Stochastic Nonparametric Framework for Basin Wide Streamflow and Salinity Modeling

1
Stochastic Nonparametric Framework for Basin Wide
Streamflow and Salinity Modeling Application to
Colorado River basin Study Progress
Meeting James R. Prairie November 29, 2005
2
Progress Update

• Completed Flow Space-Time Disaggregation
• Submitted journal article to Water Resources
  Research
• Describes algorithm as applied to 4 key gauges
• Additionally we performed simple disaggregation
  with entire Colorado River basin and generated
  data required for CRSS
• 100 stochastic simulations ran through CRSS
• Generated values and sequences never seen before
• Basin Wide Salinity Model
• Coupling salinity model and disaggregation model
• Explored multiple methods to model data

3

• Masters Research
• Single site
• Modified K-NN streamflow generator
• Climate Analysis
• Nonparametric Natural Salt Model

Stochastic Nonparametric Technique for Space-Time
Disaggregation
Basin Wide Natural Salt Model
Incorporate Paleoclimate Information
Transition probability matrices (TPM)
Generate streamflow conditioned with TPM

• Policy Analysis
• Impact of drought
• Hydrology
• Water quality

4
Basin Wide Salinity Model

• Coupling
• Nonparametric space-time disaggregation technique
  (Prairie et al., submitted 2005)
• Statistical nonparametric natural salt model
  (Prairie et al., 2005)
• Explored various scenarios
• Direct disaggregation of total salt mass
• Compute total salt mass from flow-salt regression
  
• Simulate annual flows at Lees Ferry
• Compute the associated annual salt from
  regression
• Only space disaggregation for both flow and salt
• Space-time disaggregation for both flow and salt

5
4 Key Gauges

• Directly modeled total natural salt
• Index gauge represents fictitious sum of totals
  at Lees Ferry
• Scenarios
• Direct Disaggregation of computed natural salt
• Model natural salt with annual regression then
  disaggregate the resulting values
• Only spatial disaggregation at annual time step
• Both time and space disaggregation
• Performed 500 simulations
• Dataset natural salt 1971-2003

6
4 Key Gauges in Upper Colorado River Basin
7
Disaggregation scheme
Colorado River near Cisco, Utah
1
Green River near Green River, Utah
2
1
San Juan River near Bluff, Utah
3
2
Colorado River near Lees Ferry, Arizona
4
3
4
Index gauge
5
6
7
8
9
100
1
11
2
122
3
temporal disaggregation annual to monthly at
index gauge
4
spatial disaggregation monthly index gauge to
monthly gauge
8
Bluff
9
Lees Ferry
10
Temporal x-correlation at Lees Ferry
Spatial x-correlation
11
Lees Ferry gauge
May
Annual
12
Lees Ferry gauge Flow and salt correlation
Monthly and Annual
Annual
13

• Preserve flow-salt correlation
• Incorporate Annual Regression
• Only at Lees Ferry
• Perform time-space disaggregation to simulate
  monthly flows at 4 key gauges

14
Basics of the annual flow salt regression

• Dataset
• Annual total natural flow and salt mass from
  1971-2003
• Natural flow observed historic flow
  consumptive use
• /- reservoir regulation
• Natural salt observed historic salt
  - salt added from agriculture
  salt removed with exports
• Nonparametric regression (local regression)
• natural salt f (natural flow)
• Residual resampling

15

• Lees Ferry Gauge
• Annual Flows

Only spatial disagg
Time-Space disagg
16
Lees Ferry
17
Spatial x-correlation
18
Upper Colorado River Basin

• Modeled intervening natural salt
• Index gauge represents sum of intervening salt OR
  total Lees Ferry salt
• Explored multiple scenarios
• Direct Disaggregation of computed natural salt
• Model natural salt with annual regression then
  disaggregate the resulting values
• Only spatial disaggregation at annual time step
• Both time and space disaggregation
• Performed 100 simulations
• Dataset natural salt 1971-2003

19
Upper Colorado River Basin
20
Disaggregation scheme
Index gauge
21

• Lees Ferry
• intervening

22

• Lees Ferry
• Total sum of intervening

23

• total sum of intervening

• intervening

24

• Lees Ferry
• Intervening

May
Annual
25

• Lees Ferry
• Total sum of intervening

May
Annual
26
Lees Ferry gauge Flow and salt correlation Interve
ning
Monthly and Annual
Annual
27
Lees Ferry gauge Flow and salt correlation Total
sum of intervening
Monthly and Annual
Annual
28
Summary

• Coupling the annual flow-salt regression and
  space-time disaggregation technique allowed
  improved preservation of the annual flow salt
  correlation
• Computing natural salt from regression tends to
  smooth higher order functions (i.e., PDF)
• All lower order and higher order statistics are
  adequately preserved for both the 4 key gauge and
  entire Upper Basin schemes
• Previous method did not capture the spatial cross
  correlation
• Paleo-salinity
• Climate conditioned salinity

29
Continued assistance needs

• Complete extension of Lower Basin natural salt
  out to 2003
• Lower Basin for 2004 in January 2006
• Setup annual cycle from this point forward
• Update Lower Basin salt from 1906-1970 with data
  from March 1992 Reclamation report
• Prepare CRSS to accept flow and salt scenarios
• Set up indexed MRM run
• Ensure policy runs are ready to run through CRSS
  for both flow and salt

30
Schedule for research activities

• Time-space disaggregation model September 30,
  2005
• Submission of space-time
• disaggregation paper September 30, 2005
• Basin wide salinity model November 30, 2005
• Basin wide salinity model paper December, 2005
• Analysis and incorporation of
• paleoclimate data February 2006
• Incorporating paleoclimate data paper March 2006
• Policy analysis related to both
• hydrology and salinity June 2006
• Ph.D. Defense August 2006
• Policy analysis paper August 2006

31
Slides to follow are contributions
32
Anticipated Publications

• Prairie, J.R., B. Rajagopalan, U. Lall, A
  stochastic nonparametric technique for space-time
  disaggregation of streamflows, (submitted) Water
  Resources Research, 2005.
• Prairie, J.R., et al., Coupling space-time
  disaggregation model with a natural salt
  estimation model for the Upper Colorado River
  Basin, (in preparation), 2005.
• Prairie, J.R., et al., Generating streamflow
  scenarios using paleo and historic observations
  for the Colorado River at Lees Ferry, Arizona,
  (anticipated), 2006.
• Prairie, J.R., et al., Comparative policy
  analysis with various streamflow scenarios,
  (anticipated), 2006.

33
Relevant Publications

• Prairie, J.R., B. Rajagopalan, T.J. Fulp, and
  E.A. Zagona, Statistical nonparametric model for
  natural salt estimation, ASCE Journal of
  Environmental Engineering, 131(1), 130-138, 2005.
• Prairie, J.R., B. Rajagopalan, T.J. Fulp, and
  E.A. Zagona, Modified K-NN Model for Stochastic
  Streamflow Simulation, (in press) ASCE Journal of
  Hydrologic Engineering, 2005.

34
Acknowledgements

• To my committee and advisor. Thank you for your
  guidance and commitment.
• Balaji Rajagopalan, Edith Zagona, Kenneth
  Strzepek, Subhrendu Gangopadhyay, and Terrance
  Fulp
• Funding support provided by Reclamations Lower
  Colorado Regional Office
• Logistical support provided by CADSWES

35
Extra Slides Follow

36
Local regression
37
Residual Resampling
y y e
Y
X