Document Sample
acf-drought Powered By Docstoc
					  Analysis of Precipitation Variability and Meteorological
Drought in the Apalachicola-Chattahoochee-Flint River Basin


                      Gloria Arrocha

                 Department of Meteorology
                  Florida State University
                 Tallahassee FL 32306-4520

           Supervising Professor: Paul H. Ruscher

         The water flowing through Apalachicola-Chattahoochee-Flint River (ACF) River
basin is used by several cities, industries, and farms, and for wastewater dilution,
navigation for barge traffic, recreational boating and fishing, wildlife conservation and
power generation. Droughts result in adverse effects for water users in this basin. For
example, the drought of 1980-1981 caused a reduction in hydroelectric power generation,
the curtailment of navigation, reduced lake levels for recreation, and restrictions on lawn
watering and other water uses, according to U. S. Army Corps of Engineer reports.
Besides, recent water-use conflicts between the states over which the ACF River Basin
flows have made clear how important it is to monitor weather effects on stream levels.
         This study collects and analyzes cooperative and first-order precipitation
observations of stations within the ACF river basin in order to assess significant droughts
experienced in the past. Observed and replaced missing data went all the way back to
1895, giving the study a long period of time to analyze precipitation climatology. The
main emphasis will be to describe the ACF river basin’s precipitation patterns for the last
century. This includes a climatology definition or normal annual precipitation value for
each station in the basin from 1931 to 1980. The final result will show how these stations
patterns differ or relate throughout the basin. We hope to gain a perspective on the
frequency and nature of drought in the ACF basin, in order to help inform the debate and
controversy about water use. In summary, basin-wide prolonged droughts of several
years duration are frequent and are likely to significantly affect long-term decisions
regarding water allocations in the region. There are serious political and legal
repercussions that will soon be discussed in Federal courts that will affect users in three
states, as well as stakeholders in industries along the Gulf of Mexico coast.

I. Introduction

       The Apalachicola-Chattahoochee-Flint (ACF) River basin covers approximately

50,800 km2 (United States Army Corps of Engineers [USACE] 1984) and drains parts of

Alabama, Florida and Georgia. Most of the Chattahoochee and Flint River basin

(headwaters) are contained in Georgia, north of Lake Sydney Lanier. They stretch south

to Florida to form the Apalachicola River, which discharges its waters into the Gulf of

Mexico (Figure 1). The water flowing through the ACF River basin is used by several

cities, industries, and farms, and for wastewater dilution, navigation for barge traffic,

recreational boating and fishing, wildlife conservation and power generation. (Marella et

al. 1993; Figure 2)

       As the ACF river basin’s streams flows to the south, they cross the boundaries of

different climate regions, created by the National Climatic Data Center.          These

boundaries are used to group and study different climatic variables such as temperature,

rainfall and humidity, and then determine averages, minimums and maximums, and

declare any weather extremes, such as droughts, for these areas.

        Droughts adversely affect water users in this basin. For example, the drought of

1980-1981 caused a marked reduction in hydroelectric power generation, the curtailment

of navigation, reduced lake levels for recreation, and restrictions on lawn watering and

other water uses (USACE 1984). Besides, recent water-use conflicts between the states

(Yates 2003) through which the ACF River Basin flows have made clear how important

it is to monitor weather effects on stream levels. Water allocation rights between states

and stakeholders are now issues before Federal courts, in one of the first major disputes

of this kind in the eastern United States.

       The purpose of this study is to examine the frequency of meteorological drought

(using historical precipitation data) for the ACF basin. Rainfall data have been recorded

for 108 years through different agencies and cooperative observers throughout Florida,

Georgia and Alabama. The National Weather Service stations and cooperative observer

stations have made their information available to the public through the National Climatic

Data Center (NCDC). For this study, monthly precipitation data for the past hundred

years have been retrieved for the stations in the ACF Basin. Other information and data

about stream flow, river basin boundaries and pertinent material have been obtained

through the studies and databases of the U.S. Geological Survey (USGS).

       This study collects and analyzes cooperative and first-order precipitation

observations within the ACF River basin in order to assess significant droughts

experienced in the past. Observed data spanned 1895 through 2002, providing a long

period of record. We hope to gain a perspective on the frequency and nature of

meteorological drought in the ACF basin, in order to help inform the debate and

controversy about water use.

       Our reference climatology for the calculation of anomalies and departures from

normal was made using data from 1931 to 1980, the period that had the best dataset in

terms of small amounts of missing data. The entire dataset was analyzed using a 50-year

climatology because we wished to use a longer period than the traditional 30 years here;

however, 1931 also marks the beginning of the time when many stations have easily

retrievable data from NCDC in the same format. Precipitation distribution charts were

made for each station to find what type of distribution they had. Another set of charts was

prepared for the precipitation anomalies, defined as the departure from our 50-year

climatology for each station. Other statistical figures, like standard deviation and

normalized values, similar to a precipitation index, were calculated to help organize and

analyze the dataset. The 10th, 25th and 50th percentiles of observed annual total events

were calculated, and they were used to determine the degree of drought intensity.

Because the distributions did not illustrate normal distribution signatures, we will use the

median and measures of spread to define drought, rather than just a certain standard

deviation below the mean.

       We seek to define how dry any year was compared to our long-term climatology.

These calculations identify drought occurrences and separate those events that have

relatively longer and more intense dry conditions from the shorter, less intense ones. Our

study defines prolonged meteorological drought periods as those of three or more

consecutive years of precipitation that is below normal, that also had at least one year

with precipitation measured below the 25th percentile for that station. This threshold was

used because we first want to determine how common are these prolonged droughts,

especially those with at least one particularly dry year, and if these prolonged events have

any notable time and space distribution over the river basin.

       These prolonged droughts, with one or more very dry years included among

several dry years in a row, tend to have a different type of impact on many agricultural

and economical aspects than shorter, episodic, annual droughts. This is because longer

droughts would require a greater logistical plan of action to avoid larger economic losses

and to be able to recover from them faster than short droughts would. There is more

likelihood of substantial impact on agriculture, recreation, fisheries, forest health (and fire

danger), and water quality due to low water levels and low soil moisture values during

  these times. Also, people affected the most by these events probably have shaped their

  decision making according to the empirical knowledge of climate they have acquired

  over the years. They might have established agricultural and management practices

  according to the climate for the past 20 or 30 years (or even much shorter periods of

  time), but without knowing how much the weather could or has changed over the past

  decade compared to previous decades. It might be helpful to either corroborate their

  practices or alert them for any possible changes in the climate of the river basin.

II. Chattahoochee-Apalachicola-Flint river basin under study

         Because of its location, many important urban and agricultural areas take

  advantage of the waters in the ACF drainage basin to provide power generation, irrigation

  and waste disposal. For this reason, there have been several important studies conducted

  on this river basin. A few have served as sources of information for the river basin’s

  physiography presented in this paper. Some of the previous studies and papers deal with

  environmental issues, and concentrate on studying the different environmental settings

  throughout the basin; others have studied the river basin’s water flow and discharge.

  They take only a general look at precipitation data; their main focus takes into account a

  broad band of aspects that diminish their attention on studying rainfall.

         The Apalachicola-Chattahoochee-Flint Rivers drain approximately 50,800 km2 in

  Georgia, Alabama and Florida (Figure 3). The Chattahoochee River is born in northern

  Georgia, and it flows 430 miles towards the South, serving as a boundary line between

  Georgia and Alabama. The Flint River originates south of Atlanta and travels 350 miles

  through part of the state before it joins its waters with the Chattahoochee River (Couch et

   al., cited 2004). Because of rainfall distributions, the average annual discharge from the

   Chattahoochee River exceeds that of the Flint River. The Chattahoochee River also

   makes a greater contribution to peak flows in the Apalachicola River than the Flint River.

   However, during extremely dry periods, the greater flow contribution comes from the

   Flint River, whose base flow is sustained by ground water discharges (Elder et al. 1988).

   It is estimated that aquifer discharges to the Chattahoochee River are one-fifth of the

   amount that aquifers discharge to the Flint River (Torak et al. 1991). Then, they drain

   their waters into Lake Seminole, which in turn feeds the Apalachicola River. The

   Apalachicola River crosses a short extension of 106 miles over northern Florida before it

   sheds its waters in Apalachicola Bay and the Gulf of Mexico (Couch et al., cited 2004).

   Eighty percent of the Apalachicola River’s flow is contributed by the Chattahoochee and

   Flint River, 11 percent by the Chipola River and less than 10 percent from ground water

   and overland flow (Elder et al. 1988). Some recent preliminary work suggests that the

   past 25 years have suffered frequent periods of very low flow (Figure 4) (Light 2004,

   unpublished data). It is also important to mention that other factors such as agriculture

   and urban population growth might be having a greater influence in these low flow

   events, and that they might have a greater influence on these dry conditions than lack of

   rainfall does (Leitman et al. 2003).

III. Description of methodology

          Data were collected from the cooperative and first-order observation stations

   within the Chattahoochee-Flint-Apalachicola river basin from NCDC, which amounted to

   a total of 122 stations. However, few stations had consistent observations throughout a

long period of time, but some had information that went back up to the 1890’s. This long

period of record, even with its inconsistencies in data completeness, still provides

valuable information, which may be used to analyze the river basin’s climatology.

       In the selection process, the only stations kept for the study were those that had at

least 30 years of complete data. Of the 122 stations present on the original dataset, only

50 stations complied with the minimum. As was mentioned before, many of these

selected stations had data that went back to the 1890’s, but these years outside the 30

minimum had gaps of one or more (partial or full) years throughout their records.

Because of these data gaps, it was necessary to create a way of filling the missing data by

using those stations that had valid data around that area (of course radar estimates are not

available for most of our study period). Then, a method for completing the desired

longer period of record using complex interpolation techniques was required.

       A helpful tool developed by the Spatial Climate Analysis Service at Oregon State

University was used to complete the few years that made some stations’ data inconsistent

throughout time. This tool is a computer model called PRISM (Parameter-elevation

Regression on Independent Slopes Model). PRISM is a knowledge-based system that

uses point data, a high-resolution digital elevation model (DEM), and many other

geographic data sets to generate gridded estimates of monthly and event-based climatic

parameters (Daly et al. 2004). PRISM has been used extensively to map precipitation,

dew point, and minimum and maximum temperature over the United States, Canada,

China and other countries (USDA-NRCS 1998, Daly et al. 2004). The method used by

this software was strongly appealing and parallel to the study’s plan for completing the

data, because it “can provide time series of spatially distributed precipitation for a variety

  of watersheds and drainages at time intervals ranging from years to single events, serving

  as high quality input” (Daly et al. 2004). Another appeal for using the model was that

  part of its data came from the same cooperative observer stations used by this study.

         Overall, the reliability of the model has undergone extensive evaluation ever since

  it became available, and it has been used for hydrological modeling and related scientific

  studies. Several of our stations where chosen to test the model’s ability to create

  interpolations that closely resembled real observations, and some of these are shown in

  Figure 5. The program appeared to very reliably replace missing data with realistic

  values in a very consistent manner, so the PRISM program was used to fill in missing

  observed data, when monthly accumulations were missing from the NCDC dataset.

  More details can be found on the PRISM software’s web page which address is indicated

  in the references page (Oregon State University 2004).

IV. Climatology

         The most reliable period of observations occurred between the 1930’s and 1980’s.

  For this reason, our normal climatological values were calculated using the 50 years

  between 1931 and 1980 (Figure 6). In particular, many stations had reliable data

  beginning in 1931 in common formats.

         First, the climatological normal or mean values were calculated using data within

  the 50-year range, together with the standard deviation. The precipitation anomalies were

  then graphed by calculating how many standard deviations, above or below normal, the

  total annual precipitation was for a station on each given year (Figure 7). The normal

  values would help distinguish dry periods from wetter ones at the same time that it would

  make dry year trends easier to spot.

         The    entire    set   of    plots   are    available     on    a   web    site     at The next step was to calculate different

  percentiles for the stations. The 10th, 25th and 50th percentiles were calculated, and lower

  decile and lower quartile were used as initial criteria for determining significantly dry

  years. Determining drought periods just using the 10th percentile would not necessarily

  account for all years that could have been dry enough to have a major impact on the river

  basin, and any several year-long droughts that might have had a significant impact as

  well. The 25th percentile appeared to be a more acceptable threshold to use for

  determining multiple years of drought by cutting out below-normal-precipitation years

  that were too close to climatological average, while maintaining yearlong drier periods.

         The next step of this study was to analyze how many stations were in a drought

  period for each year. This would show if there is any pattern to the drought distribution

  in the river basin and if different stations coincide with similar drought periods.

  Furthermore, this classification would tell how recurrent were basin-wide and regional

  droughts for the last one hundred years. Years that had a large number of stations

  recording values below the chosen percentile will also be pointed out in this study.

V. Significant Droughts

         Table I summarizes groups of drought periods by the number of stations that

  suffered drought conditions. The 10th percentile and 25th percentile were calculated here

  to demonstrate the viability of our chosen threshold. At the lower decile, a large number

       Table I: Years with the number of stations recording below the 10th and 25th percentile values,

                   respectively, based on the sample size of 50 stations in the ACF Basin.

Year       Total Stations         Total stations
        below 10th percentile below 25th percentile
1899             16                    50
1904             21                    50
1910             15                    50
1921             19                    50
1925             18                    50
1927             20                    49
1931             20                    50
1933             14                    48
1935             10                    42
1938             18                    48
1941             11                    44
1950             15                    49
1954             20                    50
1955             17                    47
1968             18                    47
1981             13                    46
1986             12                    46
1990             9                     45
1999             16                    48
2000             19                    49

              *Boldface font indicates years which fall outside multiple-year drought periods

of years are present in which there do not appear to be basin-wide impacts (based on the

number of stations meeting the harsh threshold). The lower quartile, however, appears to

adequately represent the idea of basin-wide drought, as evidenced by the large number of

years in which all stations are in rainfall deficits. Therefore we adopt the 25th percentile

as the threshold that must be met at least once during a multiyear period to define a

prolonged drought; that is at least one of three consecutive years must have an annual

precipitation value which is no more than the 25th percentile. A more robust statistical

threshold could be applied to more adequately address questions regarding objective

thresholds appropriate for defining meteorological drought here. Based on this chosen

threshold, the major droughts are indicated in Table II. A statistical summary for all 50

stations used in this study is shown in Table III at the end of this document.

              Table II. Periods of time classified by percentage of stations (N) under 3-year long

                                             droughts (or longer)

       N < 25% of river basin       25% < N <50%           50% < N < 75 % of river basin
            1902-1904                 1908-1910                    1895-1899
            1913-1918                 1930-1935                    1949-1952
            1924-1926                 1954-1956                    1984-1988
            1938-1940                                              1998-2001

       Of the 108 years in the study, a total of 51 years are represented in the table above

as part of notable dry periods covering at least part of the ACF basin. There were four

drought periods from 1895 to 2002 in which at least one year had most of the river basin

area experiencing drought conditions (the last column in Table II). The first such drought

period began in 1895 (which is also the first year of data), in a drought that had more than

half of the selected stations in drought conditions. That is, the recorded values for a

station were below the 25th percentile. They concentrated over the northern and central

region. In 1897, 40 stations (out of 50) showed to be part of a three yearlong drought

event. The drought extent decreased the next year, by recording above the 10th percentile

total annual precipitation for all the river basin stations. Less than half of the selected

stations showed a continued drought event for the following year, concentrating mainly

on the southern regions before it faded on 1900 (Figure 8).

       The second drought period under this category occurred between 1949 and 1952.

Three years of selected stations with no-drought conditions preceded the year of 1949.

More than half of the stations suffered drought conditions that year, all evenly spread

throughout the river basin, and a quarter of the stations recorded total precipitation below

the 10th percentile for that year. The two following years suffered a river-basin extended

drought. Then, in 1952, only 25 percent of the stations experienced another year of

below normal rainfall, while the rest suffered drought for two of the next three years. A

year later, half of the river basin suffered another 3-year drought, while the rest suffered

two more years of drought out of three.

       The third drought period happened between 1984 and 1988. The two first years

of the event had more than half of the stations under the dry spell before most of the

entire basin stations fell under this condition. These stations were initially concentrated

mainly on the central region of the basin, with no stations either in Florida or the northern

part of the Georgia basin in drought conditions, according to the study definition. In

1986, 44 stations suffered the dry conditions, and in 1987 and 1988, forty-three stations

were still under the dry spell.

       The fourth period lasted from 1998 to 2001. Three of these four years had most

of the river’s drainage area suffering smaller-than-normal amounts of precipitation, and

more than half the basin in 2001 before drought anomalies disappeared. In 1998, only a

quarter of the basin’s stations experienced dry conditions. In 2002, only half of the

stations continued to have lower than normal annual precipitation. During this time (in

September 1999), Lake Jackson, a lake in Tallahassee in an enclosed basin covering

1,658 ha (4,097 acres), drained completely overnight into one of several karst sinkholes,

an event which has occurred throughout modern history during many of these dry spells

in the neighboring ACF Basin, since at least 1907. The drawdowns or drainings from

1907-1911, in the 1930s, 1950s, 1980s, and the present one all related quite closely to the

drought periods for the ACF Basin (Table II), suggesting a profound meteorological

reason for the drawdowns of this significant body of water, with historical excellent bass

fishing and recreation (Tucker 2003). The lake slowly began filling but emptied again in

2002, and still remains approximately 1.5 m below normal levels, which makes several

areas of the lakebed still dry, apparently the longest dry period for the lake since 1911.

       A few other periods of apparent drought did not meet the same threshold as the

prior periods, but they are notable, so they will be summarized here. The northern region

experienced very dry conditions from 1933 to 1935, with the former holding dry

conditions for the entire basin (Figure 9). The drought from 1954-1956 had at least half

of the studied stations under a multiple-year drought. The first two years of this period

were the driest of all. They had nearly all their stations in below-normal conditions, and

at least 17 stations were below the 10th percentile ever recorded for each year. The whole

decade seemed to be one of the driest, as it experienced two drought periods.

       The 1902-1904 drought was a northern region drought, although the entire river

basin experienced dry conditions in 1904. The 1908-1910 drought affected the southern

region of the basin (Figure 10). The year 1910 was dry below the 25th percentile for the

entire basin with 15 stations recording the first under-the-lower-decile values of the

  century. The central parts of the basin experienced one of the longest dry periods from

  1913 to 1918. The years 1913, 1914 and 1916 affected the entire region, but very few

  stations had recorded total annual precipitation below the lower decile value. The 1924-

  1927 drought affected the northern region. The years 1925 and 1927 brought the dry

  spell to the entire basin. This drought period has a slightly larger number of stations

  (about 5 to 6 more stations) under this condition than during 1902-1904. The 1938-1941

  period of time had the stations on the northern region experience 4 years of dry

  conditions. The years 1938 and 1941 affected most of the basin (Figure 11). The years

  1958 and 1960 were the harshest for the basin when 39 and 34 of the stations experienced

  dry conditions. The 1967-1969-drought period affected mostly the basin’s central region

  stations. Once again, the 1958-1960 period brought dryness to the northern regions.

VI. Summary

         The longest drought period found during the period 1895-2002 lasted 6 years,

  from 1913 to 1918. There were two droughts that lasted 5 years (1895-1899, 1984-1988)

  and 4 that lasted four years (1999-2002 being the most recent one). There were about

  twenty-three years that had at least 90 percent of the basin under dry conditions in the

  108-year period of record. The maximum number of stations reporting below the lower

  decile for a given year was 21. The decade of the 1950s was the driest, followed by the

  1930s by having at least 50 percent of the basin’s region under dry conditions for almost

  all of the 10 years. The 1970s was the least dry decade with just one year having more

  than 75 percent of the basin under the dry spell, followed by the 1990s.

VII. Discussion

          The Apalachicola-Chattahoochee-Flint river basin is an important part of the

   socioeconomic structure of Georgia’s, Alabama’s and Florida’s urban population,

   agricultural and power generation economy. They take advantage of the river’s stream

   flow to carry some of the most basic and important activities of any modern population.

   It has been demonstrated that the sources of water for the streams differ for each river.

   The Flint River, with a smaller flow, is more stable throughout the year because it has

   ground water feeding its mainstream, while the Chattahoochee River depends more on

   precipitation runoff. The Chattahoochee River possesses a much larger stream flow and

   has a greater contribution on the peaks recorded by the Apalachicola River. Mainly fed

   by the upstream rivers, the Apalachicola River also benefits from aquifer waters and

   other aquifer-fed affluent waters that flow into it. This way, it is possible to visualize that

   the Chattahoochee River will be affected the most by rainfall variability.

          Long-term droughts will not have an immediate effect on groundwater fed rivers,

   but they would significantly hurt the Chattahoochee River at first, and most likely the

   Apalachicola River, too. Prolonged droughts will also affect groundwater fed rivers like

   the Flint because its aquifers will not be able to recharge their water tables were a drought

   event to last long enough. Florida is very dependent on reliable water flows in the

   Apalachicola River, in particular, and the river established some record low flows (Figure

   12) during the most recent drought (USGS 2004). These low flows are serious

   impediments to any multi-state agreements that might solve water use allocations, and in

   fact, the ACF Basin water allocation formulas are likely to be resolved by the Federal

courts, the first time this will have occurred for the United States east of the Mississippi

River for a large multi-state river basin.

       With this study, it has been determined that the river basin has suffered from

several multiple-year-long droughts throughout the last century, sometimes even having

two events within the same decade. These periods of dryness can have a localized pattern

to them. For example, the 1902-1904 drought affected most of the basin’s northern

region, and with it, most of the Chattahoochee and Flint headwaters. Other droughts, like

the 1913-1918 drought mostly affected the central region of the basin.

       There is no obvious pattern for the return of drought periods, except

acknowledging they can happen within just one year from the other. It can also be said

that droughts have not lasted more than 6 years for any particular area, usually no longer

than two consecutive years for the entire basin, and that the likelihood of these type of

long term events is small. This assumption can be made if we take the first year used for

this study as the beginning of the first drought. Otherwise, this drought could have been

another of the longest recorded droughts in the ACF river basin. Years with extreme

events tend to be peaks of a drought period, although a few years have been recorded as

stand alone drought years. Two examples of this are the 1931 and 1981 droughts, which

experienced 20 and 13 stations, respectively, under the lower decile values, with the rest

of the basin under the 25th percentile values.

       It is tempting to compare return periods of these droughts with other quasi-

periodic events, such as was already done for the Lake Jackson drawdowns. It is well

known, for example, that the cool phase of the El Niño-Southern Oscillation (ENSO)

known as La Niña, is generally associated with drier than normal conditions in the

southeastern United States (Green et al. 1997). In comparing our drought years from

Table II to published La Niña years from the FSU Center for Ocean-Atmosphere

Prediction Studies (COAPS, cited 2004; data are classified beginning in 1900), we find

the connection somewhat plausible, with 29.4% (15 out of 51) of our dry years (from

Table II) also occurring during La Niña. Dry years also corresponded with seven El Niño

events (13.7%), while the remaining dry years were associated with neutral years.

Curiously, while none of the four prolonged extreme drought periods corresponded to La

Niña, two of the three droughts which affected between 1/4 and 1/2 of the basin matched

exactly prolonged La Niña periods (1908-1910 and 1954-1956). In addition, the years

1998 and 1999, half of the most recent devastating prolonged drought, were both La Niña

years. Further study of this relationship is suggested, given that long-term droughts are

generally longer than La Niña events, which will usually be preceded and followed by

neutral years, if not El Niño (warm phase).

       We have demonstrated that the Apalachicola-Chattahoochee-Flint river basin is

susceptible to multiple year droughts, and droughts receiving precipitation below the 10th

and 25th percentile that last up to three or more years have happen rather frequently.

Extremes can also happen during years that are not part of multiple years of drought.

This variability makes it difficult to estimate a single plan of action for these events, and

it is difficult to rely on the recurrence of certain patterns. As the generally wetter 1970s

and 1990s show, it can be an entire decade before another multiple-year drought occurs.

No decision-making can rely on the increased wetness of a decade, for this characteristic

may not last in the next one.       There are also many ways to define drought, in

meteorological and non-meteorological terms. Future studies of both physical and

sociological aspects of water use are recommended to resolve the issue of water supply

for the important ACF river basin.

Acknowledgements: Sincere gratitude to my honor's thesis committee members, Dr.
Henry Fuelberg and Dr. Sergio Fagherazzi for being a part of my honors thesis defense.
Thanks also to the departmental faculty committee, whose members provided some very
helpful comments to improve the manuscript. Special thanks to my family, that even
though far away from where I’ve worked and studied, always gave me moral support to
continue my pursuits, including this little project. We thank the Spatial Climate Analysis
Service at Oregon State University for permission to use the PRISM software, and thank
USGS for the use of their unpublished data. And last, but not least, my thanks to Dr. Paul
Ruscher, for all the time dedicated and the trust he put in me in doing this project.

VIII. References

    COAPS, cited 2004: ENSO Index According to JMA SSTA (1868-present). Retrieved
         online June 2004, from the Center for Ocean-Atmosphere Prediction Studies, at
         Florida State University, from

    Couch, Carol A., et al., cited 2004: Influences of Environmental Settings on Aquatic
          Ecosystems in the Apalachicola-Chattahoochee-Flint Basin. U.S. Geological
          Survey Water-Resources Investigations Report 95-4278.

    Daly, C., Gibson, W.P., M. Doggett, J. Smith, and G. Taylor. 2004. Up-to-date monthly
           climate maps for the conterminous United States. Proc., 14th AMS Conf. on
           Applied Climatology, 84th AMS Annual Meeting Combined Preprints, Amer.
           Meteorological Soc., Seattle, WA, January 13-16, 2004, Paper P5.1, CD-ROM.

    Green, P. M., D. M. Legler, C. J. Miranda V., and J. J. O'Brien, 1997: The North
           American climate patterns associated with the El-Niño-Southern Oscillation.
           Unpublished manuscript, COAPS Report Series 97-1, Center for Ocean-
           Atmosphere Prediction Studies, Florida State University, Tallahassee. Available
           online at

    Light, Helen. 2004. Increase in frequency and duration of very low flows (graph) USGS.
           Unpublished data.

    Marella, Richard L., Fanning, Julia L. and Mooty, Will S., 1993: Estimated Use of Water
           in the Apalachicola-Chattahoochee-Flint River basin during 1990 with State
           Summaries from 1970 to 1990, U.S. Department of Interior, WRI 93-4084
           Tallahassee, Florida, pg 1-58

    Torak, L.J., Davis G.D., Herndon, J.G., and Strain, 1991: Geohydrology and evaluation
           of water-resource potential of the Upper Floridan aquifer in the Albany area,
           southwestern Georgia: U.S. Geological Survey Open-File Report 91-52, 85 p.

    Tucker, T., 2003: "Fate of Famed Florida Lakes Remains Uncertain", Bassmaster April
           2003, retrieved online June 2004 from ESPN.COM at

    U.S.G.S., 2003a: National Water Quality Assessment (NAWQA) Program's
          Apalachicola-Chattahoochee-Flint (ACF) River basin study Home Page. U.S.
          Geological             Survey.          [Available       online       at

  U.S.G.S., 2003b: Water Watch, Current water resources conditions. U.S. Geological
        Survey. [Available online at]

  U.S. Army Corps of Engineers, (U.S.A.C.E.) 1984: Water assessment for the
        Apalachicola-Chattahoochee-Flint River Basin (Main Report): Mobile, U.S.
        Army Corps of Engineers Executive Summary, v. 1, 88 p.

  Yates L., Oliver, cited 2003: Water Wars in Drought-Ridden Southeast. [Available
        online at]

IX. Bibliography

  McKee, Thomas B., Doesken, Nolan J., and others, cited 2003: A History of drought in
       Colorado Lessons learned and what lies ahead., Colorado Climate Center,
       Colorado State University, Atmospheric Science Department [Available online at]

  Johnson, Mark E., Watson, Jr. Charles C., cited 2003: Hurricane Return Period
        Estimation,                 [Available                 online        at]

X. Other Data Sources

  National Climatic Data Center. Collection of Cooperative Observer Station’s Monthly
         Precipitation, Temperature data.

  “Spatial Climate Analysis Service, Oregon State                        University,, created 10 March 2004.”

                     Table III. Statistical summary of the 50 stations in the ACF river basin

          Climo Value (1931-2002)   Normal (1895-2002)                                                 Standard
Station     hundredths of an inch   hundredths of an inch         10 th
                                                                             25th               50th   Deviation
10422            5521.63                  5409.50                4349.40   4835.25         5306.500     913.68
12372            5322.06                  5351.74                4254.00   4780.50         5189.500     948.17
12730            5094.65                  5098.10                4199.70   4565.50         4979.500     814.65
14502            5492.00                  5431.50                4458.60   4835.50         5382.000     859.74
14884            5330.53                  5265.79                4126.00   4723.00         5253.500     903.73
15439            5282.57                  5196.89                4348.40   4600.25         5124.000     869.28
16129            5459.41                  5395.38                4405.60   4823.75         5332.000     890.04
18438            5297.10                  5327.90                4117.90   4842.25         5252.500     927.86
80211            5293.53                  5335.98                3991.80   4598.00         5181.500    1217.31
85372            5605.24                  5663.12                4552.60   5018.00         5606.500     922.29
87424            5496.39                  5505.41                4176.60   4794.75         5418.000     996.23
87429            5612.00                  5575.57                4284.20   4877.50         5613.500    1089.37
89566            6696.35                  6492.79                5221.70   5711.00         6240.000    1239.19
90140            4919.33                  5147.29                4021.60   4316.00         4905.000     889.91
90219            5047.75                  5112.72                4166.80   4548.00         5016.000     791.86
90253            4901.88                  4878.07                3844.70   4355.25         4793.000     890.66
90451            4839.61                  4861.03                4004.40   4396.25         4808.000     754.72
90581            5177.35                  5117.18                3947.70   4466.25         5036.000     969.96
90979            5467.39                  5373.60                4099.60   4661.00         5253.500     999.62
91372            4900.33                  4899.17                4016.40   4323.25         4746.500     788.61
91463            5189.45                  5167.15                4010.90   4598.00         5186.500     973.74
91500            5084.53                  5119.54                4011.50   4430.50         5034.000     871.21
91640            5222.39                  5187.44                4156.00   4527.50         5166.500     843.55
92006            6514.16                  6191.32                5045.70   5795.50         6403.000     974.53
92166            5007.78                  5228.72                3808.40   4336.00         4767.500     866.65
92198            4988.71                  4802.04                3976.00   4318.00         4733.000     771.19
92266            4539.80                  4538.52                3560.60   4094.50         4463.500     811.65
92450            5131.39                  4959.43                4023.60   4587.75         5024.000     929.35
92475            6287.45                  6069.16                5033.90   5531.00         6368.500    1035.81
92485            5183.76                  5570.37                4251.50   4743.50         5371.000     871.62
92570            5039.80                  5075.00                3869.80   4416.25         4895.500     871.13
92578            5932.59                  5649.40                4733.00   5341.50         5913.000     980.23
93147            5407.06                  5638.68                4360.50   4906.50         5463.000     862.44
93271            5084.16                  5076.09                4006.90   4333.50         4804.000     857.28
93516            5365.96                  5274.13                4083.30   4747.50         5237.500     945.20
93570            5233.71                  5122.38                4195.30   4688.75         5081.500     793.89
93621            5486.55                  5430.21                4371.00   4904.00         5551.000     902.44
93936            5122.82                  5139.93                4065.80   4344.00         4827.500     865.26
94700            4983.69                  4930.31                3916.10   4299.50         4831.000     876.89
94949            5319.24                  5105.02                4163.80   4596.75         5126.000     868.29
95979            4683.16                  4814.14                3738.70   4004.50         4406.000     790.85
96043            5333.67                  4993.79                4064.20   4702.00         5324.500     921.04

96335              5198.00                  5201.16                4102.20   4605.75      5182.500    880.92
96407              5149.57                  5142.70                4235.70   4567.25      5196.000    841.08
97087              4836.18                  5019.72                3869.90   4367.75      4729.000    814.30
97201              4978.53                  4861.73                3867.20   4319.50      4807.000    842.37
98535              5247.67                  4861.08                3865.90   4332.50      4816.500    954.67
98661              4973.16                  4943.75                3981.80   4309.00      4751.000    782.12
99291              5241.18                  5087.00                4193.40   4597.75      5074.500    848.74
99506               5088.75                 4966.22                4033.40   4449.00      4814.500    781.45
            Climo Value (1931-2002)   Normal (1895-2002)                                             Standard
Station       hundredths of an inch   hundredths of an inch                  Percentile              Deviation
      Table III continued.


Shared By: