Reclamation Mid-Term Operational Modeling Seasonal to Year-Two Colorado River Streamflow Prediction Workshop CBRFC March 21-22, 2011 Katrina Grantz, PhD Upper Colorado Region Hydraulic
Reclamation Mid-Term Operational Modeling Seasonal to Year-Two Colorado River Streamflow Prediction Workshop CBRFC March 21-22, 2011 Katrina Grantz, PhD.
Dec 24, 2015
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Reclamation Mid-Term Operational Modeling
Seasonal to Year-Two Colorado River Streamflow Prediction WorkshopCBRFC March 21-22, 2011
Katrina Grantz, PhDUpper Colorado Region Hydraulic Engineer
Reclamation Operational ModelingOverview
• “Mid-Term” operations for the Colorado River– Operations of major reservoirs in the monthly to 2-
year and beyond timeframe
• 2 operational models– 24-Month Study (deterministic, official)– Mid-Term Ops Model (probabilistic,
additional analysis)
24-Month Study
• Reservoir Operations– 12 major reservoirs (9 UB,
3 LB)
• Monthly timestep, ~2 years, updated monthly
• Used for best guess at mid-term reservoir conditions (storage, elevation, release, hydropower)
24-Month Study: “Official model”
• Annual Operating Plan (AOP) for all reservoirs
• Determines operating tier for Lake Powell– August run of the 24-Month Study (sometimes April)
• Official model projection for determining Lower Basin shortages– Secretary declares a shortage
3 categories of model assumptions
• Inflows• Reservoir operations• Demands
24 Month Study: Inflows
Upper Basin• Forecasted inflows issued by RFC/NRCS• Unregulated inflow • 1 trace
– (3 if min/max month)
Lower Basin• 5-year average for side inflows
Month Issued Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov
Jan RFC RFC RFC
Coord A-J
Coord A-J
Coord A-J
Coord A-J
inter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Feb RFC RFC RFC
Coord A-J
Coord A-J
Coord A-J
inter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Mar RFC RFC RFC
Coord A-J
Coord A-J
inter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Apr RFC RFC RFC
Coord A-J
inter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
May RFC RFC RFCinter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Jun RFC RFC RFCinter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Jul RFC RFC RFCinter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Aug RFC RFC RFCESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
inter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Sep RFC RFC RFCESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
ESP Aug
inter-polate
inter-polate
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
76-05 Avg
Oct RFC RFC RFCESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
inter-polate
inter-polate
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
Nov RFC RFC RFCESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
inter-polate
inter-polate
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
Dec RFC RFC RFCESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
ESP Oct
inter-polate
inter-polate
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
81-10 50%
* Light grey text indicates that the model is run in this month, however, only results for the first 24 months of the model run (black text) are published in the 24 Month Study report
ESP Aug/Oct values are generated using the RFC ESP forecasted volume for the water year using either Aug or Oct initial hydrological conditions. The Aug issued ESP forecast is used for Aug and Sep runs of the 24-month Study. The Oct issued ESP forecast is used for Oct, Nov and Dec runs of the 24-Month Study.
April-July Unregulated Inflow
Interpolated values are calculated by UCBOR and are based on percent of the 76-05 average. The method takes the precent of average of the previous month's RFC or Coord A-J forecast value and interpolates that percentage over two months to 100 percent of average in order to smoothly transition between the end of the current water year and the next water year.
WY 2011 Source of Monthly Unregulated Inflow for Upper Colorado Reservoirs in the 24 Month Study
RFC values are issued by the Colorado Basin River Forecast Center (RFC) as the official forecast values for the next three-month period of time. The values are calculated using Statistical Water Supply (SWS) modeling and/or Ensemble Streamflow Predictions (ESP) modeling. This official forecast has the least amount of error associated with it.
Coord A-J values are official coordinated forecast values issued by the RFC and NRCS for the April-July runoff period using SWS, ESP and VIPER. Forecast error increases with increasing distance from the four-month official forecast period.
76-05 Avg values are the monthly average inflow values generated from water years 1976-2005 calculated using the database maintained by the Bureau of Reclamation Upper Colorado Region (UCBOR). A water year begins October 1 and ends September 30.
Most Probable
24-Month Study: UB Inflows and Model Run Duration (Most Probable)
24-Month Study: UB Inflows and Model Run Duration (Max/Min Prob)
Month Issued Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov
Jan RFC RFC RFC
Coord 90th %ile
Coord 90th %ile
Coord 90th %ile
Coord 90th %ile
inter-polate
inter-polate
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
inter-polate
inter-polate
75th %ile
of 76-05
FebMar
Apr
Coord 90th %ile
Coord 90th %ile
Coord 90th %ile
Coord 90th %ile
inter-polate
inter-polate
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
90th %ile
of 76-05
inter-polate
inter-polate
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
75th %ile
of 76-05
MayJunJul
Aug RFC RFC RFC
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
90th %ile Aug ESP
inter-polate
inter-polate
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
Sep
Oct RFC RFC RFC
90th %ile Oct ESP
90th %ile Oct ESP
90th %ile Oct ESP
90th %ile Oct ESP
90th %ile Oct ESP
90th %ile Oct ESP
90th %ile Oct ESP
90th %ile Oct ESP
90th %ile Oct ESP
inter-polate
inter-polate
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
75th %ile
of 81-10
NovDec
Interpolated values are calculated by UCBOR and are based on percent of the 76-05 average. The method takes the precent of average of the previous month's RFC or Coord A-J forecast value and interpolates that percentage over two months to 100 percent of average in order to smoothly transition between the end of the current water year and the next water year.
90th %ile Aug/Oct ESP values are generated using the RFC ESP forecasted volume for the water year using either Aug or Oct initial hydrological conditions. Monthly values are disaggregated from the total water year volume using the monthly distribution described above for the 90th %ile of 76-05. The Aug issued ESP forecast is used for Aug and Sep runs of the 24-month Study. The Oct issued ESP forecast is used for Oct, Nov and Dec runs of the 24-Month Study.
* Light grey text indicates that the model is run in this month, however, only results for the first 24 months of the model run (black text) are published in the 24 Month Study report
WY 2011 Source of Monthly Unregulated Inflow for Upper Colorado Reservoirs in the 24 Month Study
Maximum ProbableApril-July
Unregulated Inflow
RFC values are issued by the Colorado Basin River Forecast Center (RFC) as the official forecast values for the next three-month period of time. The values are calculated using Statistical Water Supply (SWS) modeling and/or Ensemble Streamflow Predictions (ESP) modeling. This official forecast has the least amount of error associated with it.
Coord 90th percentile values are the official coordinated forecast of the total April-July volume issued by the RFC and NRCS for the April-July runoff period using SWS, ESP and VIPER. Monthly values are disaggregated using the 50th percentile distribution. Forecast error increases with increasing distance from the four-month official forecast period.
90th %ile of 76-05 values are the monthly 90th percentile (90% exceedance) inflow values generated from water years 1976-2005 calculated using the database maintained by the Bureau of Reclamation Upper Colorado Region (UCBOR). A water year begins October 1 and ends September 30.
24-Month Study: Reservoir Operations
• Up-to-date operations input by reservoir operators each month– Manual process: for each reservoir evaluate
inflows, set releases, re-evaluate (sometimes an iterative process)
– Coordination between Powell and Mead
24-Month Study: Demands
Upper Basin• Implicit in unregulated inflow forecast
– Based on assumptions in RFC models (consider historic and current use patterns)
– Adjusts for wet/dry years
Lower Basin• Actual approved water orders for
the year– adjusted for ICS, paybacks, etc
24-Month Study: Output
• AOP (written document)• 24-Month Study Report (mostly tabular data),
monthly update to the AOP
Mid-Term Operations ModelMotivation
• 24-Month Study currently a deterministic model– Upper Basin driven primarily by
most probable inflow forecast – Lower Basin driven by scheduled
demands
• Need to better quantify range of possible operations in the Colorado River Basin
Mid-Term Operations Model
• Model currently in development
• Based on current 24-Month Study model• Accomodates ensemble forecast rather than
most probable inflow forecast• Uses “rules” (prioritized logic) to
set UC reservoir releases rather than manually set by operators
• Model input is range of probable inflows– CBRFC’s ESP forecasts (30 traces) will drive
first and second years of model – Ongoing research to develop
forecasting techniques for beyond 2 years (2-10 yrs)
Mid-Term Operations ModelInflows
• Model currently uses unregulated inflow ESP forecasts – Depletions are implicit in the forecast
• Eventually want to move to natural inflow– Explicitly model water use
Mid-Term Operations ModelInflows
• Rules have been written, tested, and verified to set releases for all upper basin reservoirs
• Good exercise, added documentation, transparency
• Lower basin reservoirs are demand driven– No new rules needed to
be written
Mid-Term Operations Model Operations
Mid-Term Probabilistic Ops Model Model validation
• Compared 24-MS official results against MTOM to verify reservoir rules
• Evaluated min, most, max model runs for months in 2010
• Evaluated elevations and releases using
Mid-Term Operations Model
Mid-Term Operations Model
Mid-Term Operations Model
Questions we hope to better answer…
• Back-to-back 8.23 years? Probability of equalization next year? Balancing? Shortage? What about two years out?
Mid-Term Ops Model: Expected Output
• Probabilistic information and plots– Range of reservoir elevations– Range of reservoir releases– Probability of equalization– Probability of lower basin shortages
Colorado River Hydrology Workgroup
• Research to improve Reclamation’s operations and planning on Colorado River
• Focus on “applied” research
Extra Slides Follow
Regulated Inflow vs. Unregulated Inflow
slope = 1.02R² = 0.99
0%
20%
40%
60%
80%
100%
120%
140%
160%
0% 20% 40% 60% 80% 100% 120% 140% 160%
Regu
late
d A
pr-J
ul I
nflow
(% o
f av
g)
Unregulated Apr-Jul Inflow (% of avg)
Powell InflowRegulated vs Unregulated
ESP run – CDF Powell WY Release
0
2
4
6
8
10
12
14
16
18
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Rele
ase
(acr
e-ft
)M
illio
ns
Percent Exceedance
Lake Powell Water Year Release
Probability of Equalization Estimate Current Methodology
55% Probability of Equalization
Distribution of Observed Inflow Volumes for Remainder of WY
(Provided by RFC and Based on ESP Model Output)
9.52 MAF
Volume determined from October 2009 Most Probable 24-Month Study. Volume required to trigger Equalization in WY2010
Additional Analysis RequestSWE Equalization
Upper Basin SWE Powell Unregulated Inflow
• Significant error in April 1st SWE – Inflow relationship
• Need this info well before April R² = 0.44
0%
20%
40%
60%
80%
100%
120%
140%
160%
0% 20% 40% 60% 80% 100% 120% 140% 160%
Lake
Pow
ell A
pr-J
ul U
nreg
Inflo
w (
% o
f avg
)
Upper Colorado Basin April 1st SWE (% of avg)
April 1st SWE - Apr- Jul Inflow(1990-2009)
What we can provide:
• Regulated inflow volume that would likely trigger equalization – % of average inflows to Powell
that (if forecasted in April) could trigger equalization
• Stakeholders can relate that to other variables
Additional Analysis Request24-Mo Study out-year min and max
Min and Max Runs - Current Practice
• Run in August, October, January, April• Min and Max probable inflows for current
year only– Current year: 10th and 90th percentile official
unregulated UB inflow forecast– Out-year: average historic (1976-2005) UB inflows– LB side inflows use 10th and 90th percentile of last
5 years (current year) and 5-yr avg (out-year)
Min and Max Runs - Request
• Continue Min and Max probable analysis into the out-year – Current year: 10th and 90th forecast– Out-year: 25th and 75th of historic (1976-2005)
inflows • Simulates dry year following
dry year and wet year following wet year
Quick Analysis of Natural Flows
• Considered bottom 10% and top 10% natural flow at Lee’s Ferry (1906-2007)– Following year: wet, normal, or dry (terciles)?
• Makes sense to take min/max analysis into out year (for more reasons than one…)
Dry DryDry NormDry Wet
6 (of 10)2 (of 10)2 (of 10)
Wet DryWet NormWet Wet
0 (of 10)6 (of 10)4 (of 10)
Related Documents
