[Manual] [Contents] [Prev] [Next]
In the Scenario Mode, seasonal flows for unregulated, i.e. un-dammed, streams are identified on a daily basis. These can be set by the user simply by drawing headwater seasonal flows or they can be generated from modulators that distribute the total annual headwater runoff according to the historical seasonal patterns.
Unregulated headwater flows connect directly to the river mainstem or to storage reservoirs. For storage reservoirs, the user can set the schedule of outflow according to constraints of the volume of the reservoir and the inflow. System flows are determined by unregulated stream flows and regulated flows from storage reservoir dams.
Headwater Modulation
In the Scenario Mode, flow from unregulated headwaters are modeled by the following equation:
(16) where
- t = Julian day (t = 1 to 365)
- Yt = estimated daily flow
- m = mean annual flow computed over a 10 year period
- p = fraction of mean annual for the scenario
- et = stochastic error term
- Ft = Fourier term
(17)
- ak, bk = Fourier coefficients estimated for each river
= 2
/365
The equation given for Ft above is a smooth Fourier estimate for the annual stream flow for each river, in units of multiples of the mean. For each scenario, an error term is randomly generated to incorporate the expected fluctuations. There tend to be more pronounced deviations from the modeled curve in the wet season (spring), when the exact fluctuations are more difficult to predict. For this reason, the error component is generated from a low variance normal distribution in the dry season, and a higher variance normal distribution in the wet season. Also, since daily flows tend to be highly correlated, the generated (independent) error estimates (rt) are artificially correlated according to the following equation:
(18) where
- rt = randomly generated variable from a normal distribution centered on 0 with variance appropriate for dry and wet years as described above. The switch from dry year to wet year variance parameters occurs at p = 0.4.
- e0 = 0.
The user chooses the type of year to be modeled relative to an average year, which is designated by p = 1. CRiSP.1 multiplies this proportion of the appropriate average flow parameter, m times (Ft + et), which yields an estimate for daily flow for the Scenario Mode flow.
Reservoir Volume and Flow
The storage reservoirs receive flows from the headwaters which are set by the Scenario Flow Modulators or directly by the user. The flow out of the storage reservoirs can be set by the user under constraints established by the maximum and minimum volume of the storage reservoirs. The equation describing the reservoir usable volume is
(19) where
- dV = change in reservoir volume in acre-ft
- dt = time increment, typically 1 day
- FU = unregulated natural flow into the reservoir in kcfs
- FR = regulated flow out of the reservoir, which is controlled by the user under volume constraints in kcfs
The volume for each reservoir is determined a reservoir time step increment from a numerical form of the volume equation
(20) where
- V(i) = reservoir volume time step i with units of acre-ft
t = one day increment
- FU and FR = unregulated and regulated flows in kcfs
- c = 1983.5 is a conversion factor
- acre-ft = (86400 s/d) * (0.023 acre-ft/ k ft3) * (k ft3 / s) * (d)
- V = (86400) * (0.023) *(F) * (
t)
- V = 1983.5 * (F) * (
t)
The user requests reservoir output FR with the following constraints: The user is allowed to draw any flow curve for reservoir withdrawal as long as the reservoir is between minimum and maximum operating volumes. If a request requires a volume exceeding the allowable range, CRiSP.1 alters the request to fit within the volume constraints. The algorithm is
(21)
with constraints on reservoir outflow and volume defined by the algorithm
________________________________________________
- if Vrequest(i+1) > Vmax then
- Vrequest(i+1) = Vmax
- FR(i) = FU(i) + [V(i) - Vmax ] / c
- else
- if Vrequest(i+1) < Vmin then
- Vrequest(i+1) = Vmin
- if Frequest(i) > FU then
- FR(i) = FU(i)
- else
- FR(i) = Frequest(i)
- else
- FR (i) = Frequest(i) (22)
_________________________________________________
where
- FR = outflow from reservoir according to the constraints
- FU = unregulated inflow to reservoir
- Vrequest = requested outflow from reservoir
- Frequest = requested outflow from reservoir
- V(i) = reservoir volume in reservoir time step i
- Vmax = maximum reservoir volume
- Vmin = minimum reservoir volume
Theory for Parameter Estimation
Average daily flow (designated flow_mean) was computed for all available years. Each daily flow was divided by that year's average. Elements of the resulting series were denoted by
, where t = day_of_year. Next, the first nine terms of a Fourier series were computed with a fast Fourier transform. Since the mean of each series was 1, corresponding to the normalized annual mean flow, it follows a0 = 1.0. The remaining Fourier coefficients were estimated according to the equations
(23) where
= 2
/365
- k = value between 1 and 4
The residual time series, Rt were computed by the equation
(24)
The residuals were split into high-variance and low-variance parts, and sample standard deviations computed. mod_start_hi_sigma and mod_end_hi_sigma are the Julian day when high flow variance begins and ends. Period average high and low standard deviation are mod_hi_sigma and mod_lo_sigma, respectively.
Data
The daily flow from Hydrodata, a CD-ROM database marketed by Hydrosphere, Inc., were obtained for the following locations and dates:
- Clearwater River @ Orifino, Idaho: Oct. 1980 - Sept. 1989
- Salmon River @ Whitebird, Idaho: Oct. 1980 - Sept. 1989
- Grande Ronde River @ Troy, Oregon: Oct. 1980 - Sept. 1989
- Imnaha River @ Imnaha, Oregon: Oct. 1980 - Sept. 1989
Flow modulator parameter estimates derived from flow data listed above were compared to modulator parameters estimated from flows over the previous 10 years at the same location (Oct 1970-Sep 1980). The parameters were slightly different, but graphs of smooth flow curves were nearly identical for Clearwater, Salmon, and Imnaha rivers. The Grande Ronde had a different shape, so for this river the parameters were adjusted to include all data from 1970 to 1989 data.
Table 5 shows parameters estimated for the unregulated headwater modulators. Parameters mod_coeffs_a and mod_coeffs_b correspond to ak and bk respectively. Table 6 shows data for regulated headwaters, i.e., Columbia above Grand Coulee Dam, North Fork Clearwater above Dworshak Dam, and Snake River above Brownlee Dam. Daily mean flow observations for each year were obtained from the US Army Corps of Engineers, North Pacific Division and processed as in Table 6. Data were obtained for the following locations and dates:
- North Fork Clearwater River Oct. 1973 - Sept. 1991
- Grand Coulee Dam Oct. 1971 - Sept. 1991
- Brownlee Dam Oct. 1981 - Sept. 1991
Maximum Unregulated Flows
Observed maximum flows in the tributaries were obtained from the peak flow data of Hydrodata, a CD-ROM database marketed by Hydrosphere, Inc. The data record length was variable (Table 7).
Storage Reservoirs Parameter Values
Storage reservoirs volumes are obtained from Project Data and Operating Limits (1989 a and b) and are given in Table 8.
| Table 8 Storage reservoirs. * used in model. |
| Reservoir |
Max Pool ft |
Min Pool ft |
Usable Storage in acre-ft |
Powerhouse Hydraulic Capacity (kcfs) |
| Grand Coulee |
1290 |
1208 |
5,185,500 |
280 |
| Libby Dam |
2459 |
2287 |
4,979,599 |
24.1 |
| Hungry Horse |
3565 |
3336 |
3,161,000 |
8.9 |
| Duncan |
1897 |
1794 |
1,398,600 |
20 |
| Mica |
2478 |
2320 |
7,770,000a |
41.6 |
| *Coulee totalb |
- |
- |
22,494,699 |
- |
| *Dworshak |
1605 |
1445 |
2,015,800 |
10.5 |
| *Brownlee |
2080 |
1976 |
975,318 |
34.5 |
Desired reservoir elevation levels for flood control, obtained from Project Data and Operating Limits (1989 a and b), are presented in Table 9. This is not used by CRiSP.1 at the present time.
[Manual] [Contents] [Prev] [Next]
Columbia River Salmon Passage Model CRiSP.1.5 Theory, Calibration & Validation Manual
Copyright © 1996, Columbia Basin Research. All rights reserved.
web@cbr.washington.edu