The MFEelRiver_RecordExtension example demonstrates the usage of the Record Extension Analysis in order to extend a shorter annual maximum flow time series using a site with a longer period of record.

Input Data

In this example, time series representing annual maximum flow at two locations within the Eel River watershed in California are analyzed.  The primary (i.e. long record) time series is from the Eel River at Scotia, CA stream gage, which has a drainage area of approximately 3113 square miles (sq mi) and a period of record from 1911 - 2021.  The secondary (i.e. short record) time series is from the Middle Fork Eel River near Dos Rios, CA stream gage, which has a drainage area of approximately 745 sq mi and a period of record from 1966 - 2021.  The location of these stream gages is shown in Figure 1.  The time series are plotted within Figure 2 and tabulated within Table 1.

Figure 1. Stream Gage Locations

Figure 2. Input Time Series for MFEelRiver_RecordExtension Example

Table 1. Input Time Series for MFEelRiver_RecordExtension Example

DateFlow (cfs)
ScotiaDos Rios
20 Jan 1911136000
26 Jan 1912170000
18 Jan 1913150000
22 Jan 1914309000
02 Feb 1915351000
25 Feb 1917292000
07 Feb 191878600
17 Jan 1919149000
16 Apr 192062000
19 Nov 1920148000
19 Feb 1922123000
28 Dec 192273400
08 Feb 192473400
06 Feb 1925127000
04 Feb 1926176000
21 Feb 1927221000
27 Mar 1928233000
04 Feb 192941000
15 Dec 1929120000
23 Jan 193187000
27 Dec 1931127000
17 Mar 193358100
29 Mar 193450900
08 Apr 193579900
16 Jan 1936216000
05 Feb 1937134000
11 Dec 1937345000
03 Dec 1938133000
28 Feb 1940305000
24 Dec 1940150000
06 Feb 1942209000
21 Jan 1943315000
04 Mar 194457800
03 Feb 194599100
27 Dec 1945239000
12 Feb 194786100
08 Jan 1948114000
18 Mar 1949140000
18 Jan 1950117000
22 Jan 1951249000
27 Dec 1951262000
09 Jan 1953215000
17 Jan 1954245000
31 Dec 195452400
22 Dec 1955541000
25 Feb 1957153000
25 Feb 1958202000
12 Jan 1959145000
08 Feb 1960343000
11 Feb 1961113000
14 Feb 1962107000
01 Feb 1963252000
21 Jan 1964178000
23 Dec 1964752000
04 Jan 1966
70200
05 Jan 1966311000
05 Dec 1966154000
29 Jan 1967
42600
14 Jan 1968
40000
15 Jan 1968148000
13 Jan 1969223000
20 Jan 1969
52300
23 Jan 1970
90500
24 Jan 1970310000
04 Dec 1970234000
16 Jan 1971
59900
23 Jan 197213300033000
16 Jan 197315200049200
16 Jan 197438700089500
18 Mar 1975231000
25 Mar 1975
41700
26 Feb 197610900023700
21 Feb 1977
1470
10 Mar 19775790
09 Jan 1978
35800
17 Jan 1978169000
11 Jan 19799610025100
13 Jan 1980
61700
14 Jan 1980226000
28 Jan 198198700
14 Feb 1981
25100
20 Dec 1981300000
16 Feb 1982
67000
26 Jan 1983
76800
27 Jan 1983296000
10 Nov 1983
23600
09 Dec 1983112000
12 Nov 1984133000
08 Feb 1985
20200
17 Feb 198636400093100
12 Mar 1987
25000
13 Mar 198794500
10 Dec 198711800027000
23 Nov 198813700033900
08 Jan 199010200018800
04 Mar 1991
22700
05 Mar 1991105000
20 Feb 19925420010800
20 Jan 1993
33700
21 Jan 1993290000
23 Jan 1994
7220
24 Jan 199448500
09 Jan 199536800065000
12 Dec 199515500030000
01 Jan 1997360000135000
12 Jan 1998
40300
17 Jan 1998170000
08 Feb 1999125000
09 Feb 1999
27500
14 Feb 200016600034600
05 Mar 20015900013000
02 Jan 200211900024800
16 Dec 200222600053700
17 Feb 2004
59700
18 Feb 2004217000
08 Dec 200493800
18 May 2005
32600
30 Dec 2005
107000
31 Dec 2005307000
26 Dec 2006
20600
27 Dec 200674900
04 Jan 2008
42500
05 Jan 2008142000
24 Feb 200964300
02 Mar 2009
18300
25 Jan 2010
17700
26 Jan 201094400
29 Dec 201011900024700
21 Jan 2012
21800
28 Mar 201298000
02 Dec 201220200070300
29 Mar 20147950019400
11 Dec 2014
33000
07 Feb 2015148000
18 Jan 201617700027000
08 Jan 2017
61600
11 Jan 2017252000
07 Apr 201810100026000
27 Feb 201927200062900
26 Jan 20205020017500
02 Feb 2021301005560
24 Oct 2021
28400

Inspecting the 56 years of overlapping data, it is evident that some of the annual peaks are not "truly" concurrent (i.e. the annual peak for a given water year at the long term site occurred during a different runoff event when compared to the short term site), as shown in Figure 4. Per Bulletin 17C: "For the purposes of record extension, concurrent flood peaks are those that occurred in the same water year, not on the same flood event."  As such, the entire overlapping record was considered concurrent within this example.

The Correlation Analysis within HEC-SSP can be used to identify any peak flow data points that are many days apart. For an example of how to create a Correlation Analysis for this purpose, see Using a Correlation Analysis to Select a Suitable Nearby Gage for Record Extension.

Figure 3. Example of Annual Peaks That Are Not From the Same Event

General Tab

A Record Extension Analysis has been developed for this example. To open the analysis, either double-click on the analysis labeled MFEelRiver_RecordExtension from the study explorer or from the Analysis menu select open, then select MFEelRiver_RecordExtension from the list of available analyses. When MFEelRiver_RecordExtension is opened, the General tab within the Record Extension Analysis editor will appear as shown in Figure 4. For this analysis, the MOVE.3 (Bulletin 17C) computational method and the Annual Maximum (One Per Year) data type was selected.  The Scotia time series was selected as the Primary (Long Record) while the Dos Rios time series was selected as the Secondary (Short Record).  The default Year Specification (Water Year) was left unchanged. No modifications were made to the time window or the output labeling.

Figure 4. General Tab

Data Tab

The Data tab contains several tables detailing the complete, concurrent (i.e. overlapping), and non-concurrent (i.e. non-overlapping) records.  Additionally, a plot showing the concurrent record is included.  The  primary record (Scotia) is shown on the x-axis while the secondary record (Dos Rios) is shown on the y-axis.  Prior to a successful compute, the plot will only contain the concurrent record (blue circles), as shown in Figure 5.  Following a successful compute, the extended record (red circles) will be added to the plot, as shown in Figure 6.

Figure 5. Data Tab Prior to a Successful Compute

Figure 6. Data Tab Following a Successful Compute

Computing the Analysis

Once all of the General information has been selected and the data reviewed on the Data tab, the user can press the Compute button to perform the analysis.  Once the computations have been completed, a message window will open stating Compute Complete.

The Record Extension Analysis will use all values within the concurrent time period when computing correlation coefficients, estimators for augmentation, and the extended record.  No values will be removed from consideration even though they may not have resulted from the same runoff event, as was mentioned above.

Record Extension Tab

The Record Extension tab contains several tables detailing the Matalas-Jacobs Estimators (MJ-Mean and MJ-Variance), the Estimators for Augmentation, and statistics of the extended record.  Additionally, a plot showing the input time series is included.  Prior to a successful compute, the plot will only contain the primary record (Scotia; green circles) and the secondary record (Dos Rios; blue circles), as shown in Figure 7.  Following a successful compute, the extended record (red circles) will be added to the plot, as shown in Figure 8.

Figure 7. Results Tab Prior to a Successful Compute

Figure 8. Results Tab Following a Successful Compute

Since this example made use of the MOVE.3 (Bulletin 17C) computational method and the Annual Maximum (One Per Year) data type, following a successful compute, the effective record length (ne) of the variance will be used to select data for extension.  Specifically, the most recent ne years will be selected and the Save Extended Record to New Data Set button will become enabled.

The user may select a different portion of the extended record to save to a new data set.  However, the Estimators for Augmentation were computed using the most recent ne years.  As such, choosing a different portion of the extended record will not result in the Matalas-Jacobs estimators being reproduced within the extended record.  It is recommended that the user selects the default extended record when saving to a new data set.


The Save Extended Record to New Data Set button will only be active following a successful compute.  

When the Save Extended Record to New Data Set is clicked, a dialog will be presented allowing for the definition of a Name, Description, Short ID, and DSS Pathname Parts for the new data set, as shown within Figure 9.

Figure 9. Save As New Data Set Dialog

Upon clicking the Next button, a Save As New Data Set Summary dialog will be shown allowing the user to review the metadata of the new data set as well as a tabulation and plot of the data that is about to be imported, as shown in Figure 10.

Figure 10. Save As New Data Set Summary Dialog

Once the extended record is saved to a new data set, it can be used to compute updated flow-frequency information.  An example of this type of application is shown here.

Report File

In addition to the tabular and graphical results, there is a report file that echoes the input data, selected computational options, and results. To review the report file, press the View Report button at the bottom of the analysis window. When this button is selected a text viewer will open the report file and display it on the screen, as shown in Figure 11. 

Different types and amounts of information will show up in the report file depending on the data and the options that have been selected for the analysis.

Figure 11. Report File