FLOOD07067

• County: Larimer
• Community: Big Thompson Canyon
• Stream Name: Big Thompson
• Basin: South Platte
• Title: Peak Discharge Estimates In Refinement of the Big Thompson Storm Analysis
• Date: 5/5/1978
• Prepared For: Conference on Flash Floods
• Prepared By: American Meteorological Society
• Floodplain - Doc Type: Flood Documentation Report

..

/'.

,.

Wllh4VtA 'P :S~ t..V1

Reprintd frca PrepJ1.r:l't Vol,..: CQ1lterlIlCI on Flalh Flood.:
Ib'4rOMtlorologlC&l A.peetl, Nay 2-5. 1918. IQI Angltl..,
Calif. Pu.blhhed by thl.Amll'lcUl MltllOl'ololleal Society.
Ioltotl.. Ma...

PEAK DISCHARGE ESTIMATES USED IN REFINEMENT OF THE BIG THOMPSON STORM ANAlYSIS

Donald L. Miller, Clarence E. Everson, James A. Mumford, and Frederick A. Bertle

Flood Section, Hydrology Branch, Division of Planning Technical Services
Bureau of Reclamation, Denver, Colorado 80225

1.

INTRODUCTION

The severe storm which caused the
disastrous flash flood in the Big TholllJlson Canyon
near Estes Park, Colorado on 31 July 1976, ranks
among the major storms of record in the United
States. The meteorological aspects of this storm
have been studied in considerable detail and
results published by Maddox et al. (1977).
Meteorological factors leading up to and during
the storm were covered in detail by them with
elllJlhasis on identifying similar situations in the
future for forecasting ,purposes. The Bureau of
Reclamation started to analyze this storm in
detail soon after its occurrence for the purpose
of integrating it into future storm and flood
studies for spillway design purposes. It was
soon realized that storm analysis based only on
the available rainfall would not explain the peak
discharge data estimated by the Geological Survey
(Grozier et al., 19.7fH.~ R<<.....tructed ,puk.s
based on preliminary rainfall analysis were on
the order of 25 to 50 percent of the Geological
Survey estimates for many locations. Therefore,
efforts were directed toward derivation of a
storm pattern and time distribution of rainfall
which would produce discharges more in agreement
with the Geological Survey estimates. Many
difficulties were encountered because of complex-
ities inherent in such a problem. Judgment,
experience, and trial and error procedures were
utilized. The use of automated hydrograph compu-
tations (Mumford, 1975) and streamflow routing
(Mumford, 1978) made this possible. The area
studied was limited to the contributing areas on
the main stem of the Big Thompson River and on
the North Forte as indi cated by the hatched
boundary line in Figure 1. The most severe part
of the storm occurred within this area and
although extension of the study to other drain-
ages to the north and south would have been
desirable, manpower limitations did not permit
their inclusion.

2.

DISCUSSION

Thunderstorm conditions developed
over the B1g Thompson River Basin Saturday
afternoon, 31 July 1976. Intense rainfall
occurred from 1830 to 2230 MDT between Estes
Parte and Drake. Additional burst-type rain
occurred between midnight and 0300 or 0400 MDT on
Sunday, 1 August. Light rain occurred on Sunday
and Sunday night, in some locations becoming
intense. The National Weather Service prepared
an isohyetal map of total precipitation for the
period 31 July through 2 August 1976, soon after
the event and revised the pattern slightly in

April 1977 (National Weather Service, 1977).
This pattern is too general for identification of
specific rainfall causing the flash flOOding of
Saturday night in the Big Thompson drainage
basin. Peak discharges occurred at Glen Comfort
about 2000 MDT, at Drake about 2100 MDT, and at
the mouth of the canyon at 2300 MDT on 31 July
1976.

Original bucket survey data sheets
were examined in this study and total storm
rainfall and the amounts occurring Saturday night
were li sted. When only total 2-day rainfall was
reported, Saturday night rainfall was estimated
from locations where both Saturday night and
total storm rainfall were indicated.

Peak discharge estimates indicated
that the maximum rainfall must have occurred over
subbasins 5, 6, and 15. However, buck.et survey
~4Ma~ic.ted ~.~-"ight centers near Glen
Comfort to the south and Glen Haven to the north
of these areas. Both of these indicated centers
were about 10 inches. At this point, both the
location of the centers and their magnitude were
questioned an'd reexamined. In connection with a
study of the storm potential for two small
drainage areas in the desert region of Southern
California, Clarte (1971) related duration to
precipitation/moisture ratio for severe thunder-
storms. A curve on log-log paper with a slope
of about 0.5 based on data for a number of
thunderstorms in the Western United States was
used in that study. Based on dewpoint data for
Goodland, Kansas on 31 July 1976, and precipi-
table water from the surface to 18,000 feet
(assuming saturation), this curve would indicate
a potential of 15 inches maximum point rainfall
in 3 hours for the Big Thompson storm. The
precipitation pattern sh~n in Figure 2 for the
8-1/2 hour period 1830 MDT 31 July to 0300 MDT
1 August 1976, was developed by trial, using the
above considerations for location and intensity
of the centers and testing rainfall estimates by
ru noff computati ons.

Rainfall centers estimated as exceed-
ing 14 inches were positioned over subbasins
5 and 15 because of the high discharges estimated
from those streams. Bucket survey amounts
guided the positioning of decreasing isohyets.
The flood-producing area was assumed to be
that enclosed by the 2-inch isohyet.

There were no hourly rainfall data
recorded in the Big Thompson drainage basin
within the intense storm area. Two brief showers
totaling less than 0.5 inch wer~ measured by an

135