DWR_3066022

• Title: WESTERN DAM ENGINEERING NEWSLETTER, VOLUME 5, ISSUE 1 MAY 2017
• Author/Source: AECOM
• Keywords: EMERGENCY RESPONSE, EROSION, CALIBRATION, HYDROLOGIC MODELING, EMBANKMENT DAMS, SEEPAGE, SINKHOLES
• Document Type - Reference Library: Research, Thesis, Technical Publications
• Document Date: 5/1/2017
• Year: 2017
• Team/Office: Dam Safety

Western Dam Engineering
Technical Note

May 2017

18
2. Nash-Sutcliffe Coefficient (𝑁𝑠),
𝑁𝑠=1 −∑[(𝑁𝑖−𝐵𝑖)2
(𝑁𝑖−𝑁̅)2]𝑛
𝑖=1

Where: 𝑁̅= mean of observed data. The 𝑁𝑠
describes the deviation from unity of the ratio
of the square of difference between the
observed and modeled values and the variance
of the observations. The corresponding
accuracy classification is:

0.75 < 𝑁𝑠 ≤ 1.00 Very good
0.65 < 𝑁𝑠 ≤ 0.75 Good
0.50 < 𝑁𝑠 ≤ 0.65 Satisfactory
0.40 < 𝑁𝑠 ≤ 0.50 Acceptable
𝑁𝑠 ≤ 0.40 Unsatisfactory
A model is considered verified and appropriate for use
for a range of conditions and scenarios with a
classification ranging between very good and
satisfactory for both of these statistical comparisons.
Additionally, the MT ESWG suggests that a model can
be considered verified if its results are within one to
two standard deviations from observed or predicted
data.
Common Questions and Additional
Advice
Hydrologic model verifications require that an
engineer juggles a collection of interdependent data,
which can become complex and onerous work. While
developing a fully verified model can be a daunting
task, we’re here to help! Here are a few common
questions and answers to guide you through the
process:
Q1: Can a calibrated model based on measured data
from a single storm event be utilized for other
storm event simulations?
Answer/Suggestion: Yes, provided that antecedent
and initial conditions are similar between the two
events.
Q2: I was able to achieve satisfactory model accuracy;
however, a channel Manning’s n value of 0.005
was adopted during the calibration process. Can I
consider my model to be verified and reasonable?
Answer/Suggestion: No, the adopted Manning’s n
value falls well outside the range of expected and
reasonable values. All model watershed input
parameters must fall within a reasonable range of
values when being varied as part of a calibration
process.
Q3: No extreme storm events, like the PMP, were
observed or measured in my project watershed.
How can I develop an applicable hydrologic model
for an IDF based on an extreme flood event like
the PMF?
Answer/Suggestion: Generally, an infrequent event
with a return period frequency of 100 years or less
can be used to calibrate a hydrologic model for
extreme events. As such, a model calibrated from
gaged data that corresponds to 100-year, 200-year
and 500-year events should be suitable for a PMP
hydrologic simulation. If gage data are not
available for these infrequent events, a pseudo-
calibration utilizing local and regional regression
equations is recommended.
Q4: How often should I review and recalibrate a
hydrologic model?
Answer/Suggestion: Model review is recommended
when the following occur:
 Significant precipitation and/or flood events
have occurred within the project watershed
and/or the region. New data should be used to
revalidate the model. If model accuracy is
unsatisfactory based on the revalidation
process, the model should be recalibrated.
 Watershed conditions have changed (e.g., new
development, wildfire, etc.). In this case, the
model accuracy should be reviewed and the
model should be recalibrated, if necessary,
based on the new watershed conditions.
 The hydrologic model software version and
associated code/computation engine have
been revised/updated. Model results
commonly change as a result of revisions and
updates to software; therefore, existing model
calibrations should be checked within revised
and updated model versions to ensure that