Impacts of Climate Change and Land Use  on the Southwestern United States

Climatic variability

Climate and Droughts

Alan L. McNab
Thomas R. Karl
National Oceanic and Atmospheric Administration

This paper is reproduced in its entirety from Hanson, R.L., 1991, Evapotranspiration and Droughts, in Paulson, R.W., Chase, E.B., Roberts, R.S., and Moody, D.W., Compilers, National Water Summary 1988-89--Hydrologic Events and Floods and Droughts: U.S. Geological Survey Water-Supply Paper 2375, p. 99-104..


figure 1

Figure 1. Propagation of precipitation defecits through other components of the hydrologic cycle. (Modified from Changnon, 1987).

A drought is a complex phenomenon that can be defined from several perspectives (Wilhite and Glantz, 1987). The central theme in the definitions of a drought is the concept of a water deficit. A drought is difficult to define because of the need to specify the component(s) of the hydrologic cycle affected by the water deficit and the time period associated with the deficit. The simultaneous occurrence of a long-term deficit in deep ground-water storage and a short-term surplus of soil water in the root zone is an example of the complexity encountered in defining a drought.

Changnon (1987) illustrates how the definitions of a drought are related to specific components of the hydrologic cycle and how precipitation deficits are related to drought (fig. 1). Figure 1 shows how the effects of two hypothetical precipitation deficits are propagated over time through the surface-runoff, soil-moisture, streamflow, and ground-water components of the hydrologic cycle. From this perspective, precipitation can be considered to be the carrier of the drought signal, and streamflow and ground-water levels can be considered to be the last indicators of the occurrence of a drought (Hare, 1987; Klemes, 1987).

If precipitation is the carrier of the drought signal, then climate describes the long-term characteristics of this signal. The climatic factors associated with drought, including a description of the local climate in areas that have precipitation deficits, are described in the following section. These factors are then related to atmospheric circulations that extend well beyond the local area. Finally, possible causes of drought-related atmospheric circulations and the relation of these causes to nonatmospheric factors are described.



Most definitions of a drought refer to abnormal dryness; normal dryness, such as during the summer in the southwestern United States, does not constitute a drought. The strongest drought signals are recognized during seasons when substantial precipitation is expected but fails to fall (Karl and others, 1987). For instance, the severe droughts in the Great Plains (1890's, 1910's, 1930's, and 1950's) were associated with a lack of precipitation during the spring and summer, which normally are the wettest seasons (Borchert, 1971).

Abnormally large amounts of precipitation during normally dry seasons are particularly effective in ending a drought. For instance, even though tropical cyclones do pass over the eastern and southern United States during droughts, the precipitation that results generally is insufficient to end the droughts (Cry, 1967). Tropical cyclones that pass over during months when normal precipitation is minimal, however, produce much greater drought amelioration than those that pass over those areas during months that normally are wet.

The difference between normal precipitation and deficient precipitation commonly depends on precipitation from just a few storms. The 1930-60 precipitation record for the Upper Colorado River Basin illustrates the extent to which the total precipitation in a region can depend on a few periods of substantial precipitation as opposed to numerous periods of minimal or moderate precipitation. The record shows that about 50 percent of the annual precipitation in the region results from only about 25 percent of the storms. Therefore, the lack of a few large storms during a season can be sufficient to cause drought (Riehl, 1965).

Drought commonly is perceived to be an abnormally long period without precipitation. A decreased frequency of precipitation, however, is not the only climatological factor that causes precipitation deficiencies. Droughts also are associated with weather systems that result in only minimal precipitation. Bergman and others (1986) and Karl and Young (1987) note that minimal precipitation was delivered by storms along the Gulf and East Coasts during the severe drought of 1986 in the southeastern United States.

figure 2

Figure 2. Precipitation frequency (A) and intensity (B) at the Central Park Observatory, New York, N.Y. Anomalies (A and B) calculated with respect to 1925-65 average annual precipitation (41.73 inches), (A) average number of days per year that had precipitation (119 days per year), and (B) average precipitation per day that had precipitation (0.35 inch per day). (Sources: Data from Spar l968, and Namias l968.) Click to view full-size image.

The anomalies in precipitation frequency and intensity associated with dry and wet years for 1925- 65 at the Central Park Observatory, New York, N.Y., are shown in figure 2. Dry (precipitation below normal) and wet (precipitation above normal) years are defined in figure 2 as annual-precipitation anomalies (negative or positive departures from the 1925-65 average annual precipitation). Precipitation frequency is defined in figure 2A as the number of precipitation- days anomaly (the number of days that had precipitation minus the average number of days per year that had precipitation during 1925-65). Intensity is shown in figure 2B as the precipitation per precipitation-day anomaly (annual precipitation divided by the annual number of days that had precipitation less the 1925-65 average of this quantity).

There is a slight tendency for precipitation to be less frequent than normal during dry years (beige quadrant of fig. 2A). The substantial number of points in the blue quadrant of the graph, however, indicate that dry years can have as many days that have precipitation as wet years. With respect to intensity, this precipitation record shows a tendency for substantial daily precipitation during wet years (blue quadrant of fig. 2B) and minimal daily precipitation during dry years (beige quadrant of fig. 2B). Similar results have been reported in the central United States. An analysis of data for central lowa showed about the same number of days having precipitation for July during the drought of 1976 as for the normal July of 1977 (White and Vaughan, 1982). However, the size and number of precipitation radar echoes were substantially less during July 1976 than in July 1977. Changnon (1980) used a data set for 1931-68 to demonstrate that typical dry periods in July and August in Illinois have a normal number of days having minimal precipitation but few, if any, days having moderate to substantial precipitation.


Droughts, and particularly summer droughts or the summer periods of multiyear droughts, generally are associated with higher than normal surface- air temperatures. For example the drought of 1986 in the southeastern United States had associated surface-air temperatures so much higher than normal that year that Georgia, North Carolina, and South Carolina had their warmest July in the 20th century (Bergman and others, 1986; Karl and Young, 1987). Numerous other investigators (Namias, 1955, 1982a, 1983; Karl and Quayle, 1981) have remarked on the abnormally high surface-air temperatures during droughts in the summer and growing season in the Great Plains.

Not all droughts, however, are associated with higher than normal surface-air temperatures. The drought of 1962-65 in the northeastern United States is a well-documented example of a major drought that was associated with lower than normal surface-air temperatures in all seasons (Namias, 1966, 1968; Mitchell, 1968). To examine the simultaneous occurrence of abnormally dry and abnormally hot weather, regional temperature ranks for the 10 driest years between 1896 and 1988 were compiled (table 1). In the winter season, for example, in the Northeast, the 5th driest winter season also was the 36th warmest during the period. In the summer season, for example, for the Nation, the 1st driest season also was the 21st warmest for the same period. These data indicate that, on a national scale, there is a well-defined association between dry weather and higher than normal surface-air temperatures during the summer and a less well defined, but still apparent, similar association during the winter. This association, however, is not apparent on a regional scale. In the Northeast and West, dry summers are about equally likely to be associated with either higher than or lower than median surface- air temperatures. In the northeast and central United. States, dry winters are more likely to be associated with lower than median surface-air temperatures; several of the driest winters ranked within the 10 coolest winters. In the South and Southeast, dry summers consistently are ranked among the warmest.

Table 1. Regional and national (conterminous United States) winter- and summer-season ranks for the 10 driest winter and summer seasons, 1896-1988
[Areas: NE, northeast (Connecticut, Delaware, Maine, Maryland, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont); ENC, eastern north-central (lowa, Michigan, Minnesota, Wisconsin); C, central (Illinois, Indiana, Kentucky, Missouri, Ohio, Tennessee, West Virginia); SE, southeast (Alabama, Florida, Georgia, North Carolina, South Carolina, Virginia); WNC, western north-central (Montana, Nebraska, North Dakota, South Dakota Wyoming); S, south (Arkansas, Louisiana, Kansas, Mississippi, Oklahoma, Texas); SW, southwest (Arizona, Colorado New Mexico, Utah); NW, northwest (Idaho, Oregon, Washington); W, west (California, Nevada); U.S., conterminous United States. *=1988 temperature rank]
Temperature rank for indicated area
19375 8875 356619 404260
23170 5943423 60614484
3912 9150133 845326
4724138 2363724 888052
5*365956 22246333 347428
6171030 46191720 12862
7865383 531981 60145
8902547 8440223 917027
9711482 3774171 774615
10873252 8543281 763419
 ------ -------- ------
Average...6738634321 3334585134
1544418 2143 131621
248647 2351534 12208
366251 2810624 407940
447355 1832314 392623
53980*26 138022 321425
652*116 6648131 37432
713522 8624737 42221
8141866 252677 2574*9
9856914 9171149 11256
10711370 3631819 10990
 ------ -------- ------

Namias (1983) and Chang and Wallace (1987) constructed maps of the conterminous United States correlating precipitation and temperature for the summer. These maps show negative correlations stronger than minus 0.5 for the interior States (with strongest correlations of about minus 0.7 to minus 0.8) and almost zero to slightly positive correlations in the West Coast, Northeast, and Mid-Atlantic States.

Namias (1983) cites several case studies that show that increased surface-air temperatures during droughts extend into the midtroposphere. Chang and Wallace (1987) compiled data for 63 summers that show the air-temperature anomalies at 500 millibars are of the same sign as the surface-air temperature anomalies, except smaller by a factor of 2 or more.


Droughts commonly are referred to as "dry" in the sense that not only does less precipitation fall, but also the air is drier than usual. Several single- station case studies support this concept.

Relative humidity is a commonly used measure of atmospheric water vapor. Relative humidity depends on two factors-the absolute quantity of water vapor in the air (absolute humidity) and the air temperature. Baldwin (1957) presents a case study of 2 dry years and 2 wet years at San Antonio, Tex. (table 2). The mean relative humidity for the dry years was 56 percent, compared to 64 percent for the wet years. Potter (1958) compiled normal relative- humidity data for May collected at five stations in central Canada and compared these data with the data for the dry May of 1958. The five-station average relative humidity for May 1958 was about 50 percent compared to average normal relative humidity for May of about 65 percent.

Table 2. Relative humidity at 950 millibars and precipitation at San Antonio, Texas.
[data from Baldwin, 1957]
Month Relative humidity
Dry years
Wet years
  Dry years
Wet years
January5371 0.461.76
February4674 0.461.76
March4562 0.281.43
April6162 2.015.20
May6470 1.231.22
June6064 2.455.61
July5568 0.632.3
August6056 2.083.43
September5060 1.501.38
October6261 2.525.41
November5148 1.180.56
December5266 0.821.51
Total   15.6232.23

This decrease in relative humidity during droughts extends through all levels of the atmosphere that contain substantial water vapor. Namias (1966, 1978b) and Spar (1968) present several case studies showing that the relative humidity at less than 500 millibars is less during droughts than during nondrought periods.

Huff and Changnon (1963) conclude that, during summer droughts in Illinois, the decreased relative humidity is not well correlated to the absolute quantity of water in the atmosphere; that is, the increased temperature is the primary factor in causing the decrease in relative humidity. Namias (1966), however, presents some evidence that the decrease in relative humidity also can be caused by a lack of water vapor. He compiled a table of absolute-humidity data (grams of water vapor per kilogram of air) for Washington, D.C., for the drought of 1962-65, compared to a more normal period (1946-55) . The absolute-humidity values from the land surface to 10,000 feet during the drought were about 10 to 20 percent less than the values for the normal period during all seasons. At least in this case study, the air during the drought was drier throughout the lower atmosphere not only because the temperatures were higher but also because there was less water vapor in the atmosphere.



A drought is associated with persistent or persistently recurring atmospheric circulation patterns (Namias, 1985). For example, summer droughts in the southeastern United States generally are associated with the frequent recurrence of high-pressure (anticyclonic) circulations. Daily circulation patterns associated with a drought (such as the anticyclonic circulation of the preceding example) are not notably different from daily circulation patterns that occur during nondrought periods. A drought is associated with persistent or persistently recurring circulations that produce little or no precipitation and is not associated with any discernibly unique feature of the individual daily circulation patterns.

figure 3

Figure 3. Average height of 700-millibar pressure surface over North America for June 1933-52 and the anomalous height of the surface for October 1952. (Modified from Winston, 1952, 1953.) Click to view full-size image.

Three-dimensional depictions of the atmospheric circulation patterns at a single instant (synoptic maps at several constant-pressure levels) do not have features that are unique to droughts. Such depictions of atmospheric circulation patterns averaged for a month or more during a drought, however, do have anomalies, compared to climatological averages.

Maps of the October average height of the 700-millibar pressure surface and the anomalous height of the surface for October 1952 (fig. 3) and the June average height of the 700-millibar pressure surface and the anomalous height of the surface for June 1953 (fig. 4) are presented to show the position of atmospheric anomalies that occurred during the well- documented drought of 1952-54 (Klein, 1952a,b, 1953a,b; Winston, 1952, 1953; Namias, 1955). October 1952 was dry throughout the conterminous United States; all States except Florida had less than normal precipitation for the month. In June 1953, major precipitation deficits occurred predominantly in the south-central United States.

figure 4

Figure 4. Average height of 700-millibar pressure surface over North America for June 1933-52 and the anomalous height of the surface for June 1953. (Modified from Klein, 1952a,b; 1953a,b; Namias, 1955.) Click to view full-size image.

Just as the contours on a daily map of the 700- millibar pressure surface indicate the approximate atmospheric streamlines (because the atmosphere flows subparallel to the lines showing the altitude of the 700- millibar pressure surface at an altitude of about 10,000 feet for that day), monthly average contours of the 700-millibar surface (figs. 3 and 4) indicate the average streamlines for the month. Similarly, the contours of anomalous height in figures 3 and 4 indicate anomalous streamlines for the 2 months, compared to a long-term average for those months. Streamlines on the 700-millibar pressure surface represent the approximate low to midtropospheric flow and, consequently, the approximate source(s) of moisture in situations that are not markedly baroclinic (that is, situations in which the streamlines based on the contour patterns do not change direction with height).


The most prominent features in figure 3 are the abnormally strong ridge (that is, local high- pressure area) over the mountains of North America and the deep trough (that is, local low-pressure area) over the eastern Pacific Ocean and eastern North America. Anomalies showing the difference between the average height of the 700-millibar pressure surface for October 1952 and the average height of the surface for October during the previous 20 years have been plotted as gray dashed lines. The close spacing and north-south orientation of the anomalies indicate an abnormally strong flow from the north during October 1952. This flow is consistent with the recurrent movement of dry air from Canada southward into the central United States, shown on daily weather maps (Winston, 1952). In addition, the abnormally deep trough over eastern North America was the result of daily patterns that had minimal flow components from the Gulf of Mexico (except over Florida), which means the northward transport of moist air from the gulf was inhibited. Klein (1953b), in his analysis of the circulation for September 1953, presents a similar discussion of how an abnormally strong monthly average western ridge and eastern trough are the result of circulations that cause the southward movement of dry polar air into the United States and that inhibit the northward movement of moist air from the gulf.

The average atmospheric circulation pattern at the height of the 700-millibar pressure surface for 1953 (fig. 4) was quite different from the cool-season June circulation pattern typified by October 1952 (fig. 3). Instead of the abnormal ridge in the west and trough in the east, the circulation pattern for June 1953, which is typical for summer months, was characterized by a much stronger than normal anticyclone (that is, local high pressure area surrounded by closed height contours) over the southern United States plus troughs along the West and East Coasts.

The abnormally strong trough along the West Coast that accompanied the anticyclone was associated with frequent advection of air from the hot, dry desert regions in the Southwest into the central part of the country (Winston, 1953). Also, the abnormally fast westerly component of the air flow along the northern part of the anticyclone is an indication that cool air from Canada did not move southward very far into the United States. The advection of cool air in a warm season produces lifting of the warm air, which is favorable for precipitation. The absence of cool, destabilizing air from the north means frontal showers occurred less frequently than usual.


Large-scale vertical movement of air is a major factor in the occurrence of precipitation. Air that ascends over a large region favors precipitation, whereas air that descends over a large region inhibits precipitation. Namias (1983) identifies persistent and persistently recurrent descending air as the "immediate drought-producing mechanism'' in the sense that descending air inhibits precipitation. When air descends it may be adiabatically compressed, which will increase the temperature. The vertical profile of temperature usually is such that the descent causes midtropospheric temperatures to increase more than lower tropospheric temperatures. As a result, descending air increases the vertical stability of the atmosphere. The increase in temperature also decreases the relative humidity. The increased vertical stability and the decreased relative humidity tend to inhibit precipitation.

Vertical movements of air, such as descending air, are related directly to the quasi-horizontal pressure-surface patterns that appear on daily synoptic charts. During dry, cloudless days, a diagnosis of the vertical flow required to balance the horizontal flow indicated by the pressure-surface contours commonly indicates a downward movement of air associated with midtropospheric ridges and anticyclones. Although the relations between downward movement of air and ridges and anticyclones are based on atmospheric dynamics applicable to synoptic charts for individual days, similar relations are observed for monthly average pressure-surface patterns.

Winston (1952) associates the abnormally strong western ridge for October 1952 with descending air over the western United States. Similarly, Namias (1978b) relates the winter drought of 1975-76 in California to descending air beneath an abnormally strong winter-season ridge in the western United States. The large, upper-level anticyclone that occurred during the summers of 1952-54 (for example, see fig. 4) produced widespread downward movement of air that inhibited precipitation throughout much of the United States (Klein, 1952a; Namias, 1955).


Circulation patterns that inhibit precipitation can be described in several ways. Upper-atmosphere pressure-surface charts show the importance of descending air, dry-air advection, and the absence of destabilizing temperature gradients during a drought. Similar or equivalent information in terms of storm tracks, frontal passages, and anticyclone frequency also can be presented.

Klein (1952a,b) presents diagrams of the frequency of anticyclone passages during June and July 1952 that are consistent with the occurrence of a strong anticyclone pattern in monthly average-pressure surfaces (for example, fig. 4). Similarly, the strong ridge over the western United States during the drought of January 1976 indicated that storms were following a track northward to Alaska rather than eastward toward the West Coast (Namias, 1985). Although such storm tracks are useful to highlight or clarify the importance of a particular aspect of an atmospheric circulation pattern, they do not indicate the basic causes of atmospheric circulations any more than does the use of synoptic pressure surfaces.

Namias (1966, 1978a, 1982b) carefully points out that trough and ridge patterns, storm tracks, frontal passages, and so forth are descriptions, not causes, of atmospheric circulations associated with less than normal precipitation. The underlying question is: What causes these drought-related atmospheric circulations, and why do they persist for months and sometimes years?



Recurrent periods of less than normal precipitation are best described as climate anomalies, not weather anomalies. The distinction is useful because it suggests an appropriate perspective for understanding the atmospheric drought signal. A climatic perspective on drought is a global, or at least a hemispheric, perspective that includes interactions between the atmosphere and its ocean and land boundaries. A search for the causes of a drought cannot be restricted to the atmosphere above the area affected by a drought. The correlation of regional atmospheric-circulation features across global distances means that anomalous circulations or boundary conditions can be manifested at great distances. These long-distance connections can, of course, interact with feedback mechanisms within an area affected by a drought.

A climatic perspective also means that a drought needs to be understood in terms of time and space statistics. A knowledge of the causes of a drought cannot provide an a priori description of the specific sequence of dry and wet days or any other similar daily details about a drought. Such day-to-day changes in precipitation, temperature, wind, and so forth are defined as weather, not climate. Weather is not predictable, even in principle, for periods longer than about 2 weeks (Thiele and Schiffer, 1985). As a consequence of this limit on predictability, the causes of drought, when discovered, will not define the actual sequence of dry and wet days during a drought.

Finally, a climatic perspective is consistent with the monthly, seasonal, and sometimes longer persistence and quasi-periodicity associated with droughts. Such persistence and quasi-periodicity can result from external causes, such as solar perturbations, or from climate interactions that are not dependent on external causes, or on both.

Pittock (1978, 1983) has reviewed extensively the literature on solar variability and climate. He concludes that the approximately 20-year drought cycle in the western United States is related jointly to the 22-year Hale double sunspot cycle and the 18.6-year lunar nodical cycle. He regards the evidence as merely indicative, however, because the signals are difficult to identify in the climate record. Furthermore, detailed causal mechanisms connecting the extraterrestrial variations to the atmosphere have not been determined.

Climate models that exclude the types of external factors mentioned above can still simulate quasiperiodic and even chaotic conditions of the general nature associated with climatic disturbances, such as drought. This complexity is due to the nonlinear terms in governing equations that are used in simulating interacting physical processes within the atmosphere (Lorenz, 1964; Hunt, 1988). Gordon and Hunt (1987) were able to simulate droughts by using a 10-year integration of climatic data in a general circulation model, and they suggested that the droughts were a consequence of nonlinear, dynamic interactions.

Anomalous surface-boundary conditions commonly have been cited as being principal factors in causing and maintaining a drought. Although this approach avoids the question of what caused the boundary anomaly, it provides an important understanding of the global aspects of drought-producing processes.


Oceans exchange energy with the atmosphere via evaporation and turbulent transfer of sensible heat. The atmosphere and ocean temperature difference or sea-surface-temperature (SST) anomaly is an important controlling factor in these exchanges. Consequently, extensive and persistent SST anomalies (typically measured to be 2 to 3 degrees Fahrenheit throughout large areas for several consecutive months) can produce substantial variations in atmospheric heating. Variations in sea-ice boundaries also affect the atmosphere and ocean heating but to a much lesser degree than do SST anomalies because of the much smaller areas involved.

Namias (1978a) discusses the winter drought of 1976-77 in the western United States as an example of how an anomalous SST pattern contributes to the maintenance of a drought. The 700-millibar pressure surface for January 1977 had a strong ridge over the West Coast that was associated with descending air over the western United States and storm tracks that generally were in a more northerly direction toward Alaska, instead of southerly toward the coast of California.

figure 5

Figure 5. Sea-surface-temperature anomalies over the central and eastern parts of the North Pacific Ocean, winter 1977 (December 1976 through February 1977). (Modified from Pitcher and others, 1988). Click to view full-sized image.

The SST anomalies for the winter of 1976-77 (fig. 5) included a large, cold SST anomaly over the central part of the North Pacific Ocean and a warm SST anomaly just off the coast of California. Once established, the west-to-east SST-anomaly gradient helped maintain a southerly atmospheric flow that resulted in the circulation pattern for January 1977 that was characterized by the strong ridge over the West Coast. This effect of the ssT-anomaly gradient on the atmosphere is based on the assumption that the SST- anomaly gradient was transmitted to the overlying atmosphere and enhanced the high pressure over the warm SST anomaly and the low pressure over the cold SST anomaly.

Namias (1978a) also discusses the drought- causing ridge over the West Coast as part of the Pacific Ocean-North American (PNA) teleconnection. A teleconnection is a correlation between contemporaneous meteorological features at widely separated points over the surface of the Earth. The negative correlation between the height of the 700-millibar pressure surface over the West Coast and the height of the 700-millibar pressure surface over the central part of the North Pacific Ocean is an example of the PNA teleconnection. The 700 millibar pressure surface over the extratropical Pacific Ocean also is correlated to the underlying SST anomaly; that is, an estimate of the height of the 700-millibar pressure surface can be obtained from the underlying SST anomaly by using a regression equation.

These relations apply to the winter of 1976-77 as follows. A strong, cool SST anomaly was present over the central part of the North Pacific Ocean. An application of the correlation between the SST anomaly and the height of the 700-millibar pressure surface shows an anomalously low height for the surface above the SST anomaly. Because the height of the 700- millibar pressure surface over the central part of the North Pacific Ocean is negatively correlated to the height of the 700-millibar pressure surface over the West Coast, the anomalous low height of the surface over the central part of the North Pacific Ocean caused the anomalous strong ridge of the surface over the West Coast via the PNA teleconnection. Namias (1978a) emphasizes the importance of using numerical models to simulate and further understand this complex set of relations.

Voice and Hunt (1984) used a global general- circulation model to study the dynamics of droughts as a response to a specified tropical SST anomaly. Although they specified an unrealistically extensive SST anomaly in the tropics to focus on droughts in Australia, their discussion illustrates some of the mechanisms that are important in describing atmospheric and ocean relations. The experiment showed that the SST anomaly changed the evaporation rate above the anomaly. Evaporation anomalies then began to perturb the atmospheric circulations because of their effect on precipitation; that is, because of latent heating of the atmosphere. The contribution of sensible-heating anomalies due to the SST anomaly was-much less than the contribution due to latent heating. Significant perturbations of the atmospheric circulation occurred over distant parts of the globe and produced areas of both increased and decreased precipitation. Voice and Hunt (1984) noted that the results can be sensitive not only to the characteristics of the SST anomaly but also to the season in which the anomaly develops and to perturbations occurring at the time the anomaly develops.

Pitcher and others (1988) studied the effects of SST anomalies on atmospheric circulation during January 1977 by using a general-circulation model. One set of their experiments consisted of specifying SST anomalies that were plus and minus one and two times the midlatitude SST anomaly discussed by Namias (1978a; fig. 5). They determined that the model response to the specified cold SST anomaly over the central part of the North Pacific Ocean and the specified warm SST anomaly over the eastern part of the North Pacific Ocean was the PNA teleconnection pattern; that is, the pattern associated with the drought- causing ridge over the West Coast. Their experiment with a specified warm SST anomaly over the central part of the North Pacific Ocean and a specified cold SST anomaly over the eastern part of the North Pacific Ocean (minus one and two times the temperature of the basic anomaly) did not produce a PNA pattern of the opposite sign. Because of this perplexing result, they suggest that caution is warranted in studies of correlations between SST anomalies and midlatitude upper-air pressure-surface anomalies.


Just as SST anomalies are important in determining the atmospheric-heating variations due to the ocean surface, soil-moisture anomalies are important in regulating the evapotranspiration and, consequently, the latent heating over land. Variations in snow cover also are an anomalous boundary-forcing mechanism over land, because of the difference in reflectance between snow- and nonsnow-covered ground.

Because deficits in precipitation directly affect the soil moisture and snow cover, there is the possibility of a feedback (that is, reinforcing interaction) between drought-related land-surface conditions and the atmospheric circulation producing the drought. Although definitive evidence of this feedback has not been established, several studies have been performed to help determine its importance.

The emerging consensus is that land-surface anomalies are less important in short-term climatic fluctuations than are SST anomalies (Walsh, 1986). Three factors limit the relative importance of land- surface anomalies compared to SST anomalies (Walsh and others, 1985). First, soil-moisture or snow-cover anomalies typically are smaller in area than are SST anomalies. Second, they are much less persistent than are SST anomalies. Third, the substantial moisture content and weak vertical stability over the ocean favor redistribution of the SST anomaly throughout a thick layer of the atmosphere, which enhances the global effect of the SST anomaly. In contrast, the lesser effects of the land-surface anomalies on the upper atmosphere limit their effects to local areas.

Namias (1960) investigated the effect of soil moisture by constructing a contingency table that related temperature and precipitation during the spring in the Great Plains to the temperature and precipitation during the following summer. The analysis showed that a warm spring is more likely to be followed by a dry summer than by a wet summer by a ratio of 87:58. When precipitation during the spring was included as one of the categories, the results were that a warm and dry spring was more likely to be followed by a dry summer than by a wet summer by a ratio of 49: 14. Namias (1960) suggested that the dry soil developed during dry springs enhances the persistence of upper-level anticyclones associated with warm, dry summers.

More recently, Walsh and others (1985) and Karl (1986) also have obtained evidence that moisture anomalies are associated with subsequent land-surface temperature anomalies. Van den Dool and Klein (1986) determined that soil-moisture anomalies are weakly associated with subsequent anomalies of the 700-millibar pressure surface.

Several numerical experiments (Gilchrist, 1982; Rind, 1982; Shukla and Mintz, 1982) have been made to study the effect of anomalously dry land-surface- boundary conditions on subsequent precipitation patterns. The basic result of these experiments is that, when soil moisture is set and held at an anomalously small value, the subsequent precipitation that falls in the area affected by the anomaly is decreased. These results, however, need to be used with caution because of the unrealistically large soil-moisture deficits that were specified-Gilchrist (1982) set the soil moisture to zero throughout all of Europe; Rind (1982) set the soil moisture of the entire United States to one-fourth of a control value; and Shukla and Mintz (1982) decreased the evapotranspiration to zero on all land surfaces throughout the world. Gilchrist (1982) reported that not only would precipitation throughout Europe be decreased, but also the deficit would extend into North Africa. Gilchrist (1982) and Rind (1982) mentioned that the results may be dependent on the initial conditions into which the moisture anomaly was introduced.


Precipitation anomalies are a naturally recurring feature of the global climate. These anomalies affect various components of the hydrologic cycle to produce a drought. Climatologies of precipitation, temperature, and atmospheric moisture provide an indication of the frequency and intensity of precipitation, the correlation of precipitation and temperature, and the atmospheric drying that occurs during droughts.

Climatologies of atmospheric circulation patterns illustrate that drought is associated with persistent or persistently recurring circulation patterns that produce little or no precipitation; a drought does not occur as a result of discernibly unique daily circulation patterns. Monthly average circulation patterns indicate the importance of descending air, dry-air advection, and the absence of destabilizing temperature gradients during a drought. These factors, however, only provide descriptions of the atmosphere during a drought. They do not answer the underlying question: Why do these drought-related circulations arise, and why do they recur more frequently than in normal years?

The search for causes of atmospheric drought signals goes far beyond the immediate area affected by drought, because of the global (or at least hemispheric) nature of atmospheric circulations that produce sustained periods of less than normal precipitation. Many current investigations are focused on anomalous conditions at the boundaries of the atmosphere with the ocean and the land surface, such as SST and soil- moisture anomalies. Nonlinear processes within the climate system, as well as external solar and lunar variations, are being investigated as basic causes of drought.


Baldwin, J.L., 1957, Drought and cloud seeding: U.S. Department of Agriculture, Weekly Weather and Crop Bulletin, January 10, p. 8.

Bergman, K.H., Ropelewski, C.F., and Halpert, M.S., 1986, The record southeast drought of 1986: Weatherwise, v. 39, no. 10, p. 262-266.

Borchert, J.R., 1971, The Dust Bowl in the 1970s: Annals of the Association of American Geographers, v. 61, no. 1, p. 1-22.

Chang, F.C., and Wallace, J.M., 1987, Meteorological conditions during heat waves and droughts in the United States Great Plains: Monthly Weather Review, v. 115, no. 7, p. 1253-1269.

Changnon, S.A., 1980, Removing the confusion over droughts and floods-The interface between scientists and policy makers: Water International, v. I, no. 10, p. 10-18.

___ 1987, Detecting drought conditions in Illinois: Illinois State Water Survey Circular 164-87, 36 p.

Cry, G.W., 1967, The effects of tropical cyclone rainfall on the distribution of precipitation over the eastern and southern United States: U.S. Department of Commerce, Environmental Science Services Administration Professional Paper 1, 67 p.

Gilchrist, A.R., 1982, Aspects of the simulation of climate and climate variability in middle latitudes: Physical Basis for Climate Prediction, World Meteorologic Organization/lSCU Study Conference, Leningrad, WCP-47, p. 129-150.

Gordon, H.B., and Hunt, B.G., 1987, Interannual variability of the simulated hydrology in a climatic model- Implications for drought: Climate Dynamics, v. 1 no. 1, p. 113-130.

Hare, F.K., 1987, Drought and desiccation-Twin hazards of a variable climate,in Whilhite, David, Easterling, William, and Wood, David, eds., Planning for drought: Boulder, Colo., Westview Press, p. 3-9.

Huff, F.A., and Changnon, S.A., 1963, Drought climatology of Illinois: Illinois State Water Survey Bulletin 50, 68 p.

Hunt, B.G., 1988, Nonlinear influences-A key to short-term climatic perturbations: Journal of the Atmospheric Sciences, v. 45, no. 3, p. 387-395.

Karl, T.R., 1986, The relationship of soil moisture parameterizations to subsequent seasonal and monthly mean temperature in the United States: Monthly Weather Review, v. 114, no. 4, p. 675-686.

Karl, T.R., and Quayle, R.Q., 1981, The 1980 summer heat wave and drought in historical perspective: Monthly Weather Review, v. 109, no. 10, p. 2055-2073.

Karl, T.R., and Young, P.J., 1987, The 1986 Southeast drought in historical perspective: Bulletin of the American Meteorological Society, v. 68, no. 7, p. 773-778.

Karl, T.R., Quinlan, Frank, and Ezell, D.S., 1987, Drought termination and amelioration-Its climatological probability: Journal of Climate and Applied Meteorology, v. 26, no. 9, p. 1198-1209.

Klein, W.H., 1952a, The early summer drought of 1952: Weatherwise, v. 80, no. 10, p. 111-113.

___ 1952b, The weather and circulation of June 1952: Monthly Weather Review, v. 80, no. 6, p. 122.

___ 1953a, The weather and circulation of August 1953: Monthly Weather Review, v. 81, no. 8, p. 246-254.

___ 1953b, The weather and circulation of September 1953: Monthly Weather Review, v. 81, no. 9, p. 304-308.

Klemes, V.E., 1987, Drought prediction-A hydrologic perspective, in Wilhite, David, Easterling, William, and Wood, David, eds., Planning for drought: Boulder, Colo., Westview Press, p. 81-94.

Lorenz, E.N., 1964, The problem of deducing the climate from the governing equations: Tellus, v. 16, no. 1, p. 1-11.

Mitchell, J.M., 1968, Further studies of drought over the northeastern United States: Conference on the Drought in the Northeastern United States, New York, N.Y., New York University, Proceedings, no. TR-68-3, p. 129-159.

Namias, Jerome, 1955, Some meteorological aspects of drought with special reference to the summers of 1952-54 over the United States: Monthly Weather Review, v. 83, no. 9, p. 199-205.

___ 1960, Factors in the initiation, perpetuation and termination of drought: International Association of Scientific Hydrology Commission on Surface Waters Publication 51, p. 81-94.

___ 1966, Nature and possible causes of the northeastern United States drought during 1962-65: Monthly Weather Review, v. 94, no. 9, p. 543-554.

___ 1968, Further studies of drought over the northeastern United States: Conference on the Drought in the North eastern United States, New York, N.Y., New York University, Proceedings, no. TR-68-3, p. 57-94.

___ 1978a, Multiple causes of the North American abnormal winter 1976-77: Monthly Weather Review, v. 106, no. 3, p. 279-295.

___ 1978b, Recent drought in California and western Europe: Reviews of Geophysics and Space Physics, v. 16, no. 3, p. 435-457.

___ 1982a, Anatomy of Great Plains protracted heat waves (especially the 1980 U.S. summer drought): Monthly Weather Review, v. 110, no. 1, p. 824-838.

___ 1982b, Case studies of long period air-sea interaction relating to long-range forecasting: Physical Basis for Climate Prediction, World Meteorologic Organization/lSCU Study Conference, Leningrad, WCP-47, p. 293-325.

Namias, Jerome, 1983, Some causes of United States drought: Journal of Climate and Applied Meteorology, v. 22, no. 1, p. 30-39.

___ 1985, Hydrologic aspects of drought, in Beran, M.A., and Rodier, J.A., eds., Unesco-World Meteorological Organization, A Contribution to the International Hydrologic Programme, Paris, France, p. 27-64.

Pitcher, E.J., Blackmon, M.L., Bates, G.T., and Munoz, S.R., 1988, The effect of North Pacific Ocean sea surface temperature anomalies on the January climate of a circulation model: Journal of Atmospheric Sciences, v. 45, no. 2, p. 173-188.

Pittock, A.B., 1978, A critical look at long-term sun-weather relationships: Reviews of Geophysics and Space Physics, v. 16, no. 3, p. 400-420.

___ 1983, Solar variability, weather, and climate-An update: Quarterly Journal of the Royal Meteorological Society, v. 109, no. 462, p. 23-55.

Potter, J.G., 1958, An unusually dry spring in central Canada: U.S. Department of Agriculture Weekly Weather and Crop Bulletin, July 7, p. 8.

Riehl, Herbert, 1965, Introduction to the atmosphere: McGraw-Hill, New York, N.Y., p. 285-292.

Rind, David, 1982, The influence of ground moisture conditions in North America on summer climate as modeled in the GISS GCM: Monthly Weather Review, v. 110, no. 10, p. 1487-1494.

Shukla, J.O., and Mintz, Yale, 1982, Influence of land- surface evapotranspiration on the Earth's climate: Science, v. 215, no. 4539, p. 1498-1500.

Spar, Jerome, 1968, The drought in the Northeast, 1962-1966: Conference on the Drought in the North eastern United States, New York, N.Y., New York University, Proceedings, no. TR-68-3, p. 1-31.

Thiele, Otto, and Schiffer, R.A., 1985, Understanding climate-A strategy for climate modeling and predictability research: National Aeronautics and Space Administration Reference Publication 1158, 31 p.

Van den Dool, H.M., and Klein, W.H., 1986, The geograpical distribution and seasonality of persistence in monthIy mean air temperatures over the United States: Monthly Weather Review, v. 114, no. 3, p. 546-56.

Voice, M.E., and Hunt, B.G., 1984, A study of the dynamics of drought initiation using a global general circulation model: Journal of Geophysical Research, v. 89, no. D p. 9504-9520.

Walsh, J.E., 1986, Surface-atmosphere interactions over the continents-The Namias influence: Namias Symposium, La Jolla, Calif., Scripps Institution Oceanography Reference Series 86-17, p. 121-131.

Walsh, J.E., Jasperson, W.H., and Ross, Becky, 1985, Influences of snow cover and soil moisture on monthly air temperature: Monthly Weather Review, v. 113 p. 756-768.

White, G.R., and Vaughan, H.C., 1982, Comparison of normal and below normal July precipitation patterns in central lowa: Journal of Climatology, v. 2, no. 9 p. 331-338.

Wilhite, David, and Glantz, M.R., 1987, Understanding the drought phenomenon-The role of definitions, in Wilhite, David, Easterling, William, and Wood, David, eds., Planning for drought: Boulder, Colo., Westview Press, p. 11-27.

Winston, J.S., 1952, The weather and circulation of October 1952: Monthly Weather Review, v. 80, no. 10, p. 190-194.

___ 1953, The weather and circulation of June 195 Monthly Weather Review, v. 81. no. 6, p. 163-168.

Link to USGS home page

Accessibility FOIA Privacy Policies and Notices

Take Pride in America home page. logo U.S. Department of the Interior | U.S. Geological Survey
Page Maintainer: Randy Schumann
Page Last Modified: Mon 5-Aug-2013 17:21:13 EDT