|
Predicting
the Amount of Available Water
| Predictions: Click on the year or
information of interest
|
Data
and Information Inputs
The amount of
water available in any year to a specific grower is based on many factors:
-
The
precipitation since the start of the water year (measured as snowpack).
-
The water
carried over from the previous water year.
-
The water
rights priority of the distributing canal company.
-
Early
season demand as affected by precipitation.
-
Seasonal
demand as affected by temperature and wind.
Using snowpack
measurements and carryover amounts the NRCS makes monthly
predictions of the amount of stream flow for the coming season. In order
for a grower to be able to predict the amount of water available for
their specific area their water rights priority and historical use must
also be factored in as probability functions.
SNOTEL data
showing seasonal Snow Water Equivalent values. For more SNOTEL
sites check here, for more information check the links page.
Producers start
making production decisions in January so timely information is
needed. Because of the vagaries of the climate, a description of the
probability of water shortage and an estimate of the expected severity
would be beneficial.
Using the NRCS
Surface Water Supply Index (SWSI) stream flow predictions as a starting
point predictions of the probable April through September flow.
The distribution of water rights goes with the classic "First in
time, first in right. different canal companies and users have more senior rights
and hence are first in line for available water. In a typical year the
irrigation demand usually exceeds the natural flow by July first. Each
district's
stored water makes up the deficit for as long as it will last. The
predictions of when the canal will run out of water this year can be found
in the links or here.
Natural flow and diversions from 1993 irrigation
season. By July the demand exceeds supply so storage makes up the
difference
Month by month storage of the upper Snake River reservoirs
showing variability from year to year
Water Right Priorities
for Major Upper Snake Canals
Priority of water rights by decade for the
larger upper Snake canals.
Note Teton Island canal has very senior
rights and should not be in jeopardy. Compare with Twin Falls North
canal where all rights are after 1900. If water rights after 1900 are
cut (as they usually are) then storage water will have to be used if it
is available. To see predictions for your canal, for this year see links
at upper right.
Return
to top
| Statistical
Methodology for Assessing
Probability of Surface Water Supply Shortage
This presentation discusses the statistical
method behind the specific predictions for each canal.
|
Slide 1: Showing the typical natural flow
through Milner if no dams or diversions were on the river. Also
shown is the demand for the water which exceeds the supply at the
end of June. The deficit is made up by storage from the larger
reservoirs of Jackson, Palisades and American Falls. 1980
irrigation season.
|
 |
| In the 1992
irrigation season the supply was much lower so storage was
used almost the whole season |
|
|
|
|
|
Total
Basin Storage is the sum of: •
October
31st reservoir
carryover and Basin
Initial Storage from basin runoff. It varies from 2
million acre feet to 3.5 million acre feet. It is largely a function of the spring flow from
snowpack. The wetter the winter the more water is
available.
|
|
|
| Using
the NRSC predictions for streamflow we can estimate the
amount of water that will be available for the Upper Snake
system. The 95% interval can cover about a million acre feet
of variability. |
|
|
| Using
the Aberdeen Canal as an example the amount of storage available
depends on the total basin storage. |
|
|
|
The amount of storage water
required for Aberdeen can be as much as 200,00 acre feet but
has an inverse dependency on the amount of runoff. In wet
years the demand on storage is reduced.
|
|
|
|
So a prediction interval can
be determined for the storage water required a based on the
system runoff |
|
|
|
When the combined allocated
storage and required storage graphs are combined it becomes
evident that if basin runoff is low there will be a deficit.
In this scenario with only 0.3 million acre feet of
carryover. If basin runoff is below 4.3 million acre feet
there will probably not be enough water to meet demand. If
there were more carryover from previous seasons then the
probability of shortage would be reduced. The amount of
overlap between the prediction intervals is the shortage
probability. |
 |
|
When there is only 300,000
acre feet of water for carryover No matter how much stream
flow is predicted, there will likely be shortages. |
|
|
|
With more carryover (1
million acre feet) the probability of shortage is greatly
reduced. |
 |
|
With 2 million acre feet of
carryover the probability of shortage (overlap) almost disappears.
Even in the driest years there won't be a shortage. The
worst shortages happen in multiple reduced water years when
the storage is depleted.
|
 |
|
Statistically, we are looking
at the overlap in the tails between two normal curves or
storage required and storage allocated . If storage
allocated is much above the required amount then the
probability of the deficit is small. |
|
|
|
If the Allocated amount is
close to the Storage amount then the probability of deficit
is greatly increased.
|
|
|
The
probability Pof a water shortage is the probability that
storage water requirement will exceed storage water
allocated.
•P(Shortage)
= P(Required > Allocated)
•P(Shortage)
= P(Allocated – Required)
< 0
The area of the
"tail" in the lower left corner represents the
probability of a shortage. |
|
|
System wide the probability
of a shortage is certain if flow is below 3.5 million acre
feet, but disappears if flow is above 5 million acre feet. |
 |
|
Here it is expressed as a
percentage.
|
 |
|
Runoff cumulative
probability |
 |
|
Multiplying Basin Runoff by
Water Shortage Probability |
 |
The
forecast shortage probability is calculated as a
conditional probability:
Forecast Shortage Probability =
∫
P(Runoff) x
P(Shortage|Runoff)
|
 |
The
data requirements for forecasting Shortage Probability
are:
• October
31 reservoir storage carryover
•NOAA-Northwest
River Forecast Center estimates of April through
September runoff volumes for different probabilities of
exceedence.
|
Return to top |
|
