Droughts, naturally occurring events, have affected the northern Great Plains climate in short and long increments during the past 2000 years. Two of the most severe droughts to affect the northern Plains in the recent past occurred from 1988 to 1992 and during the ?Dust Bowl? in the 1930s. Recent studies, though, have now shown that these two periods of intense drought are not a rarity in the northern Great Plains climatic history. Lake salinity and tree ring data reveal that a series of "Megadroughts" occurred prior to AD 1200. These droughts encompassed decades and in some instances centuries at a time with very dry conditions. The intent of this paper is to construct a drought history for the northern Plains while studying their definition, length, and intensity and in some cases their formation processes.

There is no clear-cut definition of drought because it occurs in virtually every climatic zone and its characteristics vary significantly depending on the region. According the Drought Mitigation Center ( http://enso.unl.edu/index.html ), definitions are broken down into two different kinds: conceptual and operational. Conceptual definitions help people to understand the concept of drought in order establish drought policy for an area while defining the conditions needed for assistance to be awarded. The operational definitions try to help identify the beginning, end, and degree of severity through the use of climatology on an hourly, daily, monthly, or other time scale.

Stormfax ( http://www.stormfax.com/drought.htm ) and the Drought Mitigation Center break drought into four different types: meteorological, agricultural, hydrological, and socioeconomical. A meteorological drought is described as being a deviation from normal precipitation conditions over a period of time for a specific area. Severity is measured by the Palmer Drought Severity Index (PDSI) (Wilhite, 1987). Time scales may consist of hourly data all the way to annual data.

An agricultural drought is a result of inadequate soil moisture essential for certain crops to grow and thrive. An important part of an agricultural drought is the varying degrees of susceptibility of crops for different stages of crop development. The main severity index is the Crop Moisture Index (CMI) (Wilhite, 1987).

The effects of a hydrological drought are not as immediate as a meteorological or agricultural drought. It is defined as occurring when water supplies in streams, lake, rivers, and reservoirs are depleted. Its frequency and severity is defined on a watershed or river basin scale by the Surface Water Supply Index (SWSI) (Wilhite, 1987).

A socioeconomic drought is the negative affect that a physical water supplies deficit has on a community caused by a meteorological, hydrological, and agricultural drought. Its severity depends on the time and space processes of supply and demand for some economic good.

The one common thread in all of the definitions is that drought originates from a deficiency of precipitation over an extended period of time when water supplies are less than demand. The next dilemma is concerned with measuring a drought?s severity. Severity is aggravated by climatic factors such as high temperatures, high winds, and low relative humidity.

The PDSI is widely used to express the severity and longevity of a drought (Table 3). The PDSI takes into account current conditions as well as the recent past conditions in order to get a full scope of the effect that the duration of a drought or a wet spell has had on a region. PDSI values are then weighted on scale of 11 categories (Wilhite 1987).

The most recent severe drought in the Northern Plains history was the drought of 1988-92. It was characterized by an inverted U-shape drought area that was indicated by extremely dry conditions in the Northern Plains, Pacific coast, and the Southeast (Woodhouse 1998). The primary cause of this drought is explained as being changes in Sea Surface Temperatures (SSTs) in the tropical Pacific Ocean. Winds from the north caused the California Ocean current to carry cold air toward the equator. When the colder air mixed with the warmer water, the air quickly became warmer and was forced upward, which in turn became warmer and was also forced upward. This cold ocean current flow caused the lowering of SSTs, prompting a trend to strong anticyclonic flow in the upper troposphere over the continental United States in April of 1988 persisting through June of that same year (Trenbeth 1992). This change displaced the jet stream and storm tracks well to the north of its normal position while also weakening the jet stream over the western United States. The configuration diverted the moisture to the north while strong anticyclones blocked access to Gulf of Mexico moisture (Namias 1991, Woodhouse 1998). What followed was some of the driest and warmest conditions that the northern Plains, especially North Dakota, had seen since the ?Dust Bowl? in the 1930?s (Williams-Sether 1994). The dry conditions are evident in the PDSI map in Figure 2. According to the Weather Channel ( http://www.weather.com) , damages from the 1988-92 drought in the Northern Plains were estimated at $40 billion with over 5,000 directly related fatalities.

PDSI maps have been reconstructed by the use of instrumentation data back to 1895. PDSI maps have also been constructed on a year-by-year basis for tree ring data back to 1700 ( Figures 2 and 3 ). A complete series of PDSI maps can be found on the NOAA/NESDIS website ( http://www.ngdc.noaa.gov/paleo/drought.html ) . When the tree ring data reconstructions are compared to the instrumentation reconstructions, it is found that the tree ring reconstructions do not capture the severity of the drought as well. The duration of drought, however, is strongly correlated. For that very reason it is hard to compare the severity of droughts. Table 4 and Figure 4 show a list of droughts that have occurred since 1700 and their estimated severity as given by the constructed NOAA/NESDIS PDSI maps.

The "Dust Bowl" in the 1930's is considered to be the most extreme drought and longest in duration of recent history. It affected 2/3rds of the United States while producing a 40% reduction in corn and wheat yields. The biggest enemy of agriculturists was the strong winds out of the west and southwest. Dry, sandy topsoil was blown away by "dusters" and "black blizzards" and deposited it as drifts in feed stacks. Cattle died from starvation and suffocation as dust covered pastureland. Between 1930 and 1934 alone there was an estimated $5 billion loss in agriculture. Lake Michigan and Huron were also measured to be at their all time lowest levels during the late 1930?s (Tannehill, 1947).

Even though severity cannot be concluded from the available data, important trends can be seen. From Figure 5, evidence of an approximate cycle with a 20-year mean time interval of severe droughts can be seen. Causes of this cycle have not yet been determined.

Records of data for the Northern Plains prior to the AD 1700 are very sparse. Weakley reports droughts lasting from 1539-1564 and 1587-1605 from the reconstruction of tree ring data in southwestern Nebraska (Wedel 1986). These estimates are also verified by salinity data from Moon Lake (Laird 1996) and tree ring data from Arkansas (Stahle 1988). According to Weakley, possibly the most severe and persistent drought of the past 600 years occurred in a 38-year period between AD 1276 and AD 1313. It is considered the last in a chain of four "Megadroughts" to affect the Northern Plains in the past 2000 years.

Little climatic data for the Northern Great Plains prior to AD 1200 has been collected. One of the more extensive and thorough sets is the Moon Lake, North Dakota salinity data set. Laird reconstructed drought intensity and frequency for the past 2,300 years based on lake salinity fluctuations inferred from fossil diatom assemblages. The assumption that lower lake levels and therefore higher salinity would reflect negative water balance was made. The findings consistently agree with documentation of recent droughts since AD 1700, Weakely's reconstruction from tree ring data for southwestern Nebraska, and Fritts's (1965) reconstruction of the Western climate using tree ring data from the California Sequoia Forests.

Laird showed that droughts of greater frequency and intensity than the past 800 years occurred before AD 1200. It was shown that extreme droughts across the Northern Plains might have occurred from AD 200-370, AD 700-850, and AD 1000-1200 which directly agreed with the California Sequoia tree ring data. The AD 1000-12000 drought also corresponds to record of the Medieval warm period of European records.

Droughts have existed throughout the course of history, and will continue to exist in the future. Yet we know very little about what causes them. In order to learn more and how to plan for them in the future, we must first continue to construct drought histories. These data will provide a better "feel" for the conditions that brought on the onset of major droughts, while also being able to access the complex processes that end a drought.