S., 1988, Solar luminosity variations in solar cycle 21: Nature, v. 332, p. 810-812.