NAWQA Ecological Synthesis and Studies
  Internal USGS Access Only
NAWQA
Intranet / Public
Overview
NAWQA Ecology
Contacts
Work Plans &
Guidance
Publications
Meetings and
Presentations
Technical and
Laboratory Support
Data
Tools
Training
Related Links

Data > Hydrologic Variability Data > Parameter Descriptions   

Hydrologic Variability Data:

Hydrologic Variability Parameter Descriptions

This page describes and defines the variables which are included in the hydrologic variability datasets. Hydrologic variability parameters are calculated from historical records of mean daily discharge (cfs) over the period of record for each stream gaging station. The variables are grouped into these categories:

  1. Descriptors of the basin and streamflow record
  2. Overall flow variability and predictability
  3. Thresholds and patterns of high/low extremes
  4. Magnitude and duration of extremes
  5. Timing of extremes
  6. Frequency and duration of extremes
Some comments on the types of statistics used here

Two of the parameters included in group 3 above, and most of the parameters in groups 4-6, include both a measure of central location and a measure of spread or dispersion.  Due to the tendency for streamflow data to have a skewed distribution, the central location of the data is indicated by the median rather than by the mean. The spread of the data is described by a coefficient of dispersion. Each variable name is listed only once in these descriptions, whereas in the database those parameters which have a median and coefficient of dispersion will be listed in two forms: the variable name plus an ending of M indicates the median, while the variable name plus an ending of CD indicates the coefficient of dispersion. The coefficient of dispersion is calculated as the (75th percentile value minus the 25th percentile value) divided by the 50th percentile (median) value for all of the parameters except those calculated from dates (see DTMIN, DTMAX below). The CD of these date parameters is calculated as (75th%ile - 25th%ile) / 183 days (Nature Conservancy and Smythe, 1997).

The median and coefficient of dispersion are nonparametric measures for center and spread of data and are usually appropriate when data distributions do not follow a normal curve.  Hydrologic data often have a strongly skewed distribution rather than a normal distribution and are thus often better described by nonparametric measures (Nature Conservancy and Smythe, 1997; Helsel and Hirsch, 1992).

Descriptors of the basin and streamflow record

SUID

 

Study unit 4-letter identification code.

STAID

 

Station identification number.

PO

 

Two-letter state abbreviation (postal identification code) for station location.

STNAME

 

Station name.

YEARB

 

Beginning water year (WY) of the streamflow period of record.

YEARE

 

Ending water year (WY) of the streamflow period of record.

YEART

 

Total number of years in the period of record. In some cases, where there are small gaps in the record, YEART may not equal the difference between YEARE and YEARB. These cases are noted by the following 'flag' variable, YRNOTE.

YRNOTE

 

This is a 'flag' data field, in which an entry of '1' indicates that one or more years of flow data are missing from the period of record which was used to calculate the hydrologic variability statistics. An entry of '0' indicates that there are no gaps or missing years; i.e. that the flow record is continuous from YEARB to YEARE, inclusive. For entries of '1,' details are listed in the comment header lines of any applicable data files.

AREA

 

Watershed area, in square miles, of the stream gage's natural basin, from USGS  NWIS data.


Overall flow variability and predictability

ANNQ

 

Mean annual flow, in cfs. Calculated from annual average mean daily flow values over the entire period of record.

Q_AREA

 

The mean annual flow (ANNQ) per unit of watershed area (AREA). This is a size scalar for comparing streams of different sizes and different watershed areas.

ANNCV

 

Coefficient of variation of ANNQ. In the Nonparametric IHA calculations, this variable is calculated as the (25th%ile-mean)/mean.

PREDQ

 

Flow Predictability based on Colwell's (1974) predictability parameter. PREDQ designates the predictability for all 24-hr mean flows over the period of record. Each value ranges from 0 to 1 and is composed of two additive components, constancy (C) and contingency (M) (Colwell, 1974; Poff and Ward, 1989):

 

 

  • Constancy (C) is a measure of invariance throughout the year. C is maximum when the magnitude of streamflow does not vary throughout the year and is minimum when the flow fluctuates greatly throughout the year.
  • Contingency (M) is a measure of periodicity or seasonality. M is maximum when any particular level of flow is completely dependent on a particular season and minimum when the probability of occurrence of a particular magnitude of flow is independent of the season (i.e. any level of flow can occur at any time of the year).

 

 

 

Maximum predictability (PREDQ) can be the result of any of the following: 

  • Complete constancy (C): the mean daily flow is the same for all seasons in all years, such as the flow approached by an extremely dependable spring-fed stream with uniform flow throughout the year (Poff and Ward, 1989);
  • Complete contingency (M): the mean daily flow is constantly changing throughout the year, but the pattern is the same for all years; i.e. in a particular season the same magnitude of flow always occurs from year to year, and any particular flow magnitude is strongly tied to a particular season in all years.
  • A combination of C and M: the pattern is the same for all years, but some of the same magnitudes of flow occur at different times of the year.

C_PREDQ

 

Constancy/Predictability. This is the proportion of Colwell's total predictability that is a result of constancy (C). Therefore, 1.0 - C_PREDQ = proportion resulting from contingency (M).


Thresholds and patterns of low/high extremes

LOWQ 
HIGHQ

 

Low flow or high flow thresholds, defined in these nonparametric statistics as the 25th and 75th percentiles, respectively. These threshold levels are used in determining the variability parameters of FLOD60D, FLDFREE, and the medians and coefficients of dispersion of the parameters LONUM, HINUM, LODUR and HIDUR.

FLOD60D

 

An index of seasonal flood predictability; specifically, the maximum proportion of all floods from the period of record that occur in any 60-day window common to all years in the record. Ranges from 0.167 (completely random) to 1.0 (perfectly predictable).

FLDFREE

 

Flood free season. An index of seasonal predictability of non-flooding; specifically, the maximum number of days, common to all years in the record, during which no floods have occurred.

ZDAY_

 

Average number of days during the year with zero flow.  There are two ZDAY parameters, ZDAYM and ZDAYCD, which refer to a median (M) and a coefficient of dispersion (CD) for the ZDAY variable.  See the comments above for a further explanation.

BFI_

 

Base flow index, calculated as the 7-day annual minimum flow divided by the annual mean (Nature Conservancy and Smythe, 1997).  There are two BFI parameters, BFIM and BFICD, which refer to a median (M) and a coefficient of dispersion (CD) for the BFI variable.  See the comments above for a further explanation.


Magnitude and duration of extremes (Medians and Coefficients of dispersion)

D1MIN_ 
D7MIN_ 
D90MIN_ 
D1MAX_ 
D7MAX_ 
D90MAX_

 

 

The 6 parameters at left are values for 1-, 7-, and 90-day (D1__, D7__, and D90__) average annual extremes (minimum, __MIN__, and maximum, __MAX__) of streamflow. Again, for each one, there is a long-term median, ending in M, and coefficient of dispersion, ending in CD, resulting in a total of 12 parameters in this group (M and CD for each variable listed at the left). For example, D7MAXM, is the long-term median of the highest 7-day average daily streamflow in each year. 



Timing of Extremes (Medians and Coefficients of dispersion)

DTMIN_ 
DMAX_

 

The numerical day of the year on which the one-day minimum flow (DMIN_) and one-day maximum flow (DMAX_) occurred. Based on Julian calendar day numbering, with January 1 = day 1, and December 31 = day 366 (February 29 included). Note:  the CD of these dates is calculated as (75th%ile-25th%ile)/183 days. 

DTNOTE

 

A warning flag of '1' indicates that the IHA software gave a warning message saying "Dates for maximum and minimum flow are widely distributed throughout the year. Use statistics with caution." Entry '0' indicates that this warning does not apply.  Note:  there is no M or CD for this variable.


Frequency and Duration of Extremes (Medians and Coefficients of dispersion)

LONUM_ 
HINUM_

 

Numerical count of the number of low (LONUM) and high (HINUM) flow pulses, with a pulse defined as a period when the water conditions are greater than the high pulse threshold (HIGHQ) or less than the low pulse threshold (LOWQ) defined above. 

LODUR_ 
HIDUR_

 

The duration, in days, of the low and high flow pulses defined above (LOWQ, HIGHQ). 



U.S. Department of the Interior, U.S. Geological Survey
NAWQA Ecological Synthesis and Studies (INTERNAL USGS ACCESS ONLY)
Please note our privacy statement and disclaimer
Accessibility
Please direct comments and questions to the Webmaster
Last modified on Friday, 20-May-2005 13:03:08 EDT
The URL for this document is http://water.usgs.gov/nawqa-only/ecology/hydro_variability_descriptions.html