justin t. maxwell*, peter t. soulé
TRANSCRIPT
CLIMATE RESEARCHClim Res
Vol. 38: 95–104, 2009doi: 10.3354/cr00772
Printed February 2009Published online January 13, 2009
1. INTRODUCTION
The 2007 drought in the southeastern and south-western United States attracted considerable mediaattention. Reports in the southeast claimed thatthe drought of 2007 was the worst in a century (NewYork Times, www.nytimes.com/2007/10/16/us/16drought.html; Time Magazine, www.time.com/time/magazine/article/0,9171,1684513,00.html). In addition,reports of damaged crops, diminishing water supplies,and large wildfires extended from early summer intolate fall. The drought affected pasture and hay produc-tion most severely, making it so difficult and expensiveto feed livestock that the USDA allowed farmers toapply for national disaster relief (USDA 2007). In Geor-gia, the drought caused an estimated $787.2 millionin agricultural production losses and $1.3 billion intotal economic losses (Flanders et al. 2007). Numerouscounties throughout the southeast, southwest, moun-
tain west, and northern Minnesota were considered‘primary natural disaster areas’ due to losses from ex-treme drought (USDA 2007). In addition, the droughtof 2007 severely reduced water levels in many reser-voirs, thereby affecting water supplies for munici-palities, industry, and hydroelectric and nuclearpower generation (P. O’Driscoll & L. Copeland, USAToday, www.usatoday.com/weather/news/2007-10-19-drought_N.htm; Duke Energy, www.duke-energy.com/news/releases/2007103001.asp). In the south-west, water levels in Lake Mead were 54% belowmaximum capacity at the end of October (K. Dewey,www.hprcc.unl.edu/nebraska/Lake-Mead-2007.html),and the southeastern states of Georgia, Florida, andAlabama were involved in federal lawsuits over con-trol of water releases from Lake Lanier (NBC 11 Alive,www.11alive.com/news/article_news.aspx?storyid=110717). The reduced supply of water forced manysoutheastern states to implement restrictions on wa-
© Inter-Research 2009 · www.int-res.com*Email: [email protected]
United States drought of 2007: historical perspectives
Justin T. Maxwell*, Peter T. Soulé
Department of Geography and Planning, Rankin Science Building, Appalachian State University, Boone, North Carolina 28608, USA
ABSTRACT: The impacts of the United States drought of 2007 to both society and ecosystems weresubstantive and included multi-billion dollar agricultural losses and the second worst wildfire seasonon record. The purpose of this paper is to place the 2007 drought in historical perspective relative tothe climate record from 1895–2007 to increase our understanding of this hazard and contribute toimprovements of drought mitigation plans. We compared the 2007 drought historically against theclimatic record (1895–2007) using the Palmer Drought Severity Index (PDSI). We then examined thetemporal progression of the 2007 drought and placed the peak month of drought severity (November)in historical perspective using rankings of severity and statistical recurrence intervals. Moreover, weexamined the climatic factors (e.g. geopotential height anomalies) that contributed to both abnor-mally dry and wet conditions recorded within the continental United States. While there were regionsthat experienced the worst drought on record both annually and in November during the calendaryear 2007, this year was not as severe as other notable drought years. November 2007 ties (with 5other years) for the 12th worst on record in terms of the number of climatic divisions experiencing theworst November drought. Statistically, drought/wetness conditions in November 2007 were notexceptionally extreme, with almost all of the calculated statistical recurrence intervals being muchless than the 113 year period of record.
KEY WORDS: PDSI · Historical rank · Recurrence interval · Media · Geopotential height · Anomalies
Resale or republication not permitted without written consent of the publisher
Clim Res 38: 95–104, 2009
tering lawns, washing automobiles, and other non-critical uses of water (CBS News; www.cbsnews.com/stories/2007/05/29/national/main2861151.shtml?source=search_story). Further, drought conditionshelped produce one of the worst wildfire seasons onrecord, with 3.8 × 106 ha burned nationwide in 2007,which is the second largest amount in recorded history(NOAA 2007a). Wahlquist (2003) found that mediacoverage of drought influences public perceptions andcommunity mitigation plans. She argues that most ofthe public receives information about drought condi-tions through the media; thus, the way the mediareports on a current drought will influence how thepublic prepares for future droughts.
Drought occurs throughout the entire United Statesand can potentially be severe anywhere (Cook et al.1999). Research on drought in the United States isdiverse and includes examinations of intensity, fre-quency, causes, and spatial and temporal patterns(Oglesby & Erickson III 1989, Soulé 1990, Knapp et al.2002, Hoerling & Kumar 2003, Salas et al. 2005). Whensevere, drought is a natural hazard that can causebillions of dollars in damages (Wilhite 2000). Continuedresearch on the causes and patterns of drought will in-crease our understanding of this hazard and contributeto improvements of drought mitigation plans. For ex-ample, the Carolinas Dynamic Drought Index (Carboneet al. 2008) uses a combination of indices at differenttemporal scales to help decision makers understand thebehavior and patterns of drought, which is needed toimprove mitigation plans in North and South Carolina.
Analyzing drought from a historical perspective hasbeen a frequent topic in drought research (Karl &Quayle 1981, Cook et al. 1988, Pielke et al. 2005, Her-weijer et al. 2006). Various techniques have been usedto characterize drought conditions, including principalcomponents analysis (Karl & Koscielny 1982, Cook et al.1999, Knapp et al. 2002), tree-ring analysis (Cook et al.1988, Cook et al. 1999, Timilsena et al. 2007), synopticanalyses (Trenberth et al. 1988, Palmer & Brankovic1989, McCabe et al. 2004), and analyses of frequency,duration, and drought rank / recurrence intervals (RI)(Diaz 1983, Karl & Young 1987, Soulé 1992, González &Valdés 2006). Pielke et al. (2005) argue that the spatialscale of the study is important in determining the his-torical rank. At finer scales (e.g. county level), a droughtcan be historically significant relative to a broader scale(e.g. state level). For this study, we chose to use climaticdivisions because they are small enough to show intra-state variations but large enough to easily interpret.
The purpose of this paper is to place the 2007drought in historical perspective relative to the climaterecord from 1895–2007. We discuss the progression ofthe 2007 drought by determining the month when thedrought peaked, and further examined that month by
comparing it historically. In addition, we examined theclimatic factors (e.g. geopotential height anomalies)that contributed to extreme drought (or wet) conditionsin many regions of the United States.
2. DATA AND METHODS
We obtained monthly data of the Palmer DroughtSeverity Index (PDSI; Palmer 1965) for the period1895–2007 for each of the 344 climatic divisions in thecontiguous United States (NOAA 2007b). This PDSIdata set is updated monthly and calibrated using the1931–1990 period (NOAA 2007c), and we obtained thedata in January 2008. The PDSI data are frequentlyused to characterize drought conditions in the UnitedStates (Soulé 1990, Hidalgo 2004, Bordi et al. 2006,Goodrich & Ellis 2006, Easterling et al. 2007), and havebeen used in historical comparisons (Soulé 1993, Cooket al. 1999, Herweijer et al. 2006, Timilsena et al. 2007).The PDSI is water balance-based and provides a mea-surement centered on zero that characterizes the mois-ture conditions of the current month relative to the cli-matic normals of the location (i.e. the climatic division).The index for any given month considers the moistureclimate of the current and preceding months, makingthe index moderate in response to changing moistureconditions. Because the index is based on a ‘moisturedeparture from normal’ within a given climatic division(Karl 1983; p. 1357), a month recording PDSI values of–3.0 (severe drought) in southwestern Arizona can becompared to a similar magnitude drought in south-eastern Georgia, since the magnitude of moisture sev-erity relative to long-term climatic normals is similarfor both locations.
To place the drought of 2007 in historic context, wedetermined and mapped the spatial pattern of the rankof average annual drought severity based on PDSIvalues using the full period of record (1895–2007) foreach climatic division. Thus, the driest year on recordreceived a rank of 1, while the wettest year received arank of 113. We used choropleth maps of drought sever-ity based on PDSI paired with 500 hPa geopotentialheight anomaly maps obtained from the NCEP/NCARreanalysis (NCEP/NCAR 2008a) to examine the tem-poral progression of the 2007 drought in the UnitedStates. To determine the month of maximum droughtseverity in 2007, we calculated the average monthlyvalue of drought severity across all climatic divisions(n = 344). To aid our interpretation of the temporal pat-terns of drought severity, we examined monthly meananomaly composite maps of columnar precipitablewater, 1000 mb temperature, 1000 mb vector wind, and1000 mb outgoing longwave radiation (NCEP/NCAR2008b) (data not shown).
96
Maxwell & Soulé: USA drought of 2007
After determining that drought conditions peaked inNovember, we placed this month in historical perspec-tive in 2 ways. First, we used the same procedures as forthe annual drought conditions and determined the rankof drought severity in November for each climatic divi-sion. Second, we calculated the statistical RI for thismonth for each climatic division using Kite’s (1988)methodology, which allows an assessment of the statisti-cal rarity of events based on standardized Z scores. Wefirst tested the monthly PDSI data from each climatic di-vision for normality using the Shapiro-Wilk test (Shapiro& Wilk 1965), a null hypothesis of no significant differ-ence from a normal probability distribution, and α = 0.01.Of the 4128 months tested (n = 12 mo for n = 344 cli-matic divisions), the majority (87%) were normally dis-tributed, allowing us to use the data directly withouttransformation. We separated these recurrence intervalsbased on whether an individual climatic division had pos-itive or negative moisture anomalies for the month, andpresent the recurrence intervals as a choropleth map.
3. RESULTS
While the drought of 2007 in the contiguous UnitedStates received most of the media attention, record-
setting moisture conditions were experienced on bothends of the spectrum (Fig. 1). In the 113 year studyperiod, 2% of the climatic divisions (7/344) across thecountry recorded their worst annual average droughtconditions in 2007, and 10.2% of the divisions haddrought conditions that were among the 5 worst yearsever. In contrast, 3 climatic divisions (0.9%) recordedtheir wettest year on record, and 5.5% had moistureconditions ranked in the top 5.
The spatial pattern of moisture conditions in 2007was consistent throughout the year, especially fromMay to November. Although drought intensity in-creased during the year, areas that began the year indrought (wetness) generally ended it in drought (wet-ness) (Fig. 2). Perhaps most intriguing is that the con-sistency of moisture patterns was not matched byconsistency of geopotential height anomaly patterns.In January (Fig. 2a), the flow was conducive for cyclo-genesis in the southern Great Plains, with the areabeing between a western trough and an eastern ridge.In March (Fig. 2b), virtually the entire country experi-enced positive geopotential height anomalies andabove normal temperatures (not shown), and droughtconditions worsened in both the mountain west andsoutheast. In May, most of the country again experi-enced positive geopotential height anomalies (Fig. 2c),
97
3 9
3
3
8 8
2
4
9
7
41
15
16
98
6
3
9
9
3
80
33
91
106
108
111
7
1
4
29
102 2
14
29
5
31
4
6
95
50
73
43
3
71 2
89
8
565
4
97
1
21
108
2
55
4
112
4
63
26
54
57
9
6
106
24
16
7
8
2
25
2
3
3
56
55
30
7
20
12
101
49
6
14 59
63
90
18
9945
14
5
25
1
18
14
37 14
23
84
48
23
16
27
38
88
98
63
35 26
73
97
24
2
16
110
31
83
88
92
20
1
59
31
6
27
81
85
76
104
6
14
17
6
50
14
111
31
22
1
93
66
35
88
14
49
1
18
19
86
77
66
42
67100
101
106
23
69
78
12 12
74
8983
15
12
31
13
13
13
98
93
17
69
43
113
10
98
79
72
100
48
87
63
104
101
35
25
46
10373
20
112
104
86
92
18
23
12
80
13
60
104 34
65
63
11
18
33
105
103
1060
94
18
27
104105
41
96
17
18
104
76
23
107 85
49 112
23
95
31
111
108
31
2158
10471
75 47
33
82
101
59
47
104
105
7794
88
11025
103
104
13
48
107104
68
94 37
97
59
76
50
109
54
105
21
2011
30
82
109
105
65
15
18
16
85
5323
22
111
46
99
91
105
14
113
41
46
111
24
90
66
4
106
12
54
30
43111
20
106
22
113
111
104
100 110112
77
106
102
33
7
21
43
98
20
105
33237
95
101
6413
8479
14
69
24276935
91
35
77
84
0 325 650 1300 km
Annual rank
1- 5
6 - 15
16 - 35
36 - 78
79 - 98
99 - 108
109 - 113
Fig. 1. Drought ranks (1 = driest, 113 = wettest) based on annual average Palmer Drought Severity Index (PDSI) values for 2007. Map created using ArcGIS 9.2 (ESRI 2006)
Clim Res 38: 95–104, 200998
035
070
014
00 k
m
Janu
ary
PD
SI v
alue
s
<–3.
00
–2.9
9 to
–1.
01
–1.0
0 to
1.0
0
1.01
to 2
.99
> 3.
00
Mar
ch P
DS
I val
ues
<–3.
00
–2.9
9 to
–1.
01
–1.0
0 to
1.0
0
1.01
to 2
.99
> 3.
00
035
070
014
00 k
m
a b
Fig
. 2 (
abov
e an
d f
ollo
win
g 2
pag
es).
200
7 m
onth
ly P
DS
I va
lues
by
clim
atic
div
isio
n (
left
pan
els)
an
d 5
00 h
Pa
geo
pot
enti
al h
eig
ht
anom
alie
s (r
igh
t p
anel
s) f
or (
a) J
anu
ary,
(b
) M
arch
, (c)
May
, (d
) Ju
ly, (
e) S
epte
mb
er, a
nd
(f)
Nov
emb
er. M
ap c
reat
ed u
sin
g A
rcG
IS 9
.2 (
ES
RI
2006
)
Maxwell & Soulé: USA drought of 2007 99
May
PD
SI v
alue
s
035
070
014
00 k
m
<–3.
00
–2.9
9 to
–1.
01
–1.0
0 to
1.0
0
1.01
to 2
.99
> 3.
00
c
July
PD
SI v
alue
s
035
070
014
00 k
m <–3.
00
–2.9
9 to
–1.
01
–1.0
0 to
1.0
0
1.01
to 2
.99
> 3.
00
d
Fig
. 2 (
con
tin
ued
)
Clim Res 38: 95–104, 2009100
Sep
tem
ber
PD
SI v
alue
s
035
070
014
00 k
m <–3.
00
–2.9
9 to
–1.
01
–1.0
0 to
1.0
0
1.01
to 2
.99
> 3.
00
e
Nov
embe
r P
DS
I val
ues
035
070
014
00 k
m <–3.
00
–2.9
9 to
–1.
01
–1.0
0 to
1.0
0
1.01
to 2
.99
> 3.
00
f
Fig
. 2 (
con
tin
ued
)
Maxwell & Soulé: USA drought of 2007
with the position of the strongest ridge axis over theGreat Lakes causing a contraction of the abnormallywet pattern in the northeast, an expansion of thedrought conditions in the southeast, and an intensifica-tion of wetness throughout the Great Plains. Thisoccurred in association with south and southeasterlyflow from the Gulf of Mexico that decreased rates ofevapotranspiration due to increased cloud cover andprecipitation (NCEP/NCAR 2008c). The Great Plainsexperienced abnormal amounts of cold, wet and cloudyconditions because the 1000 mb vector winds fromthe south continued strongly throughout the summermonths, expanding the area of wetness. In addition,tropical storm Erin moved well inland into the southernGreat Plains in mid-August, further enhancing thewetness in this region (especially Texas). In July, anenhanced western ridge was prominent (Fig. 2d), anddrought conditions intensified in the northern inter-mountain region in association with significantly abovenormal temperatures. Despite below normal 500 mbheights in the east in July, surface temperatures re-mained near normal and there was very little changein drought conditions. In both August (not shown) andSeptember (Fig 2e), weakly anomalous ridging oc-curred throughout most of the east, resulting in abovenormal temperatures and an expansion of droughtconditions. Moreover, the lack of land-falling tropicalsystems in the United States in 2007 (Unisys 2008) re-duced chances of any synoptic-scale drought-bustingsystem modifying the pattern.
The area affected by drought conditions peaked inSeptember, with more climatic divisions experiencingsevere drought conditions (≤–3.0 PDSI) in this month(Table 1). However, drought intensity for the entireUnited States peaked in November following a month(October, not shown) when geopotential height anom-alies were positive everywhere except in the coastalnorthwest Pacific. Positive height anomalies persisted
through November in the western states (Fig. 2f),resulting in some moderation of the extreme wetnessin the Great Plains.
November 2007 ties (with 5 other years) for the 12thworst on record, in terms of the number of climaticdivisions experiencing severe drought conditions(Table 2). In comparison, 1930 was far worse, with over5 times the number of climatic divisions recording theworst November drought conditions and displaying asubstantially more contiguous spatial pattern ofdrought than in 2007 (Figs. 3 & 4). Regionally, the spa-tial pattern of drought ranks in November 2007 showsthat the southeastern states and California experi-enced the worst drought conditions in terms of con-tiguous climatic divisions with top 5 rankings. Statisti-cally, drought/wetness conditions in November 2007were not exceptionally extreme, as only 2 climatic divi-sions recorded a recurrence interval extending beyondthe 113 year period of record for drought and 1 divisionfor wetness (Fig. 5). The resulting spatial pattern of therecurrence intervals reinforces the consistency ofdrought (or wetness) conditions throughout the year.
4. CONCLUSIONS
In 2007, some regions within the United States expe-rienced their worst drought conditions since 1895. Thereductions in soil moisture had multiple impacts forsociety, including reduced agricultural yields, increasedwildfire occurrence, and reduced supply in reservoirsthat service municipalities, industry, and power gener-ation. Despite these substantive impacts, the 2007drought was not exceptionally severe climatologicallyand statistically, in comparison to other droughts thathave occurred within the era of instrumental records(e.g. the drought of 1930). Further, moisture conditions
101
Table 1. Average monthly Palmer Drought Severity Index(PDSI) values (2007) across contiguous United States (n = 344)climatic divisions and number of climatic divisions recording
PDSI values ≤–3.0 (severe drought)
Month Average PDSI No. divisions ≤–3.0 PDSI
January 0.66 21February 0.34 19March 0.06 37April 0.19 35May –0.37 57June –0.43 60July –0.44 61August –0.45 74September –0.69 79October –0.51 59November –0.84 71December –0.43 61
Table 2. Comparison of years with all-time November low PDSI values
Rank Year No. climatic divisions
10. 1930 4320. 1934 4030. 1956 3940. 1954 2550. 1936 196.5 1953 116.5 1964 118.5 1965 108.5 2002 10
120.0 1897 8120.0 1924 8120.0 1976 8120.0 1988 8120.0 2007 8
Clim Res 38: 95–104, 2009102
4
9
2
2
9 8
4
9
2
9
10
8
98
4330
85
5
3
53
24
1
106
9
7
3
30
100 5108
29
9
34
1
6
80
45
58
6
66
3
83
8
92
36
98
14
105
3
57
5
113
5
15
19
40
43
8
5
7
109
13
7
34
14
1
3
10
28
9
2
15
2
2
16
6
1
63
56
6
7
36
5
2
5
19
15
2
25
76
19 54
45
1
90
15
7934
7
39
25
79
2
90
94
1
16
106
40 16
19
9
19
70
13
6
96
40
26
15
42
81
89
25
37 19
11
69
82
44
88
1
18
87
99
12
12
67
67
84
9
105
94
24
1
67
37
2
21
99
78
89
55
74
5
14
13
54
30
22
2
98
97
63
25
84
33
99 2110
13
77
87
23
32
62
97
103
97
57
7794
26 19
79
44
78
8
90
21
2
20
11 19
22
81
92 78
24
8
86
85
18
105
11
38
96
66
14
15
36
10342
89
56
10
24
77
15
63
15
65
29
24
20
107
102
2248
82
101
27
57
21
98
83
7924
41
43
12
88
105 90
70
19
21
41
19
4584
10538
110
28 57
21
90
99
32
101
59
43
38
105
17
88
106
8876
65
30
7385
23
2
96
47 24
69
57
32
24
58
18
13
50
51
95
14
20
54
1213
96
20
8
78
3
85
108
44
113
28
61
33
92
19
31
32
16
80
26
25
13
111
108
102
25
101107
100
36
867
16
39
88
8
23
79
15
41
3
19
21
15
113535
44
35
65
19
0 325 650 1300 km
2007 November Rank
1 – 5
6 – 15
16 – 35
36 – 79
80 – 98
99 – 108
109 – 113
Fig. 4. Drought ranks (1 = driest, 113 = wettest) based on average PDSI values for November 2007. Map created using ArcGIS 9.2 (ESRI 2006)
1930 NovemberPDSI values
<–3.00
–2.99 to –1.01
–1.00 to 1.00
1.01 to 2.99
> 3.00
0 325 650 1300 km
Fig. 3. Drought conditions in November 1930. Map created using ArcGIS 9.2 (ESRI 2006)
Maxwell & Soulé: USA drought of 2007
across the United States were highly variable in 2007,with a large portion of the country (e.g. southern GreatPlains) experiencing extreme wetness. Drought hasalways been a challenge for humans, and with thepopulation increases that have occurred in the UnitedStates since the worst drought years of the 1930s, evenless severe droughts have significant and long-lastingimpacts on humans and human activities. Thus, themedia attention received by the drought of 2007 ispartly a result of stresses placed on water resources byour increasing population. Further, the extensive presscoverage reinforced the need for enhanced water con-servation and drought mitigation plans at the local andregional level, and heightened the awareness of Amer-icans of our most precious natural resource.
LITERATURE CITED
Bordi I, Fraedrich K, Petitta M, Sutera A (2006) Extremedrought events comparison using PDSI and SPI indexes.Geophys Res Abstr 8:06020
Carbone G, Rhee J, Mizzell H, Boyles R (2008) A regional-scale drought monitoring tool for the Carolinas. Bull AmMeteorol Soc 89:20–28
Cook E, Kablack M, Jacoby G (1988) The 1986 drought in the
southeastern United States: How rare an event was it?J Geophys Res B 93:14.257–14.260
Cook E, Meko D, Stahle D, Cleaveland M (1999) Droughtreconstructions for the continental United States. J Clim12:1145–1162
Diaz H (1983) Some aspects of major dry and wet periodsin the contiguous United States, 1895–1981. J Clim ApplMeteorol 22:3–16
Easterling D, Wallis T, Heim R, Lawrimore J (2007) Theeffects of temperature and precipitation trends on USdrought. Geophys Res Lett 34:L20709, doi: 10.1029/2007GL031541
ESRI (Environmental Systems Research Institute) (2006)ArcGIS 9.2. Redlands, CA. Available at www.esri.com/
Flanders A, McKissick J, Shepherd T (2007) Georgia eco-nomic losses due to 2007 drought. The University of Geor-gia Center for Agribusiness and Economic Development.Center Report: CR-07-10 July 2007, p 1–8
González J, Valdés J (2006) New drought frequency index:definition and comparative performance analysis. WaterResour Res 42:W11421, doi: 10.1029/2005WR004308
Goodrich G, Ellis A (2006) Climatological drought in Arizona:an analysis of indicators for guiding the governor’sdrought task force. Prof Geogr 58:460–469
Herweijer C, Seager R, Cook E (2006) North Americandroughts of the mid to late nineteenth century: a history,simulation and implication for mediaeval drought. Holo-cene 16:159–171
Hidalgo H (2004) Climate precursors of multidecadal drought
103
6
55
-3-4
3
2
13
2
4
4
-16
36
-2
-5
2
-5
-10
-5
-4
-3
-13
2
-2
2
-9
-50
12
3
-35
-20 -17
-4
5
18
4
-9
-19
3
5
6
6
-6
-7
3
-3
8
-34
5
-8
-3
-7
3
4
-5
5
-8
5
9
10
3
-28
-7
2
48
2
2
9
3
4
3
-4
-6
-7
222
-17
-4
7
7
2
5
9
-6 -2
-3
3
5
-7
7
6
3
6
-4
4
2
-3
-5
35
4
-6
13
9
5 4
-3 -9
-7
-6
-8
-5
-3
5
-5
5
-7
-3
-4
-3
9
-16
6
2
-6
-25
-3
4
-7
-13
4
-9
3
-11
2
-5
36-3
12
-5
4
-19
-2
-51
-7
28
4
4
5
-8
5
2
-2
-4
-6
-20
-5
-4
-9
21
-6
-12
-4
4
-2
-38
-19
2
510
8
7
-2
-4 -6
11 -3
2
-7
-5
-6-6
-5
-17
-10
11
-6
225
-8
-62
-3
-13
7135
-8
-7
3
-3
-3
3
-2
-9
-5
-7
2 -12
5
-10
3
-8
8
-4
10
-4
-26
-5
-5-3
-22 57
-12-5
-6
-3
-3
-9
4
4
11
-6
-6
-13
-23
-3
-2
-14
-55
-5
-3
-3
-43
-43
-4
-304
-7
-3
-16
-6
-51
-17
-3
-3
-38
-9
-10
-6
-33
42
-11
-70
-4
-19
-15
-12
-22
-6
4
-3
-59
-2
-4
-5
2
-11
-26
-10
-3-8
-2
-18
19
-23
-10
-8
-7
-17
-3
-37
-31
-4
-63-13
-6
5
-4
-4
-9
15-4
-6
-14
10
-42
63
-9
17-39
10
-55 39
-5
-16
-57-156
-3
-14
-19
-7
-122
-3
-5
-7
-16
-3
-30
-5-3
-6
-8
-9
-12-22
-11
-4-4
-3
-4
-2
-5-3
Novemberrecurrence interval
>50 yr (Drought)
49 – 25 yr (Drought)
24 – 0 yr (Drought)
1 – 24 yr (Wet)
25 – 49 yr (Wet)
> 50 yr (Wet)
0 325 650 1300 km
Fig. 5. Statistical recurrence intervals for November 2007 calculated from 1895–2007 PDSI records. Map created using ArcGIS 9.2 (ESRI 2006)
Clim Res 38: 95–104, 2009
variability in the western United States. Water Resour Res40:W12504, doi: 10.1029/2004WR003350
Hoerling M, Kumar A (2003) The perfect ocean for drought.Science 299:691–694
Karl T (1983) Some spatial characteristics of drought durationin the United States. J Clim Appl Meteorol 22:1356–1366
Karl T, Koscielny A (1982) Drought in the United States:1895–1981. J Climatol 2:313–329
Karl T, Quayle R (1981) The 1980 heat wave and drought inhistorical perspective. Mon Weather Rev 109:2055–2073
Karl T, Young P (1987) The 1986 southeast drought in histori-cal perspective. Bull Am Meteorol Soc 68:773–778
Kite G (1988) Frequency and risk analysis in hydrology. WaterResources Publications, Colorado
Knapp P, Grissino-Mayer H, Soulé P (2002) Climatic regional-ization and the spatio-temporal occurrence of extreme sin-gle-year drought events (1500–1998) in the interior Pacificnorthwest, USA. Quat Res 58:226–233
McCabe G, Palecki M, Betancourt J (2004) Pacific andAtlantic Ocean influences on multidecadal drought fre-quency in the United States. Proc Natl Acad Sci USA 101:4136–4141
NCEP/NCAR (National Centers for Environmental Predic-tion/National Corporation for Atmospheric Research)(2008a) 500 mb geopotential height anomaly maps. Ac-cessed 14 Feb 2008. www.cdc.noaa.gov/Composites/Day/
NCEP/NCAR (2008b) Monthly mean anomaly compositemaps of columnar precipitable water, 1000 mb tempera-ture, 1000 mb vector wind, and 1000 mb outgoing long-wave radiation. Accessed 14 Feb 2008. www.cdc.noaa.gov/Composites/Day/
NCEP/NCAR (2008c) Monthly precipitation anomaly andNOAA interpolated OLR monthly mean and climatologymaps. Accessed 14 Feb 2008. www.cdc.noaa.gov/Com-posites/Day/
NOAA (National Oceanic and Atmospheric Administration)(2007a) Climate of 2007. Wildfire season summary.National Climatic Data Center. Accessed 15 Jan 2008.www.ncdc.noaa.gov/oa/climate/research/2007/fire07.html
NOAA (2007b) drd964x.pdsi.txt Accessed 15 Jan 2008. ftp://ftp.ncdc.noaa.gov/pub/data/cirs/drd964x.pdsi.txt
NOAA (2007c) Time bias corrected divisional temperature-precipitation-drought index. Accessed 15 Jan 2008. ftp://ftp.ncdc.noaa.gov/pub/data/cirs/divisional.README
Oglesby R, Erickson D III (1989) Soil moisture and the persis-tence of North American drought. J Clim 2:1362–1380
Palmer W (1965) Metrologic drought. Research Paper No. 45.US Department of Commerce, Weather Bureau, Washing-ton, DC
Palmer T, Brankovi ’c C (1989) The 1988 US drought linked toanomalous sea surface temperature. Nature 338:54–57
Pielke R, Doesken N, Bliss O, Green T and 5 others (2005)Drought 2002 in Colorado: An unprecedented drought ora routine drought? Pure Appl Geophys 162:1455–1479
Salas J, Fu C, Cancelliere A, Dustin D, Bode D, Pineda A,Vincent E (2005) Characterizing the severity and risk ofdrought in the Poudre River, Colorado. J Water ResPl-ASCE 131:383–393
Shapiro S, Wilk M (1965) An analysis of variance for normal-ity (complete samples). Biometrika 52:591–611
Soulé P (1990) Spatial patterns of multiple drought types inthe contiguous United States: a seasonal comparison. ClimRes 1:13–21
Soulé P (1992) Spatial patterns of drought frequency andduration in the contiguous USA based on multiple droughtevent definitions. Int J Climatol 12:11–24
Soulé P (1993) Hydrologic drought in the contiguous UnitedStates, 1900–1989: spatial patterns and multiple compari-son of means. Geophys Res Lett 20:2367–2370
Timilsena J, Piechota T, Hidalgo H, Tootle G (2007) Five hun-dred years of hydrological drought in the Upper ColoradoRiver Basin. J Am Water Resour Assoc 43:798–812
Trenberth K, Branstator G, Arkin P (1988) Origins of the 1988North American drought. Science 242:1640–1645
Unisys (2008) Unisys weather. Accessed 14 Feb 2008www.weather.unisys.com/
USDA (United States Department of Agriculture) (2007)USDA accepts applications for crop, feed and livestocklosses suffered Feb. 28 up to Dec. 31, 2007. Accessed 14Feb 2008 www.fsa.usda.gov/FSA/printapp?fileName=nr_20080125_rel_0020.html&newsType=newsrel
Wahlquist A (2003) Media representation and public percep-tions of drought. In: Botterill L, Fisher M (eds) Beyonddrought: people, policy, and perspectives. CSIRO Publish-ing, Collingwood Victoria, Australia, p 67–86
Wilhite D (2000) Drought as a natural hazard: concepts anddefinitions. In: Wilhite D (ed) Drought: a global assess-ment 1:3–18. Routledge, New York
104
Editorial responsibility: Bryson Bates, Wembley, WA, Australia
Submitted: March 4, 2008; Accepted: August 5, 2008Proofs received from author(s): December 1, 2008