Professional Documents
Culture Documents
Depth-Duration Frequency of
Precipitation for Texas
By William H. Asquith
Austin, Texas
1998
U.S. DEPARTMENT OF THE INTERIOR
Bruce Babbitt, Secretary
Second printing
District Chief
U.S. Geological Survey
8011 Cameron Rd.
Austin, TX 787543898
ii
CONTENTS
Abstract ............................................................................................................................................................................... 1
Introduction ......................................................................................................................................................................... 1
Purpose and Scope ................................................................................................................................................... 1
Previous Studies ....................................................................................................................................................... 2
Climate of Texas ...................................................................................................................................................... 2
Data Base of Precipitation Annual Maxima ............................................................................................................ 3
Data Sources ................................................................................................................................................ 3
Correction of Precipitation Annual Maxima for Fixed-Interval Recording ................................................. 6
Acknowledgments ................................................................................................................................................... 11
Regionalization of Precipitation Annual Maxima ............................................................................................................... 11
L-moments ............................................................................................................................................................... 12
Generalized Logistic and Generalized Extreme-Value Distributions ...................................................................... 16
Spatial Averaging of L-coefficient of Variation and L-skew and Estimation of L-scale ......................................... 17
Contouring Distribution Parameters ........................................................................................................................ 21
Depth-Duration Frequency of Precipitation for Texas ........................................................................................................ 21
Examples of Depth-Duration Frequency Computation ........................................................................................... 23
Precipitation Intensity-Duration Frequency Curve .................................................................................................. 24
Comparison to Previous Studies .............................................................................................................................. 25
Summary ............................................................................................................................................................................. 25
Selected References ............................................................................................................................................................ 26
Appendix 1. Fifteen-Minute Precipitation Stations in Texas With at Least 10 Years of Annual Maxima
Data Through 1994 ........................................................................................................................................ 72
2. Hourly Precipitation Stations in Texas With at Least 10 Years of Annual Maxima Data
Through 1994 ................................................................................................................................................. 75
3. Daily Precipitation Stations in Texas With at Least 10 Years of Annual Maxima Data
Through 1994 ................................................................................................................................................. 83
PLATE
(Plate is in pocket)
1. Map showing locations of 15-minute, hourly, and daily recording precipitation stations in Texas with at
least 10 years of annual maxima data through 1994
FIGURES
1. Graph showing relation between precipitation depth and storm duration for world record, Texas record,
and Texas data-base maxima ............................................................................................................................... 4
2. Map showing modified National Weather Service climatic regions of Texas ..................................................... 5
3. Boxplots showing record-length distribution for 15-minute, hourly, and daily precipitation stations in
climatic regions of Texas ..................................................................................................................................... 7
47. Graphs showing:
4. Time series of 1-day precipitation annual maxima for long-term stations (1778, 2266, and 2444)
and distribution of available record in corresponding climatic regions of Texas ..................................... 8
5. Time series of 1-day precipitation annual maxima for long-term stations (3262, 3873, and 3992)
and distribution of available record in corresponding climatic regions of Texas ..................................... 9
6. Time series of 1-day precipitation annual maxima for long-term stations (4058, 5272, and 9522)
and distribution of available record in corresponding climatic regions of Texas ..................................... 10
7. L-moment ratio diagrams for selected precipitation durations in Texas .................................................. 14
89. Boxplots showing:
8. Effects of spatial averaging on the statewide distribution of L-coefficient of variation for each
precipitation duration ................................................................................................................................ 19
9. Effects of spatial averaging on the statewide distribution of L-skew for each precipitation
duration ..................................................................................................................................................... 20
CONTENTS iii
1018. Maps showing location () parameter of generalized logistic (GLO) distribution for:
10. 15-minute precipitation duration in Texas ................................................................................................ 28
11. 30-minute precipitation duration in Texas ................................................................................................ 29
12. 60-minute precipitation duration in Texas ................................................................................................ 30
13. 1-hour precipitation duration in Texas ...................................................................................................... 31
14. 2-hour precipitation duration in Texas ...................................................................................................... 32
15. 3-hour precipitation duration in Texas ...................................................................................................... 33
16. 6-hour precipitation duration in Texas ...................................................................................................... 34
17. 12-hour precipitation duration in Texas .................................................................................................... 35
18. 24-hour precipitation duration in Texas .................................................................................................... 36
1923. Maps showing location () parameter of generalized extreme-value (GEV) distribution for:
19. 1-day precipitation duration in Texas ........................................................................................................ 37
20. 2-day precipitation duration in Texas ........................................................................................................ 38
21. 3-day precipitation duration in Texas ........................................................................................................ 39
22. 5-day precipitation duration in Texas ........................................................................................................ 40
23. 7-day precipitation duration in Texas ........................................................................................................ 41
2432. Maps showing scale () parameter of generalized logistic (GLO) distribution for:
24. 15-minute precipitation duration in Texas ................................................................................................ 42
25. 30-minute precipitation duration in Texas ................................................................................................ 43
26. 60-minute precipitation duration in Texas ................................................................................................ 44
27. 1-hour precipitation duration in Texas ...................................................................................................... 45
28. 2-hour precipitation duration in Texas ...................................................................................................... 46
29. 3-hour precipitation duration in Texas ...................................................................................................... 47
30. 6-hour precipitation duration in Texas ...................................................................................................... 48
31. 12-hour precipitation duration in Texas .................................................................................................... 49
32. 24-hour precipitation duration in Texas .................................................................................................... 50
3337. Maps showing scale () parameter of generalized extreme-value (GEV) distribution for:
33. 1-day precipitation duration in Texas ........................................................................................................ 51
34. 2-day precipitation duration in Texas ........................................................................................................ 52
35. 3-day precipitation duration in Texas ........................................................................................................ 53
36. 5-day precipitation duration in Texas ........................................................................................................ 54
37. 7-day precipitation duration in Texas ........................................................................................................ 55
3846. Maps showing shape () parameter of generalized logistic (GLO) distribution for:
38. 15-minute precipitation duration in Texas ................................................................................................ 56
39. 30-minute precipitation duration in Texas ................................................................................................ 57
40. 60-minute precipitation duration in Texas ................................................................................................ 58
41. 1-hour precipitation duration in Texas ...................................................................................................... 59
42. 2-hour precipitation duration in Texas ...................................................................................................... 60
43. 3-hour precipitation duration in Texas ...................................................................................................... 61
44. 6-hour precipitation duration in Texas ...................................................................................................... 62
45. 12-hour precipitation duration in Texas .................................................................................................... 63
46. 24-hour precipitation duration in Texas .................................................................................................... 64
47. Graph showing precipitation intensity-duration curves of 100-year storm for selected localities in Texas ........ 65
48. Map showing depth of 100-year storm for 6-hour precipitation duration in Texas ............................................. 66
49. Graph showing precipitation intensity-duration curves of 100-year storm for Amarillo and Orange,
Texas .................................................................................................................................................................... 67
TABLES
1. Correction factors for mean annual maxima of precipitation .............................................................................. 11
2. Summary of goodness-of-fit and heterogeneity measures for Texas ................................................................... 15
3. Summary statistics of distribution-parameter maps for each precipitation duration in Texas ............................. 22
iv
Depth-Duration Frequency of
Precipitation for Texas
By William H. Asquith
Abstract 1
2, 3, 5, and 7 days were investigated. The frequencies In recent years, other investigations of precipita-
(expressed as recurrence intervals) range from 2 to 500 tion frequency have been conducted. Schaefer (1990)
years. Selection of a bound on recurrence interval is and Parrett (1997) presented precipitation frequency
difficult. In general, as recurrence interval increases, investigations based on the generalized extreme-value
the accuracy of DDF estimate decreases. An upper (GEV) distribution for Washington State and Montana,
recurrence interval of 500 years is considered reason- respectively. Also, the GEV distribution was selected
able since most hydraulic design criteria for engineered by Huff and Angel (1992) to model the frequency of
structures and controls incorporating DDF require annual precipitation maxima for durations of 5 minutes
recurrence intervals equal to or less than 500 years. The to 10 days for the midwestern United States (Illinois,
reciprocal of a recurrence interval is an exceedance Indiana, Iowa, Kentucky, Michigan, Minnesota,
probability. For example, a 100-year precipitation has Missouri, Ohio, and Wisconsin). For the western United
an annual exceedance probability of 0.01, which by def- States (Arizona, Nevada, New Mexico, Utah, and parts
inition means that there is a 1-percent chance in each of California, Colorado, Idaho, Oregon, Texas, and
and every given year that this precipitation will be Wyoming), the generalized Pareto distribution was
equaled or exceeded. selected to model the frequency of annual precipitation
maxima for durations of 1 hour to 7 days (L.T. Julian,
Accurate DDF analysis using data from any
National Weather Service, written commun., 1997).
one station is difficult because the data for one station
The NWS study covers only a small part of the Texas
represent a poor spatial and (or) temporal sampling of
Panhandle and the northern one-half of the Trans-
precipitation. For example, storms occur over areas
Pecos region of West Texas. Wilks and Cember (1993)
that might or might not contain a station; and generally,
selected the Beta-P distribution to model the frequency
only short record is available at a single station. Addi-
of annual precipitation maxima for durations of 1 to
tionally, the distribution of precipitation associated
10 days for northeastern United States and southeastern
with any one station tends to be highly nonuniform.
Canada. Other studies of precipitation frequency, which
More accurate DDF estimates can be developed by are at or near completion, are for Canada (Younes Alila,
pooling or regionalizing data from many nearby University of British Columbia, Vancouver, Canada,
stations (see Stedinger and others, 1993, p. 18.33). The written commun., 1998), for southern British Columbia
regionalization procedures used are explained in the (M.G. Schaefer, MGS Consultants, Lacey, Wash., writ-
Regionalization of Precipitation Annual Maxima sec- ten commun., 1998), and for Oklahoma (R.L. Tortorelli,
tion of this report. The procedures were developed U.S. Geological Survey, written commun., 1998).
using a data base of precipitation annual maxima for
173 (15-minute), 274 (hourly), and 865 (daily) National
Climate of Texas
Weather Service (NWS) stations, each with at least
10 years of record in Texas. In total, about 3,030 cumu- The climate of Texas varies considerably across
lative years of record is available for the 15-minute the State, and the climate for any given location in
stations, 10,160 years for the hourly stations, and Texas is highly variable and nonuniform. This variabil-
38,120 years for the daily stations. ity and nonuniformity is due in part to the influences
of four principal types of air masses that periodically
Previous Studies move through the State (Bomar, 1995). Two of these
air masses generally are characterized as warm and
DDF information for the area of the United States moist. The continental tropical originates in the
east of the Rocky Mountains, including Texas, is avail- Pacific Ocean and moves into Texas from the west;
able from two principal sources. The first source, com- and the maritime tropical originates in the Gulf of
monly known as TP40 (Hershfield, 1962), presents Mexico, southeast of Texas. The maritime polar and
DDF for durations of 30 minutes to 24 hours and recur- continental polar, or Arctic, arrive from the northwest
rence intervals of 2 to 100 years. The work of Hershfield and north, respectively, and generally are characterized
was extended by Frederick and others (1977) and com- as cold and dry. Additionally, tropical cyclones, which
monly is known as HYDRO35. The HYDRO35 study also contribute to the climatic variability and nonunifor-
presents DDF for durations of 5 to 60 minutes and mity, occasionally affect Texas during the summer and
recurrence intervals of 2 to 100 years. early fall. These cyclones have produced many daily
INTRODUCTION 3
4
100
90
80
70
Alvin (7/2427/1979)
60
) (4.0)
50 tes
inu Medina and Kerr
nm Counties (8/13/1978)
40 5 ( Di (4.1)
.47
(4.1) Ingram (6/307/2/1932)
7D
2.3
30 P=
(3.5)
25 Odem (10/19/84)
(3.2) Thrall (9/610/1921)
20
DHanis (5/31/35)
15 te
Sta
(2.8) ar t of Shackelford County (8/13/1978)
rn p
ste
hea
4
1 day
1 hour
3
24 hours
60 minutes
2.5
5 10 15 30 1 2 3 6 12 1 2 3 5 7 10
MINUTES HOURS DAYS
STORM DURATION
EXPLANATION
World maximum depth-duration relation (World Meteorological Organization, 1986)Equation computes envelope curve
Maximum statewide 100-year precipitation depthderived from depth-duration frequency analysis in report
Statewide maximum at recording precipitation stations for corresponding durationderived from Texas precipitation data base
(2.8)
Selected greatest precipitation depths in Texas (John Patton, National Weather Service, written commun., 1996)Number
in parenthesis is ratio of depth to 100-year depth for corresponding locality derived from depth-duration frequency analysis,
figures 1046
Figure 1. Relation between precipitation depth and storm duration for world record, Texas record, and Texas data-base maxima.
HIGH
PLAINS
LOW
ROLLING
PLAINS
NORTH
CENTRAL
EAST
TEXAS
TRANS-
PECOS
EDWARDS
PLATEAU
SOUTH
CENTRAL UPPER
COAST
EXPLANATION SOUTH
TEXAS
Climatic region boundary
Figure 2. Modified National Weather Service climatic regions of Texas (modified from Carr, 1967).
by Hydrosphere Data Products, Inc. (1996). The 15- 12-, and 24-hour annual maxima were compiled from
minute data generally are available for 197294. Hourly the hourly data; likewise, the 1-, 2-, 3-, 5-, and 7-day
precipitation data generally are available for 194894. annual maxima were compiled from the daily data. To
Daily precipitation data are available for about 1910 clarify potentially confusing terminology in this report,
94. Additionally, the precipitation record was extended the term 60-minute duration is associated with the
for nine hourly stations from data provided by the NWS aggregation of four consecutive 15-minute data values
(John Vogel, written commun., 1996). as opposed to the 1-hour data from a single value. Sim-
The 15-, 30- and 60-minute annual maxima were ilarly, the term 24-hour duration is associated with the
compiled from the 15-minute data. The 1-, 2-, 3-, 6-, aggregation of 24 consecutive 1-hour data values as
INTRODUCTION 5
opposed to the 1-day data from a single value. For some proximity (about 3 mi north to south or 2 mi east to
years, the influence of missing record or deleted record west) is the basis for combining the records of a few sta-
caused inconsistencies in the availability of depths for tions. For this study, (1) the records of 8 fifteen-minute
all the durations. In these instances, the missing or stations were combined into 4 records; (2) the records of
deleted year was removed from the data base. Removal 30 hourly stations were combined into 15 records; and
of some years did not greatly reduce the available pre- (3) the records of 47 daily stations were combined into
cipitation data for Texas. 23 records. The combined records are not separately
Only those stations that have at least 10 years of identified; however, they can be determined from the
data were selected for inclusion in this study. The author station names in the Appendices.
believes that a minimum of 10 years of data for stations
in Texas is needed before the data become representa- Correction of Precipitation Annual Maxima for
tive of the true precipitation characteristics. A basis for Fixed-Interval Recording
the 10-year minimum record length is that droughts or
periods of abundant rainfall in Texas seldom last more The ratios of 60-minute to 1-hour maxima and
than 4 to 5 years. Therefore, using data with at least 24-hour to 1-day maxima were analyzed. Both the
10 years of record is an attempt to insure that the col- 1-hour and 1-day maxima generally are less than the
lected data are not unduly biased by the cyclical nature 60-minute and 24-hour maxima, respectively, because
of the climate. the 1-hour and 1-day maxima are collected on a fixed-
The 15-minute, hourly, and daily NWS stations interval basis (beginning to end of each hour or mid-
used for this study are listed in the Appendices. The night to midnight); whereas, the 60-minute and 24-hour
locations of the 15-minute, hourly, and daily stations are maxima are determined with four consecutive or
shown on plate 1. moving 15-minute windows or 24 consecutive 1-hour
Each of the three types of precipitation stations windows, respectively. In total, 170 stations have con-
generally are uniformly distributed around the State; current 15-minute (therefore 60-minute) and hourly
however, of the three, the daily precipitation stations maxima, which comprise about 2,900 values; and 144
have the densest network. The areal densities in stations stations have concurrent hourly (therefore 24-hour) and
per 1,000 square miles (mi2) for each of the station daily maxima, which comprise about 4,800 values. A
types are 0.646, 1.02, and 3.23 for the 15-minute, 60-minute to 1-hour weighted-mean ratio of 1.13
hourly, and daily stations, respectively. The distribution [weighted standard deviation (SD) = 0.054] and a 24-
of record length for each region and station type are hour to 1-day weighted-mean ratio of 1.13 (weighted
shown in figure 3. The figure indicates that the record SD = 0.082) were determined for the entire State. The
length is reasonably well distributed among the regions weighting is based on record length for each station.
for each of the station types. However, there are clear The spatial distribution of the two mean ratios was
differences in the record availability between the investigated. The mean ratios show no apparent spatial
station types. In general, the daily stations have the dependence.
longest records of the three station types; whereas, the Weiss (1964) considered the ratio of true to fixed-
15-minute stations have the shortest records. interval maxima on the basis of probability theory and
To illustrate the inherent variability of a series determined a theoretical correction factor of 1.143
of 1-day precipitation annual maxima, graphs are shown for the true 60-minute to hourly maxima and for the
for each climatic region using data from a long-term true 24-hour to daily maxima. Miller and others (1973)
station (figs. 46). Similar variability is indicated by used an empirically derived factor of 1.13. Both factors
graphs of the 15-minute and hourly stations (not are either reasonably close to or match those determined
included in this report). Additionally, figures 46 show for this report. On a conceptual basis, both empirical
the median and mean beginning and ending years of factors from this report would be expected to be slightly
daily precipitation stations for each climatic region. less than the theoretical factor because both the 60-
Occasionally an existing station is relocated to a minute and 24-hour maxima were developed from finite
nearby site, and new stations might be established near samples (15-minute or 1-hour intervals) and therefore
an existing station. If the distance between two or more are biased slightly downward themselves.
stations is small, then their records were assumed equiv- However, it is possible that Weisss corrections
alent and combined into one longer record. Geographic are too large, given that 1.13 has been determined by
80
60
40
(17) (14) (57) (19) (14) (29) (7) (9)
(7)
20
130
120 HOURLY STATIONS
(18)
RECORD LENGTH, IN YEARS
(26) (33)
60
40
20
130
120 DAILY STATIONS (178)
RECORD LENGTH, IN YEARS
60
40
20
0
SOUTH
UPPER
TRANS-
COAST
TEXAS
EDWARDS
ROLLING
PLAINS
CENTRAL
CENTRAL
TEXAS
PECOS
PLATEAU
EAST
PLAINS
HIGH
NORTH
SOUTH
LOW
CLIMATIC REGION
EXPLANATION
Distribution of record length
(17) Number of stations
Maximum value
75th percentile
Interquartile range Median
25th percentile
Minimum value
Figure 3. Record-length distribution for 15-minute, hourly, and daily precipitation stations in climatic regions of
Texas.
INTRODUCTION 7
8
20
HIGH PLAINS
83 STATIONS
15 MEAN RECORD: 52 YEARS
20
UPPER COAST
56 STATIONS
15
MEAN RECORD: 48 YEARS
20
EAST TEXAS
0
1895 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 1995
YEAR
EXPLANATION
1-day precipitation annual maximum for station
Figure 4. Time series of 1-day precipitation annual maxima for long-term stations (1778, 2266, and 2444) and distribution of available record in
corresponding climatic regions of Texas.
20
TRANS-PECOS
58 STATIONS
15 MEAN RECORD: 39 YEARS
20
SOUTH CENTRAL
64 STATIONS
15 MEAN RECORD: 50 YEARS
20
0
1895 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 1995
YEAR
EXPLANATION
1-day precipitation annual maximum for station
INTRODUCTION
Figure 5. Time series of 1-day precipitation annual maxima for long-term stations (3262, 3873, and 3992) and distribution of available record in
9
corresponding climatic regions of Texas.
10
20
SOUTH TEXAS
15 60 STATIONS
MEAN RECORD: 50 YEARS
20
EDWARDS PLATEAU
15 160 STATIONS
MEAN RECORD: 35 YEARS
20
0
1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 1995
YEAR
EXPLANATION
1-day precipitation annual maximum for station
Figure 6. Time series of 1-day precipitation annual maxima for long-term stations (4058, 5272, and 9522) and distribution of available record in
corresponding climatic regions of Texas.
Table 1. Correction factors for mean annual maxima of precipitation
[Note: Correction factors (Weiss, 1964) are from the equation {n/n 0.125}, where n is the number of subintervals for a given
duration; min, minutes; n/a, empirical correction factor not determined for this report; hr, hours]
Corresponding Empirical correction
Duration Weiss correction factor1
fixed intervals No. subintervals factors for Texas
(variable) (dimensionless)
or time steps (dimensionless)
15 min 15 min 1 1.143 n/a
30 min do. 2 1.067 n/a
60 min do. 4 1.032 n/a
1 hr 1 hr 1 1.143 1.13
2 hr do. 2 1.067 n/a
3 hr do. 3 1.044 n/a
6 hr do. 6 1.022 n/a
12 hr do. 12 1.011 n/a
24 hr do. 24 1.005 n/a
1 Weiss, 1964.
others, and 1.13 was determined for this report. This tion factors from Weiss (1964), pertinent to this study,
difference (between 1.143 and 1.13) might be due to are listed in table 1. Schaefer (1990, p. 127) concludes
Weisss assumption that the probability of a storm event that correction factors such as Weisss or Millers affect
occurring is equal throughout any time interval and the only the location (mean) and scale (SD) measures of a
assumption that the distribution of precipitation during given data set and do not affect higher measures of the
the time interval is uniform. In reality, in the Panhandle data such as shape (skew). The adjustment to the scale
and West Texas, large storms occur more frequently measure of the annual maxima is discussed later in the
during the late afternoon and early evening; while in Spatial Averaging of L-coefficient of Variation and
Central Texas, large storms frequently occur at night. L-skew and Estimation of L-scale section.
Disregarding Weisss assumptions could produce
slightly smaller correction factors. However, the per- Acknowledgments
cent change of 1.143 to 1.13 is only about 1.1 percent.
The theoretical correction factors developed by Weiss The author expresses his gratitude to Sue Giller
were used for this study because of (1) the small differ- and Ben Hardison of Hydrosphere Data Products, Inc.,
ence between Weisss 1.143 and Millers empirical for their compilation (on diskette) of precipitation
1.13 and (2) the fact that corrections were needed for the annual maxima for the identified durations and to John
15-, 30-, and 60-minute durations and for durations Vogel of the NWS for providing some noncomputer-
greater than 1 hour or 1 day. The correction factors are stored hourly data. Furthermore, the author expresses
applicable only on the mean annual maxima for a given his gratitude to the thorough and timely review of early
duration and are not actually applied to the individual drafts by Mel Schaefer of MSG Consultants, and
data points (Weiss, 1964, p. 79). The theoretical correc- Charles Parrett and Robert Tortorelli of the USGS.
INTRODUCTION 11
REGIONALIZATION OF PRECIPITATION L-moments
ANNUAL MAXIMA
This section presents a brief introduction to the
theory of L-moments and the technique for selecting
The concept of regionalization in the context
an appropriate probability distribution. Comprehensive
of a precipitation frequency analysis is the process
discussion pertinent to L-moments, their historical
by which the precipitation characteristics surrounding development, and distribution selection can be found
a particular station are combined or pooled together to in Hosking (1986, 1990), Hosking and Wallis (1993,
develop more accurate statistical summaries of 1997), and Vogel and Fennessey (1993).
precipitation characteristics than can be derived from a
Consider a random variable X (precipitation depth
single station. Additionally, once regionalization is in this report) with a cumulative probability distribution
completed, precipitation frequency can be estimated for function F (nonexceedance probability). The quantities
locations other than precipitation stations.
1
i j k i j k
The underlying assumption of the regionalization
approach adopted for Texas is that the parameters of the
M ijk = E [X F ( 1 F ) ]= 0 X F ( 1 F ) d F , (1)
0.4
0.3
0.2
0.1
0.6
0.4
0.3
0.2
0.1
[GLO, generalized logistic distribution; GEV, generalized extreme-value distribution; min, minutes; hr, hours]
than about 1.64; this interpretation has validity only if the GEV distribution is judged appropriate for annual
the region containing the data is homogeneous (see maxima for durations of 1 to 7 days. Various geographic
Hosking and Wallis, 1993, 1997). Homogeneous is regions much smaller in areal extent than Texas were
interpreted to mean that, except for a location-specific delineated, and the goodness-of-fit and heterogeneity
scaling factor, (1) the form of the distribution for the measures were computed. Frequently, regions with
homogeneous region is known, and (2) the distribution areal extents less than about 10 percent of the total
is exactly defined (defined by specification of its area of Texas show (1) H-statistics close to 1.0 (thus,
parameters). The heterogeneity measures (H-statistic) near homogeneous), (2) absolute values of Z-statistics
for the entire State were computed for each duration of the GLO distribution less than about 1.64 for the
following the methods of Hosking and Wallis (1993). 15-minute to 24-hour durations, and (3) absolute values
The H-statistics (table 2) are larger than 2.0, which of Z-statistics of the GEV distribution less than about
Hosking and Wallis (1993, p. 275) conclude is indica- 1.64 for the 1- to 7-day durations.
tive of a heterogeneous region; thus, strict interpretation The previous studies, TP40 (Hershfield, 1962)
of the goodness-of-fit measure is questionable. The and HYDRO35 (Frederick and others, 1977), each
second interpretation of the goodness-of-fit is that the used the two-parameter Gumbel distribution, which is a
form of the distribution with the smallest Z-statistic is special case of the GEV distribution, where 3 and 4 are
considered appropriate, but the distribution parameters equal to 0.1699 and 0.1504, respectively (Stedinger and
might require additional specification. Applying the others, 1993, p. 18.9). Either the GLO or GEV
second interpretation to the Z-statistics (table 2), the distribution fits the data better than the Gumbel distribu-
GLO distribution is judged appropriate for annual max- tion because the Gumbel distribution, which plots as a
ima for durations of 15 minutes to 24 hours; whereas single point in figure 7 (but is not shown for clarity), is
0.2
L-COEFFICIENT OF VARIATION
0
0.8
0.2
L-COEFFICIENT OF VARIATION
0
T ES T ES TES 1 HOUR 2 HOURS 3 HOURS 6 HOURS 12 HOURS 24 HOURS 1 DAY 2 DAYS 3 DAYS 5 DAYS 7 DAYS
INU I NU I NU
15 M 30 M 60 M DURATION
EXPLANATION
Statewide distribution of L-coefficient of variation
(173) Number of stations
Maximum value
75th percentile
Interquartile range Median
25th percentile
Minimum value
19
20
1.0
(173) (274) (274) (274)
0.8 (274) L-SKEW (3) FROM STATION DATA
(173) (274) (274)
(865) (865) (865)
(173) (865) (865)
0.6
0.4
0.2
L-SKEW
0
-0.2
-0.4
-0.5
1.0
0.2
L-SKEW
0
-0.2
-0.4
-0.5
T ES T ES T ES 1 HOUR 2 HOURS 3 HOURS 6 HOURS 12 HOURS 24 HOURS 1 DAY 2 DAYS 3 DAYS 5 DAYS 7 DAYS
INU INU INU
15 M 30 M 60 M DURATION
EXPLANATION
Statewide distribution of L-skew
(173) Number of stations
Maximum value
75th percentile
Interquartile range Median
25th percentile
Minimum value
Figure 9. Effects of spatial averaging on the statewide distribution of L-skew for each precipitation duration.
Contouring Distribution Parameters S C
( p ) i = ( p ) i ( p ) i, (30)
The values for 1, 2, and 3 were transformed where
into the , , and parameters (see eqs. 1214 and (p)i = error associated with parameter p (, , or )
1720) of either the GLO distribution (15-minute to at station i;
24-hour duration) or the GEV distribution (1- to 7-day S(p)i = value of the parameter for the station (control
duration). The resulting parameters then were con- point); and
toured using interpolation based on the geostatistical C(p)i = value of the parameter for the station from
method of kriging to produce 37 maps representing the the contours.
spatial magnitude and variation of the parameters. The error thus defined is only indicative of the true
A complete discussion of the kriging method error because each station was used in the development
is beyond the scope of this report. Standard geostatisti- of the contour maps. An independent means to measure
cal terminology is used in this section. A thorough error is not available.
discussion of the kriging method is available in Carr The results of the map error analysis are listed in
(1995), Davis (1986), and Deutsch and Journel (1992). table 3. The table lists the mean across the State for
A linear form of the theoretical semivariogram each parameter and its SD. Also listed are the mean
was chosen to model the semivariance of each parame- contour error, mean absolute contour error, RMSE, and
ter for all durations. A variety of search radii, sample percent change from SD to RMSE for each parameter.
sizes, and grid densities were investigated for genera- Each of these statistics is a weighted value, which
tion of the contour maps. A maximum search radius of means that record length of each station was considered
about 185 mi and a grid spacing of about 3 mi were in the computation.
used. This grid spacing translates to approximately The mean contour error is the mean difference
67,000 grid cells for the State. A sample size of 10 sta- between values of the control points and values from the
tions was used for the 15-minute and hourly data (15- map; mean contour errors near zero are desirable. This
minute to 24-hour duration), and a sample size of 20 sta- error reflects potential bias in the map. Additionally, the
tions was used for the daily data. The larger sample size RMSE was calculated by the following:
for the daily data is justifiable because the density of the n
daily stations (3.23 per 1,000 mi2) is much greater than 2
that of either the 15-minute or hourly stations (0.646 or ri [( p )i ]
1.02 stations per 1,000 mi2). i=1
RMSE = ----------------------------------
-, (31)
n
DEPTH-DURATION FREQUENCY OF
PRECIPITATION FOR TEXAS ri
i=1
The DDF for any location in Texas can be esti- where
mated from the maps (figs. 1046 at end of report) and ri = record length associated with station i; and
from procedures discussed in this section. Contour (p)i = error associated with parameter p (, , or )
maps depicting the location () parameters for durations at station i.
of 15 minutes to 7 days are shown in figures 1023. RMSE is analogous to the SD, and therefore the
Contour maps depicting the scale () parameters for RMSE is comparable to the SD of the control points.
durations of 15 minutes to 7 days are shown in figures The percent change from SD to RMSE, 100*[(RMSE
2437. Contour maps depicting the shape () parame- SD) / SD], is considered indicative of the improvement
ters for durations of 15 minutes to 24 hours are shown in parameter estimation by using the map contours
in figures 3846. The figures have been checked for rather than simply using the corresponding statewide
internal consistency; for example, parameters of one mean. Percent changes from SD to RMSE for all param-
duration are less than parameters of a larger duration eters and durations, except for 1- to 7-day , are all
for a given location in Texas. greater than 15 percent. Thus, the author believes the
Analysis of the error associated with each of the use of the contour maps results in more accurate esti-
maps was done. The error associated with a station is mates of the parameters than by simply using a state-
defined as wide mean. The statewide means in table 3 are used for
[Note: Contour error is defined as parameter value for control point minus value from contour map. SD, standard
deviation; RMSE, root mean square error; min, minutes; hr, hours; --, maps for not presented because SD to
RMSE percent change is less than 15 percent]
Control points
Contours
(stations) SD to RMSE
Duration Statewide SD of RMSE of percent
Ratio of SD Mean contour Mean absolute
mean of statewide contour change
to mean error contour error
control points mean map
Location parameter (), in inches
15 min 0.896 0.178 0.199 0.00312 0.0755 0.127 28.7
30 min 1.24 .206 .166 .00730 .0897 .119 42.2
60 min 1.58 .275 .174 .00289 .117 .154 43.9
1 hr 1.52 .317 .209 .0171 .102 .150 52.5
2 hr 1.87 .422 .226 .0229 .121 .181 57.2
3 hr 2.03 .474 .233 .00819 .137 .196 58.5
6 hr 2.33 .563 .242 .0156 .152 .219 62.2
12 hr 2.66 .667 .258 .0294 .163 .240 64.0
24 hr 3.00 .783 .261 .0270 .184 .270 65.6
1 day 2.84 .673 .237 .00536 .158 .223 66.8
2 days 3.22 .776 .241 .0293 .172 .242 68.8
3 days 3.43 .839 .245 .0438 .193 .267 68.2
5 days 3.77 .922 .245 .0134 .204 .283 69.3
7 days 4.08 1.00 .245 .0135 .216 .305 69.6
Scale parameter (), in inches
15 min .179 .0357 .199 .00400 .0193 .0276 22.5
30 min .247 .0387 .141 .000862 .0227 .0288 25.6
60 min .331 .0499 .151 .00199 .0300 .0392 21.4
1 hr .338 .0518 .153 .000967 .0281 .0393 24.1
2 hr .426 .0751 .176 .000453 .0343 .0466 38.0
3 hr .469 .0928 .198 .00105 .0377 .0506 45.5
6 hr .547 .118 .216 .000018 .0457 .0617 47.9
12 hr .624 .147 .236 .00183 .0511 .0700 52.4
24 hr .713 .178 .250 .00844 .0604 .0794 55.4
1 day 1.08 .258 .234 .00429 .0949 .121 53.2
2 days 1.25 .317 .254 .00108 .111 .145 54.4
3 days 1.36 .334 .246 .00239 .122 .157 52.9
5 days 1.49 .374 .251 .00721 .130 .168 55.1
7 days 1.61 .397 .247 .00958 .143 .185 53.4
Shape parameter (), dimensionless
15 min .132 .0782 .592 .0104 .0426 .0612 21.7
30 min .119 .0602 .506 .00307 .0342 .0438 27.2
60 min .142 .0551 .388 .00694 .0311 .0410 25.7
1 hr .165 .0540 .327 .000655 .0327 .0423 21.7
2 hr .180 .0503 .279 .00359 .0313 .0400 20.5
3 hr .190 .0508 .267 .00543 .0326 .0390 23.2
6 hr .209 .0538 .257 .000309 .0284 .0362 32.7
12 hr .213 .0502 .236 .00330 .0274 .0346 31.1
24 hr .219 .0463 .211 .000238 .0261 .00331 28.4
1 day .0954 .0765 .802 -- -- -- --
2 days .0867 .0825 .952 -- -- -- --
3 days .0862 .0818 .949 -- -- -- --
5 days .0750 .0769 1.03 -- -- -- --
7 days .0615 .0755 1.23 -- -- -- --
SUMMARY 25
and (2) improve the accuracy of LCV and L-skew Dalrymple, Tate, 1960, Flood-frequency analyses: U.S.
estimates. A corrected L-scale subsequently was esti- Geological Survey Water-Supply Paper 1543A, 80 p.
mated using the corrected mean and the spatially aver- Davis, J.C., 1986, Statistics and data analysis in geology:
aged LCV for each duration and for each station. New York, Wiley, 646 p.
Finally, the corrected mean, the corrected Lscale, and Deutsch, C.V., and Journel, A.G., 1992, Geostatistical
the averaged L-skew for each station and duration were software library and users guide: New York, Oxford
transformed into the location, scale, and shape parame- University Press, 340 p.
ters for both the GLO or GEV distributions. Fischer, R.A., 1929, Moments and product moments of
sampling distributions: Journal of London Mathematical
The location, scale, and shape parameters of
Society, 2(30), 199 p.
the GLO and GEV distributions for each station and
Frederick, R.H., Meyers, V.A., and Auciello, E.P., 1977, Five-
duration were contoured using an interpolation scheme
to 60-minute precipitation frequency for the
based on the geostatistical method of kriging to produce
eastern and central United States: Washington, D.C.,
37 maps depicting the spatial variation and magnitude National Oceanic and Atmospheric Administration
of each parameter. Contour maps of the shape parameter Technical Memorandum NWS HYDRO35, 36 p.
for the GEV distribution for durations of 1 to 7 days are Greenwood, J.A., Landwehr, J.M., Matalas, N.C., and Wallis,
not presented; the root mean square errors of prelimi- J.R., 1979, Probability weighted momentsDefinition
nary maps for 1- to 7-day shape parameters were not and relation to parameters of several distributions
appreciably smaller than the statewide standard devia- expressible in inverse form: Water Resources Research,
tion. Therefore, a single statewide mean shape parame- v. 15, no. 5, p. 1,0491,054.
ter was used for 1- to 7-day durations. The DDF for any Hershfield, D.M., 1962, Rainfall frequency atlas of the
location in Texas can be estimated using the appropriate United States for durations from 30 minutes to 24 hours
maps and distribution function. and return periods from 1 to 100 years: Washington,
To reduce the effects of potential errors associated D.C., U.S. Weather Bureau Technical Paper 40, 61 p.
with the maps, precipitation intensity-duration Hosking, J.R.M., 1986, The theory of probability weighted
frequency curves, known as IDF curves, were devel- moments: Yorktown Heights, N.Y., IBM Research
oped on the basis of the relation between precipitation Division, T.J. Watson Research Center, Research Report
RC12210, 160 p.
intensities (inches per hour) and duration. The IDF
______1990, L-momentsAnalysis and estimation of
curves for selected localities in the State are presented.
distributions using linear combinations of order statis-
The relatively smooth transition of the IDF curve for
tics: Journal Royal Statistical Society B, v. 52, no. 1,
the selected localities from short to long durations is p. 105124.
notable and encouraging. In other words, the parameter
______1991a, Approximations for use in constructing
maps produce consistent precipitation depths. IDF or L-moment ratio diagrams: Yorktown Heights, N.Y.,
DDF estimates or curves derived from this report are IBM Research Division, T.J. Watson Research Center,
considered more accurate than estimates from previous Research Report RC16635, 3 p.
studies because of (1) the greater number of stations and ______1991b, FORTRAN routines for use with the method
longer record available for this study, (2) the use of of L-moments, Version 2: Yorktown Heights, N.Y., IBM
more powerful statistics (L-moments), which were not Research Division, T.J. Watson Research Center,
available at the time the other studies were completed, Research Report RC17097, 117 p.
and (3) the use of more flexible and appropriate three- ______1992, Moments or L moments? An example compar-
parameter distributions than the two-parameter Gumbel ing two measures of distributional shape: The American
distribution used in previous studies. Statistician, v. 46, no. 3, p. 186189.
______1994, The four-parameter kappa distribution: IBM
SELECTED REFERENCES Journal of Research and Development, v. 38, no. 3,
p. 251258.
Bomar, G.W., 1995, Texas weather: Austin, University of Hosking, J.R.M., and Wallis, J.R., 1993, Some statistics
Texas Press, 275 p. useful in regional frequency analysis: Water Resources
Carr, J.R., 1995, Numerical analysis for the geological Research, v. 29, no. 2, p. 271281.
sciences: Englewood Cliff, N.J., Prentice-Hall, 592 p. ______1995, A comparison of unbiased and plotting-position
Carr, J.T., Jr., 1967, The climate and physiography of Texas: estimators of L-moments: Water Resources Research,
Texas Water Development Board Report 53, 27 p. v. 31, no. 8, p. 2,0192,025.
SELECTED REFERENCES 27
Figure 10
102 o 100 o
36 o
755
0.80
0.80
00..7
98 o
96 o
34 o
000
9 55
11..0
0.9
0.
106 o 0. 8
32 o 104 o 5 0.9
0.770
0 955
0
1 ..000
00.
00 0
.6655
00..88
00.8
.855
0.75
1.05
0
30 o 95 1.1
95
00. .
00
11.0.0
. 05
1
28 o
1.00
0.95
EXPLANATION
00..9900
0.85 Line of equal location parameter
Interval 0.05 inch
26 o
Figure 10. Location () parameter of generalized logistic (GLO) distribution for 15-minute precipitation duration in
Texas.
36 o
98 o
96 o
34 o
0 .9
5
1.00
1 .35
1.40
5
11..005
106 o
1 .3
32 o 104 o
0
0.
70 155
0.
11..1
70
00..880 .9 0 1.40
1 .3
0 0
5
1.10
1.2 0
1 .25 1.50
1
30 o
.3
0
1.45 55
1.
1.10 0
1.6
1
.4
5
28 o
EXPLANATION
Figure 11. Location () parameter of generalized logistic (GLO) distribution for 30-minute precipitation duration in
Texas.
Figure 11 29
Figure 12
102 o 100 o
36 o
98 o
96 o
34 o 1.2
106 o
1.7
32 o 104 o
3
1.
1.2 4
1.
1.
1
1.5
1 .6
1.2 1 .7
o
30
1 .8
1 .9
2 .0
1.7
28 o
EXPLANATION
1.9
26 o
Figure 12. Location () parameter of generalized logistic (GLO) distribution for 60-minute precipitation duration in
Texas.
36 o
98 o
96 o
34 o
1.1
1.7
106 o 1.2
32 o 104 o
0.
7
11..3
0 .8 9
0.
1.0
1.1
11..66
1.4
11..5 1 .7
1.2 5 2.1
1 .8
o
30
1.8
1 .0
1.9
1.1
2.0
1
.7
28 o
EXPLANATION
26 o
1.9
Figure 13. Location () parameter of generalized logistic (GLO) distribution for 1-hour precipitation duration in
Texas.
Figure 13 31
Figure 14
102 o 100 o
36 o
1.3
1.4
98 o
96 o
34 o
5
1..5
106 o
32 o 104 o
2.2
0.
0.
9
0.
8
9
1.0 1.2
.6
1
1 .1 1.7
1.1
1.8 2 .1
1.9
2 .0
2 .3
1. 2
30 o
2 .7
3
1. 1 .1
2.4
2.4
1.3 2 .2 2.5 .6
2
28 o
EXPLANATION
26 o
2.4
Figure 14. Location () parameter of generalized logistic (GLO) distribution for 2-hour precipitation duration in
Texas.
36 o
98 o
96 o
34 o 1.4
106 o
32 o 104 o
2.4
1 .0 1.6
2
1.
1.2 1.8
2.2
2.0
0
3.
30 o
2. 4
1. .2
2.4
2 .4
4
1
1.
0
1.4
2.6 .8
2
28 o
EXPLANATION
26 o
2.6
Figure 15. Location () parameter of generalized logistic (GLO) distribution for 3-hour precipitation duration in
Texas.
Figure 15 33
Figure 16
102 o 100 o
36 o
1 .4
Amarillo
98 o
96 o
34 o
1.6
106 o
2.8
104 o
1.
32 o
8
1.0
2 2 .0
1. 1.4
2.2 2.6
1.6 2.4 3.4.6
3
1. 1.2 .0
o
30
2.8
4
2 .8
1
3 .0
3.2 .4
3
28 o
EXPLANATION
2.5
26 o
2.8
3 .0
Figure 16. Location () parameter of generalized logistic (GLO) distribution for 6-hour precipitation duration in
Texas.
36 o
1.6
98 o
96 o
34 o
1.8
106 o 2
32 o 104 o
2.0 3.
1 .2 1 .6
1.4
1.4
2 .2
2.4 .0
33.0
2.6 4 .0
.6 4.2
1
30 o
2.8 3 .2
33..22
1.
1.6
2
3.4 6
1.0 3 .8
3.
28 o
EXPLANATION 2.
6
2.
6
26 o
3 .2
3 .4
Figure 17. Location () parameter of generalized logistic (GLO) distribution for 12-hour precipitation duration in
Texas.
Figure 17 35
Figure 18
102 o 100 o
36 o
1.75
2.00
98 o
96 o
34 o
75
3.
106 o
2 .2
32 o 104 o 5
3.
2 .00
50
1.25
50
1. 1.50
2.5 0
2 .7 5
5
3 .00 3.2 4.75
4.50
1. 11.5.5
30 o
50
5
3.
75 00
7
3.
0
4 .0
3 .5 5
1. 2 2
5
0 4.
28 o
3.
EXPLANATION
00
26 o 3.75
Figure 18. Location () parameter of generalized logistic (GLO) distribution for 24-hour precipitation duration in
Texas.
5
36 o
1 .7
2.00
98 o
96 o
34 o
2.25
3.50
3.0
0
2.00
106 o 2.25 2.75
32 o 104 o
3 .25
1.20
1 . 75 2 .0 2.5
1.
0 0
25
1.75
0
1.5
1.75
5 0
30 o
1.50
1.2 3 .2 4.5 4.75
5 3.50
3.25 3 .50
3 .7 5 5
1.75 4 .0 04 . 2
1.50 3 . 00
2 .75
28 o
EXPLANATION
26 o
Figure 19. Location () parameter of generalized extreme-value (GEV) distribution for 1-day precipitation duration
in Texas.
Figure 19 37
Figure 20
102 o 100 o
0
36 o
2 .0
2 .2 5
Amarillo
98 o
96 o
34 o
3 .5
0
2.25
2.00 3.0 0 4.00
25
106 o
104 o
3.
32 o
2.25
1 .3
2 .5
2.00
5
2.7
5 5
1 .5 0 1.7
1.75
2.00 5
.7
3
30 o 1
1. .5 0 3.75
25 0
1.
5
5.5
4.00
5.2
0
5
4 .2 5
5
2.00 3.0 4 .5 0 4 .7
3.5 0
0 3.25 0
1.75 5 .0
28 o
EXPLANATION
26 o
3.75
Figure 20. Location () parameter of generalized extreme-value (GEV) distribution for 2-day precipitation duration
in Texas.
2.00
36 o 5
2.2
5
2 .7 98 o
o 96 o
34
4.50
.00
3 4.25
0
3 .25
3. 5
5
3.75
106 o
2.2
32 o 104 o
0
1.5
2 .0 .5
1.40
0
0
0
1.50
2.0
2.00
1 .7 5
2 .25
5
2.75 4 .7
44..225
5
2.25 5
5 .2
30 o 11.5.75
0
1.71 4.00
5.7
50
.5
5.5
5
0
4.
5
5 0
0
4 .2 0
5.
2.00 3
.0
5
0
3.2
28 o
EXPLANATION
26 o
0
4.0
Figure 21. Location () parameter of generalized extreme-value (GEV) distribution for 3-day precipitation duration
in Texas.
Figure 21 39
Figure 22
102 o 100 o
2.25
36 o
2.5 0
98 o
2 .5 0
96 o
34 o
5.00
4.
3.00
00
2.75 2.50
2.25 3..5500
4.75
106 o
32 o 104 o
22.75 3 .7 5
0 2 . 5 .7 5
2 .0 2. 2 0
1.75
5
00
1.50
2.
1.75 3.
.2255
3
2 .25 5
4..22 5 50 4.75 5 .2
5
2.50 4.
2.75 0
.0
12 .00
5
30 o
1 .75
.5 .75
4.25
1
0
2.0
6. 0
0
6.
25
0
2.50 5
2.75 3 . 75 5 .7
3.5 0
3 .25
28 o
EXPLANATION
4.00
26 o
Figure 22. Location () parameter of generalized extreme-value (GEV) distribution for 5-day precipitation duration
in Texas.
2.7 5
98 o
96 o
34 o
5
.50 5.2
4
3.00 3.2
5 .00
5
3 .75
5
106 o
2.7
32 o 104 o
4. 2
2.7 5
1.7
2 .0 0
5 0
2 .2 2.50 3.0 4 .0
5
2 .50
5
0
1.75 2.
2.0205 3 .5
25
Guadalupe 0
2. 0 50
Mountains 3.00 .0 5.
25
5
Davis
5.
2 . 50 75
Mountains 4. 5 .255
66.2
30 o 22
1 . .0000 .7
75
5
2.75
2.
4 .50
6. 0
2.25
00
6.
75
5
2.25 25
2.50 6.
.7
3
Chisos 5
50
Mountains 4.
28 o
EXPLANATION
3.50
Balcones escarpment
0 50 100 150 200 MILES
Figure 23. Location () parameter of generalized extreme-value (GEV) distribution for 7-day precipitation duration
in Texas.
Figure 23 41
Figure 24
102 o 100 o
36 o
00. .
0.1
2200
8
0.
16
0.
16
98 o
96 o
34 o
00..1166
00..2
200
0.
18
18
32 o
106 o
104 o 0.
00..1
188
0.16 00.1 6
. 16
0.18
0.18
30 o
00..2200
.222
0.2
28 o
EXPLANATION
00.2.200
0.22 Line of equal scale parameter
Interval 0.02 inch
2 2
0.
26 o
Figure 24. Scale () parameter of generalized logistic (GLO) distribution for 15-minute precipitation duration in
Texas.
36 o
0..22
00..
06
2244
98 o
00.2.2
96 o
34 o
22
0.26
0.24
00. .
106 o
2288
32 o 104 o
00..22
44
4
0.2
0.
22
0.
22
0.222
244
30 o
00..2
00.2
4
0..
2288 0.
0..26
26
30
28 o
EXPLANATION
26 o
Figure 25. Scale () parameter of generalized logistic (GLO) distribution for 30-minute precipitation duration in
Texas.
Figure 25 43
Figure 26
102 o 100 o
36 o
0.
28
98 o
96 o
34 o
0.
32
106 o
104 o
0.3
32 o 4
0
0 .3
0
0 .3
.3
0
4
32
0.
30 o 34
0.
6
0.3
8
0.3
28 o
40
EXPLANATION
0.
Figure 26. Scale () parameter of generalized logistic (GLO) distribution for 60-minute precipitation duration in
Texas.
36 o 00..3300
98 o
96 o
34 o
0.32
0.32
106 o
32 o 104 o
.322
00.3
.28
00.2 8
00.3.300
24
0.
26
.3 4
00.34
0.
0.38
0.
38
30 o
0.3 8
.366
0.3
0
0.
40
0.
40
28 o
EXPLANATION
26 o
Figure 27. Scale () parameter of generalized logistic (GLO) distribution for 1-hour precipitation duration in Texas.
Figure 27 45
Figure 28
34
102 o 100 o
0.
36 o
0.36
98 o
96 o
34 o
0.40
0.40
0.40
0.40
106 o
0 .3
104 o 8 4
32 o
0.4
0 .36
6
0.2
0. 34
0. 46
42 0.
8
0 .32
0.2
0.52
8
0.4
30 o
00..
0 .5 0
0.50
3300
0.5
0.52
8
4
4
0. 2
0.5
0 0 .5
4
0 .554
28 o
EXPLANATION
0
0 .5
26 o
Figure 28. Scale () parameter of generalized logistic (GLO) distribution for 2-hour precipitation duration in Texas.
36 o
400
0..4
0
98 o
96 o
34 o
0.30
106 o
32 o 104 o
40
00..40
00.355
0.4
5
00.5
.500 600
0.
5555 00..6
00. .
o
30
30
555
00. .5
0.6
0 . 60
5
28 o
EXPLANATION
26 o
45
0.
0 50 100 150 200 MILES
Figure 29. Scale () parameter of generalized logistic (GLO) distribution for 3-hour precipitation duration in Texas.
Figure 29 47
Figure 30
102 o 100 o
36 o
Amarillo
0.4
0. 4
98 o
96 o
34 o
00.
6.6
106 o
32 o 104 o
0.3
0.3
0 .4
00..55
0.6
.6
0 .6
30 o
0.8
00.3
3
0.7
0 .8
0.7
28 o
EXPLANATION
26 o
7
0.
0 50 100 150 200 MILES
Figure 30. Scale () parameter of generalized logistic (GLO) distribution for 6-hour precipitation duration in Texas.
0.4
36 o
98 o
96 o
34 o
0.
.77
0
106 o
32 o 104 o
0.3
5
0..5
0.4
0 .6
0.7
30 o
1.0
0.3
0 .9
.9
0
00.8
8
00. .8
.8
0.7
28 o
EXPLANATION
26 o
8
0.
0 50 100 150 200 MILES
Figure 31. Scale () parameter of generalized logistic (GLO) distribution for 12-hour precipitation duration in
Texas.
Figure 31 49
Figure 32
102 o 100 o
36 o
0.50
98 o
96 o
34 o
0.80
106 o
32 o 104 o
0 .3 5
0.50
0
0 .4
800
0.60
0..8
0.90
0.90
0
0.70
30 o 1.20
0.4
0
0
0.3 0
1 .1
0 0. 9 00
1.
28 o
EXPLANATION 80
0.
26 o
Figure 32. Scale () parameter of generalized logistic (GLO) distribution for 24-hour precipitation duration in
Texas.
7
36 o 00. . 7
0.8
0.8
98 o
96 o
34 o
0.9
1.3
106 o
32 o 104 o
.2
0.88
1
0 .5
1.0
0.7
1.6
0.6
30 o
1.1 1.2 .3
1.7
1.2
1
3
11. .4
.3
1 .5 1. 6
1
28 o
EXPLANATION
26 o
1.2
1.4
Figure 33. Scale () parameter of generalized extreme-value (GEV) distribution for 1-day precipitation duration in
Texas.
Figure 33 51
Figure 34
0.8
102 o 100 o
36 o
Amarillo
98 o
0 .9
96 o
34 o
1.6
106 o
0.9
32 o 104 o
11.4.4
0.55
0.8
1 .1
0.6 1.5
.5
0.7 1.0
0 .8
0 1.2
1.9
.7
30 o
1. 5
1 .4
.5
1
2.2
2.0
1 1 .6
1
9
1.
.3
11.8
7
1.
EXPLANATION 28 o
Edwards Plateau
Balcones escarpment 26 o
1.4
1.5
Figure 34. Scale () parameter of generalized extreme-value (GEV) distribution for 2-day precipitation duration in
Texas.
0.8
102 o 100 o
36 o
.00
11.
98 o
96 o
34 o
6
1.
106 o
1.6
32 o 104 o
6
0.
11..22
0.8
0.8
.0
22.0
0.9
1 .0
0. 8
.0
30 o 2.2
1 .4 6
11..88
1.
1 .6
22..00
1.
2
28 o
EXPLANATION
26 o
1.6
Figure 35. Scale () parameter of generalized extreme-value (GEV) distribution for 3-day precipitation duration in
Texas.
Figure 35 53
Figure 36
102 o 100 o
36 o
1.0
98 o
96 o
34 o
11..88
106 o
32 o 104 o
11..88
0 .8 11.0
0.6
.0 1.4
1.6
2.0
0.8
0 .8
1.0
1. 0
1.2
2.2
2 .4
30 o 0 .8
7
1.
1.0
28 o
EXPLANATION
indicate depression
1 .8
26 o
Figure 36. Scale () parameter of generalized extreme-value (GEV) distribution for 5-day precipitation duration in
Texas.
1.0
102 o 100 o
36 o
1.2
98 o
96 o
34 o
1.8
2.0
106 o
32 o 104 o 1.6
1.4
0.8
1..00
2.0
1 .1
22.4.4
1..0
0 2.6
30 o
1.8
4
2.
1.
0 1.
6 2.2
28 o
EXPLANATION
1 .4
26 o
Figure 37. Scale () parameter of generalized extreme-value (GEV) distribution for 7-day precipitation duration in
Texas.
Figure 37 55
Figure 38
102 o 100 o
36 o
0.20
98 o
96 o
34 o
0.15
..1100
0.1
0
106 o
.15
32 o 104 o
.05
0
0
0
0 .2
5
.1 5
0 .1
30 o
.10
0
28 o
EXPLANATION
26 o
Figure 38. Shape () parameter of generalized logistic (GLO) distribution for 15-minute precipitation duration in
Texas.
36 o
00..2200
98 o
96 o
34 o
0
.1
5
00..10
00. .10
10
10
106 o
32 o 104 o
0
.0
5
5
1
30 o
0.
.1100
00.
28 o
EXPLANATION
26 o
Figure 39. Shape () parameter of generalized logistic (GLO) distribution for 30-minute precipitation duration in
Texas.
Figure 39 57
Figure 40
102 o 100 o
36 o
22
0.
98 o
34 o 20 96 o
0.
.15
0
0
.1
106 o
0
32 o 104 o
0
.10
0.1
5 5
.1
30 o
0 0.20
00
..115
5
28 o
EXPLANATION
26 o
Figure 40. Shape () parameter of generalized logistic (GLO) distribution for 60-minute precipitation duration in
Texas.
36 o
22
0.
98 o
96 o
34 o
.20
0
5
0.1
106 o
32 o 104 o
20
0.
.15
0.10 0
0 .20
30 o
0.15
28 o
EXPLANATION
26 o
Figure 41. Shape () parameter of generalized logistic (GLO) distribution for 1-hour precipitation duration in
Texas.
Figure 41 59
Figure 42
0.24 102 o 100 o
36 o
22
0.
98 o
96 o
34 o
.20
0
106 o
32 o 104 o
0 .15
0.1
5
0.10
.20
0
30 o
.2 0
0
5
0 .1
28 o
EXPLANATION
26 o
Figure 42. Shape () parameter of generalized logistic (GLO) distribution for 2-hour precipitation duration in
Texas.
36 o
. 22
0
98 o
96 o
34 o
.20 0.15
0
106 o
32 o 104 o
0 .1 5 .2
0
0.10 0
0
30 o .2
0
28 o
EXPLANATION
26 o
Figure 43. Shape () parameter of generalized logistic (GLO) distribution for 3-hour precipitation duration in
Texas.
Figure 43 61
Figure 44
102 o 100 o
36 o
Amarillo
0.25
98 o
96 o
34 o
5
0.1
106 o
.2 0
32 o 104 o
0 0.20
0.1
.15 25
.25 0.
0
.20
30 o 0
0.10
28 o
EXPLANATION
26 o
Figure 44. Shape () parameter of generalized logistic (GLO) distribution for 6-hour precipitation duration in
Texas.
36 o
98 o
96 o
34 o
0.25
106 o
32 o 104 o
0
.2
0
20
0.
0.15 5
.2
0.10 0
.25
0
20
30 o
0.
28 o
EXPLANATION
26 o
Figure 45. Shape () parameter of generalized logistic (GLO) distribution for 12-hour precipitation duration in
Texas.
Figure 45 63
Figure 46
EXPLANATION
Figure 46. Shape () parameter of generalized logistic (GLO) distribution for 24-hour precipitation duration in
Texas.
5
4
3
Representative best-fit
2 100-year precipitation intensity-
duration curve for El Paso, Texas
0.7
0.5
0.4
0.3
0.2
0.1
1 hour
0.05
24 hours
60 minutes
0.04
0.03
5 10 15 30 1 2 3 6 12 1 2 3 5 7 10
MINUTES HOURS DAYS
STORM DURATION
EXPLANATION
Precipitation intensity of 100-year storm for Abilene, Texas Precipitation intensity of 100-year storm for El Paso, Texas
Precipitation intensity of 100-year storm for Amarillo, Texas Precipitation intensity of 100-year storm for Lubbock, Texas
Precipitation intensity of 100-year storm for Austin, Texas Precipitation intensity of 100-year storm for Orange, Texas
Precipitation intensity of 100-year storm for Brownsville, Texas Precipitation intensity of 100-year storm for San Antonio, Texas
Precipitation intensity of 100-year storm for Dallas, Texas Precipitation intensity of 100-year storm for Texarkana, Texas
Figure 47
Figure 47. Precipitation intensity-duration curves of 100-year storm for selected localities in Texas.
65
Figure 48
102 o 100 o
5.0
36 o
4.5
98 o
96 o
34 o
6 .5
5 .0
106 o
32 o 104 o
6.0
0
3.
3 .5
4.5 5
5.
4.0 6.0
3.0
4.0
7.0
6.5
7.0
10.5
7.5
9 .0 8.5
3.5
30 o
7.0
8.5 8.0
7 .5
9.0
3 .0
9.0
0
8. 5
9. 0.0
1
10.5
EXPLANATION
28 o
Balcones escarpment
0 50 100 150 200 MILES
Figure 48. Depth of 100-year storm for 6-hour precipitation duration in Texas.
5
Representative best-fit 100-year
precipitation intensity-duration
4 curve for corresponding locality
2.5
1.5
1
0.9
0.8
0.5
1 hour
60 minutes
0.4
5 10 15 30 2 3 6 12
MINUTES HOURS
STORM DURATION
EXPLANATION
Precipitation intensity of 100-year storm for Amarillo, Texas Precipitation intensity of 100-year storm for Orange, Texas
Precipitation intensity of 100-year storm for Amarillo from HYDRO35 Precipitation intensity of 100-year storm for Orange from
(Frederick and others, 1977) HYDRO35
Precipitation intensity of 100-year storm for Amarillo from TP40 (Hershfield, 1962) Precipitation intensity of 100-year storm for Orange from TP40
Figure 49
Figure 49. Precipitation intensity-duration curves of 100-year storm for Amarillo and Orange, Texas.
67
APPENDICES
Depth-Duration Frequency of
Precipitation for Texas
APPENDICES
APPENDICES 69
Appendix 1. Fifteen-minute precipitation stations in Texas with at least 10 years of annual maxima data through
1994
[Station name: LCRA, Lower Colorado River Authority; FAA, Federal Aviation Administration; WSO, Weather Service
Office; WSMO, Weather Service Meteorological Observatory.
Appendix
Climatic 1. Fifteen-minute
region: precipitation
TP, Trans-Pecos; stationsHP,
NC, North Central; in Texas with atEP,
High Plains; least 10 years
Edwards of annual
Plateau; maxima
UC, Upper data
Coast; SC,through
South
1994Continued
Central; ST, South Texas; LRP, Low Rolling Plains; ET, East Texas.]
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
0174 Alpine 3022' 10340' TP 17 1978 1994
0206 Alvord 4 NE 3323' 9739' NC 23 1971 1994
0248 Andrews 3219' 10232' HP 23 1972 1994
0262 Anna 3321' 9631' NC 16 1979 1994
APPENDICES 71
Appendix 1. Fifteen-minute precipitation stations in Texas with at least 10 years of annual maxima data through
1994Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
4191 Hindes 2843' 9848' ST 18 1977 1994
4257 Honey Grove 3335' 9554' NC 22 1971 1994
4278 Hords Creek Dam 3151' 9934' LRP 11 1984 1994
4309 Houston Addicks 2946' 9539' UC 11 1984 1994
APPENDICES 73
Appendix 1. Fifteen-minute precipitation stations in Texas with at least 10 years of annual maxima data through
1994Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
8305 Sierra Blanca 3111' 10521' TP 17 1978 1994
8446 Somerville Dam 3020' 9632' SC 21 1974 1994
8531 Spicewood 1 S 3027' 9810' EP 11 1984 1994
8563 Springtown 4 S 3254' 9741' NC 18 1977 1994
[Station name: WSO, Weather Service Office; LCRA, Lower Colorado River Authority; FAA, Federal Aviation
Administration; WSCMO, Weather Service Contract Meteorological Observatory; WB, Weather Bureau; WSMO, Weather
Service Meteorological Observatory; WSFO, Weather Service Forecast Office; CAA; Civilian Aeronautics Administration.
Appendix
Climatic 2. Hourly
region: precipitation
LRP, Low stations
Rolling Plains; EP, in Texas Plateau;
Edwards with at least 10 years ofNC,
TP, Trans-Pecos; annual maxima
North Central;data through
HP, High 1994
Plains; ST,
Continued
South Texas; UC, Upper Coast; SC; South Central; ET, East Texas.]
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
0016 Abilene WSO Airport 3226' 9941' LRP 89 1906 1994
0050 Adamsville 3117' 9809' EP 23 1963 1985
0174 Alpine 3022' 10340' TP 24 1971 1994
0206 Alvord 4 NE 3323' 9739' NC 52 1942 1994
APPENDICES 75
Appendix 2. Hourly precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
1185 Buenavista 2 NNW 3115' 10240' TP 22 1942 1963
1246 Burleson 3 SSE 3230' 9718' NC 13 1982 1994
1267 Bushland 1 WSW 3511' 10205' HP 12 1940 1951
1304 Cadiz 2826' 9757' SC 13 1940 1952
APPENDICES 77
Appendix 2. Hourly precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
3686 Granger Dam 3042' 9720' NC 15 1980 1994
3691 Grapevine Dam 3258' 9703' NC 46 1949 1994
3771 Groesbeck 2 3132' 9632' NC 18 1977 1994
3831 Guyer 3408' 9856' LRP 13 1940 1952
APPENDICES 79
Appendix 2. Hourly precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
6136 Muleshoe 2 3413' 10244' HP 53 1941 1994
6177 Nacogdoches 3137' 9438' ET 42 1947 1994
6210 Navarro Mills Dam 3157' 9642' NC 33 1962 1994
6211 Navasota 3023' 9607' ET 12 1941 1952
APPENDICES 81
Appendix 2. Hourly precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
8944 Texarkana Dam 3318' 9410' ET 18 1955 1972
8996 Thompsons 3 WSW 2929' 9538' UC 38 1957 1994
9037 Tinnin Ranch 3118' 10400' TP 28 1942 1969
9163 Truscott 5 W 3345' 9955' LRP 55 1940 1994
9916 Wright Patman Dam and Lake 3318' 9410' ET 14 1981 1994
9976 Zapata 2653' 9918' ST 55 1940 1994
1
The record of this station was considered auxiliary and therefore combined with the identified station in the table.
Depth-Duration Frequency of
0025
0034
Acker Ranch
Ackerly
2808'
3232'
9831'
10143'
ST
HP
17
47
1978
1948
1994
1994
Precipitation
0050 Adamsville
for Texas 3117' 9809' EP 25 1963 1987
0068 Adrian 3517' 10240' HP 11 1938 1948
0081 Agua Nueva
APPENDICES 2654' 9836' ST 17 1948 1964
0120 Albany 3244' 9917' NC 94 1901 1994
APPENDICES 83
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
0408 Atlanta 3307' 9410' ET 51 1930 1994
0428 Austin WSO Airport 3018' 9742' EP 65 1930 1994
0430 Austin Dam 3018' 9747' EP 18 1948 1965
0432 Austin Montopolis Bridge 3015' 9741' EP 16 1948 1963
APPENDICES 85
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
1239 Burkett 3200' 9914' LRP 47 1948 1994
1245 Burleson 2 SSW and 12451 3231' 9720' NC 47 1948 1994
1248 Burlington 3 WSW 3100' 9702' NC 28 1948 1975
1250 Burnet 3044' 9814' EP 96 1896 1994
APPENDICES 87
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
2048 Cotulla 2826' 9915' ST 86 1901 1994
2050 Cotulla FAA Airport 2827' 9913' ST 33 1949 1981
2066 Cox Ranch 3121' 9953' EP 18 1948 1965
2080 Crandall 3238' 9628' NC 35 1960 1994
APPENDICES 89
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
2925 Ennis 3220' 9638' NC 49 1940 1991
3000 Evadale 3020' 9405' ET 42 1946 1994
3005 Evant 4 SW 3127' 9813' NC 41 1941 1994
3033 Fabens 3130' 10609' TP 30 1948 1977
3284 Fort Worth Meach WSO Airport 3249' 9721' NC 11 1946 1994
3286 Fort Worth Vickery Blvd. 3244' 9720' NC 19 1948 1994
3298 Four Notch Guard Station 3039' 9525' ET 25 1940 1964
3299 Fowlerton 2 NW 2829' 9852' ST 78 1913 1994
APPENDICES 91
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
3680 Grandfalls 3 SSE 3118' 10250' TP 81 1909 1994
3685 Granger 3043' 9726' NC 27 1968 1994
3686 Granger Dam 3042' 9720' NC 15 1980 1994
3689 Grapeland 3129' 9529' ET 40 1935 1975
APPENDICES 93
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
4525 Jacksonville 3158' 9516' ET 41 1953 1994
4538 James River Ranch 3032' 9923' EP 28 1948 1975
4556 Jarrell 3049' 9736' NC 68 1926 1994
4563 Jasper 3056' 9400' ET 54 1898 1994
APPENDICES 95
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
5308 Lometa 3113' 9824' EP 18 1948 1965
5327 Long Lake 5 SW 3137' 9551' ET 37 1915 1980
5341 Longview 3228' 9444' ET 93 1902 1994
5348 Longview WSMO 3221' 9439' ET 20 1975 1994
APPENDICES 97
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
6060 Morgan Mill 3222' 9810' NC 46 1949 1994
6070 Morse 3604' 10129' HP 53 1941 1994
6074 Morton 1 WNW 3344' 10247' HP 52 1935 1994
6078 Moscow 3055' 9450' ET 23 1947 1969
APPENDICES 99
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
7044 Pine Springs 1 NE 3154' 10448' TP 15 1939 1994
7060 Pitchfork Ranch 3336' 10032' LRP 24 1971 1994
7066 Pittsburg 5 S 3256' 9458' ET 46 1949 1994
7072 Placid 3119' 9911' EP 18 1948 1965
APPENDICES 101
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
7735 Roosevelt 2 E 3029' 10002' EP 18 1948 1965
7743 Roscoe 3227' 10032' LRP 60 1935 1994
7744 Rosebud 3105' 9658' NC 22 1965 1994
7756 Rosenberg 2933' 9547' UC 22 1915 1960
APPENDICES 103
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
8646 Stillhouse Hollow Dam 3102' 9732' NC 32 1963 1994
8658 Stockdale 4 N 2918' 9758' SC 23 1940 1994
8692 Stratford 3621' 10205' HP 73 1911 1994
8696 Strawn 8 NNE 3240' 9828' NC 46 1949 1994
APPENDICES 105
Appendix 3. Daily precipitation stations in Texas with at least 10 years of annual maxima data through 1994
Continued
Station Climatic Years of Beginning year Ending year
Station name Latitude and longitude
no. region record of record of record
9480 Warren 3037' 9425' ET 51 1935 1992
9491 Washington State Park 3020' 9609' ET 49 1915 1994
9499 Water Valley 3140' 10043' EP 48 1898 1994
9500 Water Valley 11 ENE 3142' 10032' EP 10 1949 1958
APPENDICES 107
Asquith Depth-Duration Frequency of Precipitation for Texas USGS WRIR 984044
District Chief
U.S. Geological Survey
8011 Cameron, Rd.
Austin, TX 787543898