WSP11333

• File Number: 8273.100
• Description: Colorado River Basin Salinity Control - Federal Agency Reports - BOR
• Basin: Colorado Mainstem
• Water Division: 5
• Date: 5/1/1975
• Title: Application of Stochastic Hydrology to Simulate Streamflow and Salinity in the Colorado River
• Water Supply Pro - Doc Type: Report/Study

l\:)
I-"
0')
(;"n

To do this, a massive task was undertaken. Hydrologic data were
obtained from the USGS on magnetic tapes which had to be reorganized
for the analysis programs to accept. Data had to be scanned, studied,
analyzed, and interpreted to arrive at a set of information which
would be sound for model use, yet be ready for production studies
by late 1973. A reexamination of these data is expected to provide
significant improvements in the near future.

It should be realized that such a large-scale application of
stochastic generation techniques had probably never before been
undertaken. For this reason, many setbacks and periods of discourage-
ment were encountered, particularly in the elementary stages. Because
the use of stochastic methods was new to the USBR, there was little
"in-house" experience or judgment to rely upon. In passing it should
be made clear that the problems encountered were not entirely due
to the development of the stochastic techniques or to the creation
of new river basin simulation model but rather the problems resulted
from both of these ambitious undertakings, Our experiences have
continually pointed to new areas for refinement and improvement of
both the river basin simulation model (hereafter referred to as the
model) and the stochastic generation of input data for the model
(hereafter referred to as the generation), As applications were
made and the results examined, several topics have been noted which
warrant further investigation. These will be discussed in chapter V
as recommendations for future work.

1.2 Philosophy of Approach

Before discussing the model and generation in more detail, the
philosophy which led to adopting this approach should be set forth.
Because of the complexity of a large water resource system such as
the Colorado River Basin, a computerized model is needed to mathe-
matically describe its behavior. The physical features and operating
processes of a river system usually can be adequately represented by
mathematical formulas and logic which simulate the controls and flow
of water within the system. Once the basin operation is duplicated
and the model is accepted as an accurate reflection of the real sys-
tem, a quest ion arises as to what hydrologic inflows should be used.
It should be noted that the model itself does not require stochastic
inputs and therefore must be judged as accurate based only upon its
own merits. Since the basin operation itself is often a rigid set
of rules for manipulating water, the basic inflow or supply becomes
the major variable in the analysis. Historic records of streamflows
can be used, but they suffer serious drawbacks, One of these is that
a historic trace will never again repeat itself, In addition,
historic traces are too short to study long-term system properties
or rare events which may not be adequately represented. Another

2